Multi-Level Issues in Organizational Behavior and Leadership

MULTI-LEVEL ISSUES IN ORGANIZATIONAL BEHAVIOR AND LEADERSHIP

RESEARCH IN MULTI-LEVEL ISSUES Series Editors: Francis J. Yammarino and Fred Dansereau

Previous Volumes:

Volume 1: The Many Faces of Multi-Level Issues – Editors, Francis J. Yammarino and Fred Dansereau Volume 2: Multi-Level Issues in Organizational Behavior and Strategy – Editors, Fred Dansereau and Francis J. Yammarino Volume 3: Multi-Level Issues in Organizational Behavior and Processes – Editors, Francis J. Yammarino and Fred Dansereau Volume 4: Multi-Level Issues in Strategy and Methods – Editors, Fred Dansereau and Francis J. Yammarino Volume 5: Multi-Level Issues in Social Systems – Editors, Francis J. Yammarino and Fred Dansereau Volume 6: Multi-Level Issues in Organizations and Time – Editors, Fred Dansereau and Francis J. Yammarino Volume 7: Multi-Level Issues in Creativity and Innovation – Editors, Michael D. Mumford, Samuel T. Hunter, and Katrina E. Bedell-Avers RESEARCH IN MULTI-LEVEL ISSUES VOLUME 8

MULTI-LEVEL ISSUES IN ORGANIZATIONAL BEHAVIOR AND LEADERSHIP

EDITED BY FRANCIS J. YAMMARINO State University of New York at Binghamton, NY

FRED DANSEREAU State University of New York at Buffalo, NY

United Kingdom – North America – Japan India – Malaysia – China JAI Press is an imprint of Emerald Group Publishing Limited Howard House, Wagon Lane, Bingley BD16 1WA, UK

First edition 2009

Reprints and permission service Contact: [email protected]

No part of this book may be reproduced, stored in a retrieval system, transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without either the prior written permission of the publisher or a licence permitting restricted copying issued in the UK by The Copyright Licensing Agency and in the USA by The Copyright Clearance Center. No responsibility is accepted for the accuracy of information contained in the text, illustrations or advertisements. The opinions expressed in these chapters are not necessarily those of the Editor or the publisher.

British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

ISBN: 978-1-84855-502-0 ISSN: 1475-9144 (Series)

Awarded in recognition of Emerald’s production department’s adherence to quality systems and processes when preparing scholarly journals for print CONTENTS

ABOUT THE EDITORS ix

LIST OF CONTRIBUTORS xi

ACKNOWLEDGMENTS xiii

OVERVIEW: MULTI-LEVEL ISSUES IN ORGANIZATIONAL BEHAVIOR AND LEADERSHIP Fred Dansereau and Francis J. Yammarino 1

PART I: ORGANIZATIONAL BEHAVIOR

A NEW KIND OF ORGANIZATIONAL BEHAVIOR Francis J. Yammarino and Fred Dansereau 13

THICK OR THIN? A FUNDAMENTAL QUESTION IN ORGANIZATIONAL BEHAVIOR Neal M. Ashkanasy 61

A NEWER ORGANIZATIONAL BEHAVIOR Francis J. Yammarino and Fred Dansereau 69

PART II: OUTSTANDING LEADERSHIP

CHARISMATIC, IDEOLOGICAL, AND PRAGMATIC LEADERSHIP: AN EXAMINATION OF MULTI-LEVEL INFLUENCES ON EMERGENCE AND PERFORMANCE Michael D. Mumford, Samuel T. Hunter, Tamara L. 79 Friedrich and Jay J. Caughron

v vi CONTENTS

LEVELS OF PERFORMANCE: MULTI-LEVEL PERSPECTIVES ON OUTSTANDING LEADERSHIP James G. (Jerry) Huntw and John N. Davis 117

PRESIDENTIAL LEADERSHIP STYLES: HOW DO THEY MAP ONTO CHARISMATIC, IDEOLOGICAL, AND PRAGMATIC LEADERSHIP? Dean Keith Simonton 123

CHARISMATIC, IDEOLOGICAL, AND PRAGMATIC LEADERSHIP: WHERE WE ARE, AND WHERE DO WE NEED TO GO? Michael D. Mumford, Jay J. Caughron and 135 Tamara L. Friedrich

PART III: LEADERSHIP AND SOCIAL RELATIONS

A COMPONENTIAL ANALYSIS OF LEADERSHIP USING THE SOCIAL RELATIONS MODEL David A. Kenny and Stefano Livi 147

CONSIDERATIONS IN APPLYING THE SOCIAL RELATIONS MODEL TO THE STUDY OF LEADERSHIP EMERGENCE IN GROUPS: A LEADERSHIP CATEGORIZATION PERSPECTIVE Rosalie J. Hall, Robert G. Lord and Katey E. Foster 193

THOUGHTS ON STUDYING LEADERSHIP IN NATURAL CONTEXTS Stefano Livi and David A. Kenny 215

PART IV: LEADERSHIP SIMULATION

A LEVELS-BASED LEADERSHIP SIMULATION: INSIGHTS REGARDING GROUP DECISION OPTIMIZATION Shelley D. Dionne and Peter J. Dionne 227 Contents vii

COMPARING SIMULATION RESULTS OF LEADERSHIP STYLE IMPACTS ON EMERGENT VERSUS SPECIFIC TASK OUTCOMES AND REQUIRED SIMULATION MODEL COMPONENTS Janice A. Black, Richard L. Oliver and Lori D. Paris 271

MAKING IT PRACTICAL: SIMULATION, NATURALISTIC DECISION MAKING, AND COMPLEXITY IN TEAM PERFORMANCE Jessica L. Wildman and Eduardo Salas 301

SINS OF OMISSION AND ENVY: REDEMPTION AND SALVATION THROUGH LEVELS OF ANALYSIS Shelley D. Dionne and Peter J. Dionne 321

PART V: ENVIROSCAPES

ENVIROSCAPES: A MULTI-LEVEL CONTEXTUAL APPROACH TO ORGANIZATIONAL LEADERSHIP Richard Reeves-Ellington 337

TARGETING THE CULTURAL PROCESSES OF PARTNERING FOR ANALYSIS Elizabeth K. Briody 421

ENVIROSCAPES: THE CHALLENGES OF CULTURAL PARTNERING CONCEPTS Richard Reeves-Ellington 431

PART VI: ABOUT THE AUTHORS

ABOUT THE AUTHORS 447 This page intentionally left blank

ABOUT THE EDITORS

Francis J. Yammarino, Ph.D., is SUNY Distinguished Professor of Manage- ment and Director and Fellow of the Center for Leadership Studies at the State University of New York at Binghamton. He received his Ph.D. in Organizational Behavior (Management) from the State University of New York at Buffalo. Dr. Yammarino has extensive research experience in the areas of superior–subordinate relationships, leadership, self–other agreement processes, and multiple levels of analysis issues. He has served on the editorial review boards of eight scholarly journals, including the Academy of Management Journal, Journal of Applied Psychology, Journal of Organizational Behavior, Leadership Quarterly, Organizational Research Methods,and Personnel Psychology. Dr. Yammarino is a Fellow of the American Psycho- logical Society and the Society for Industrial and Organizational Psychology. He is the author of 13 books and has published more than 100 articles. Dr. Yammarino has served as a consultant to numerous organizations, including IBM, Textron, TRW, Lockheed Martin, Medtronic, United Way, Skills Net, and the US Army, Navy, Air Force, and Department of Education. Fred Dansereau, Ph.D., is Professor of Organization and Human Resources and Associate Dean for Research in the School of Management at the State University of New York at Buffalo. He received his Ph.D. from the Labor and Industrial Relations Institute at the University of Illinois with a specialization in Organizational Behavior. Dr. Dansereau has extensive research experience in the areas of leadership and managing at the individual, dyad, group, and collective levels of analysis. Along with others, he has developed a theoretical and empirical approach to theorizing and testing at multiple levels of analysis. He has served on the editorial review boards of the Academy of Management Review, Group and Organization Management, and Leadership Quarterly. Dr. Dansereau is a Fellow of the American Psycholo- gical Association and the American Psychological Society. He has authored 12 books and more than 80 articles and is a consultant to numerous organizations, including the Bank of Chicago, Occidental, St. Joe Company, Sears, TRW, the United States Army and Navy, Worthington Industries, and various educational institutions.

ix This page intentionally left blank

LIST OF CONTRIBUTORS

Neal M. Ashkanasy UQ Business School, The University of Queensland, Brisbane, Queensland, Australia Janice A. Black Department of Management and Marketing, School of Business and Public Administration, California State University, Bakersﬁeld, CA Elizabeth K. Briody General Motors R&D, Warren, MI Jay J. Caughron Department of Psychology, The University of Oklahoma, Norman, OK Fred Dansereau School of Management, State University of New York at Buffalo, Buffalo, NY John N. Davis Hardin-Simmons University, Kelley College of Business, Abilene, TX Shelley D. Dionne Binghamton University, School of Management and Center for Leadership Studies, Binghamton, NY Peter J. Dionne Sensis Corporation, East Syracuse, NY Katey E. Foster Department of Psychology, University of Akron, Akron, OH Tamara L. Friedrich Department of Psychology, The University of Oklahoma, Norman, OK Rosalie J. Hall Department of Psychology, University of Akron, Akron, OH James G. (Jerry) Huntw Texas Tech University, Lubbock, TX Samuel T. Hunter Department of Psychology, Penn State University, University Park, PA xi xii LIST OF CONTRIBUTORS

David A. Kenny Department of Psychology, University of Connecticut, Storrs, CT Stefano Livi Department of Social and Developmental Psychology, University of Rome ‘‘La Sapienza,’’ Rome, Italy Robert G. Lord Department of Psychology, University of Akron, Akron, OH Michael D. Mumford Department of Psychology, The University of Oklahoma, Norman, OK Richard L. Oliver Department of Accounting and Information Systems, College of Business, New Mexico State University, Las Cruces, NM Lori D. Paris Department of Management and Marketing, School of Business and Public Administration, California State University, Bakersﬁeld, CA Richard Reeves- School of Management, Binghamton Ellington University, State University of New York, Binghamton, NY Eduardo Salas Department of Psychology, and Institute for Simulation and Training, University of Central Florida, Orlando, FL Dean Keith Simonton Department of Psychology, University of California at Davis, Davis, CA Jessica L. Wildman Department of Psychology, and Institute for Simulation and Training, University of Central Florida, Orlando, FL Francis J. Yammarino School of Management and Center for Leadership Studies, State University of New York at Binghamton, Binghamton, NY ACKNOWLEDGMENTS

The publication of the Research in Multi-Level Issues annual series and this volume have been greatly facilitated by Rachel Brown and Emma Smith at Emerald Publishing Group in the United Kingdom, and the staff at Macmillan Publishing Solutions, India. Closer to home, we thank our Schools of Management, the Center for Leadership Studies at Binghamton, and the Jacobs Management Center at Buffalo as well as our assistants, Marie Iobst and Cheryl Tubisz, and our copyeditor, Jill Hobbs, for their help in preparing this book for publication. Finally and perhaps most importantly, we offer our sincere thanks to our contributors. The authors of the essays, commentaries, and rebuttals in this volume have provided new ideas and insights for unraveling the challenges of dealing with multiple levels of analysis and multi-level issues in a wide variety of areas. Thank you all.

xiii This page intentionally left blank

OVERVIEW: MULTI-LEVEL ISSUES IN ORGANIZATIONAL BEHAVIOR AND LEADERSHIP

Fred Dansereau and Francis J. Yammarino

INTRODUCTION

Multi-Level Issues in Organizational Behavior and Leadership is Volume 8 of Research in Multi-Level Issues, an annual series that provides an outlet for the discussion of multi-level problems and solutions across a variety of fields of study. Using a scientific debate format of a key scholarly essay followed by commentaries and a rebuttal, we present, in this series, theoretical work, significant empirical studies, methodological developments, analytical techniques, and philosophical treatments to advance the field of multi-level studies, regardless of disciplinary perspective. Similar to Volumes 1 through 7 (Yammarino & Dansereau, 2002, 2004, 2006; Dansereau & Yammarino, 2003, 2005, 2007; Mumford, Hunter, & Bedell-Avers, 2008), Volume 8 contains five major essays with commentaries and rebuttals that cover a range of topics, but in the realms of organizational behavior and leadership. In particular, the five ‘‘critical essays’’ offer extensive literature reviews, new model developments, methodological advancements, and some data for the study of organizational behavior, outstanding leadership, leadership and social relations, leadership simulation, and enviroscapes. While each of the major essays, and its associated

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 1–9 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008018 1 2 FRED DANSEREAU AND FRANCIS J. YAMMARINO commentaries and rebuttals, is unique in orientation, all of the essays share a common bond in raising and addressing multi-level issues or discussing problems and solutions that involve multiple levels of analysis in organizational behavior and leadership.

ORGANIZATIONAL BEHAVIOR

In the first essay, following from the cutting-edge work of Stephen Wolfram in A New Kind of Science (2002), we (as authors) propose ‘‘a new kind of OB’’ (organizational behavior) based on the varient approach to theory building and testing. In particular, we offer four simple, yet comprehensive theories to account for individual behavior, interpersonal relationships, group dynamics, and collectivized processes in organizations. We believe that our approach differs from contemporary approaches to OB in two ways. First, we provide four sets of two constructs, each of which is proposed in different multi-level configurations. Second, we propose very simple theories that we suggest may underlie very complex processes. Thus, although the theory is multivariate (eight variables), the constructs, their association, and the levels of analysis are considered two at a time. While the four sets of variables and levels and potential interactions proposed here seem simple notions, we believe that taken both alone and in combination, they may explain a variety of complex phenomena and behavior in organizations. In his commentary, Ashkanasy notes some possible problems in our essay and suggests some alternative strategies. He identifies three core areas of concern. The first is that the essay sets out a ‘‘thin’’ theory, which is at odds with the idea that OB in real organizations is inherently complex and addressable only though ‘‘thick’’ descriptions. Second, while the theory covers four levels of analysis, Ashkanasy feels that we may have neglected the time dimension. Third, he suggests that the theory seems an example of ‘‘grand’’ theorizing, suggesting it might also share the disappointing fate of such theories in the past. In our reply to Ashkanasy’s commentary, we argue that we are not taking a logical positivist approach that ignores the complexity of the situation. We suggest that some basic processes may underlie what may appear to be complex processes. We also agree that the time dimension is an important part of our theory and point out where in the essay we considered this issue. Finally, we argue that even if the result of empirical work does not support a grand theory, we believe following this line of research has the potential to Multi-Level Issues in Organizational Behavior and Leadership 3 make a significant contribution to the field of OB, particularly, in terms of including multiple levels of analysis in the field.

OUTSTANDING LEADERSHIP

In the second essay, Mumford, Hunter, Friedrich, and Caughron begin by pointing out that theories of outstanding, historically notable, leadership have traditionally emphasized charisma. Recent research, however, suggests that charisma may represent only one pathway to outstanding leadership. Outstanding leadership, according to these authors, may also emerge from ideological and pragmatic leadership. In their essay, they examine the multi- level conditions influencing the emergence and performance of charismatic, ideological, and pragmatic leaders. They argue that different conditions operating at the environmental, organizational, group, and individual levels influence the emergence and performance of each of these three types of leaders. Implications for understanding the origins and impact of charismatic, ideological, and pragmatic leaders are discussed. This essay clearly extends the field by adding multiple levels of analysis to the area of outstanding leadership. In regard to the commentary by Hunt and Davis, we were deeply saddened by the death of Jerry Hunt, who was one of the most open-minded professionals whom we have ever met. It was in his biannual research series on leadership that one of the editors of this volume, Fred Dansereau, had his first article on levels of analysis published. To our knowledge, it was the first article published on levels of analysis in leadership. Jerry will be deeply missed both by the editors of this volume as a person, supporter, and colleague, and by the field for his superior intellectual capability as a scholar. In line with Jerry’s open-minded approach to the field, in their commentary, Hunt and Davis attempt to push the work of Mumford, Hunter, Friedrich, and Caughron forward by asking how scholars might use their work to make predictions about outstanding leadership and the conditions that might be ideal for the emergence of each of the three types of outstanding leadership. Three of the questions that they ask provide a direction for future research based on Mumford et al.’s work. First, Hunt and Davis state that all other things being equal, according to Mumford et al., the strongest case for the emergence and performance of outstanding charismatic leadership would seem to be when trust is present at the group level (at the time of emergence) and under conditions of 4 FRED DANSEREAU AND FRANCIS J. YAMMARINO socio-technical change and lack of elite consensus at the environmental level. They then suggest that this idea allows future researchers to follow Mumford et al.’s lead and address the question of when these conditions that favor charismatic leadership might prevail. Second, they state that all other things being equal, according to Mumford et al., the strongest case for the emergence and performance of ideological leadership would seem to be when leadership can be shared at the group level, when culture is strong at the organizational level, and in collectivist cultures facing social disruption at the environmental level. They suggest that this notion allows future researchers to build on Mumford et al.’s work and address the question of when these conditions that favor ideological leadership might prevail. Finally, Hunt and Davis state that all other things being equal, according to Mumford et al., the strongest case for the emergence and performance of pragmatic leadership would seem to be when support is based on elite reactions, when the leader possesses strong cognitive skills at the individual level, and when the leader takes action to maintain perceptions of fairness at the group level, the amount of follower professionalism at the organization level, and consensus among elites at the environmental level. As was typical of Jerry Hunt’s work, the question then becomes how one makes advances based on previous work. In this case, Hunt and Davis again suggest that this assumption allows future researchers to follow Mumford et al.’s lead and address the question of when these conditions that favor pragmatic leadership might prevail. In his commentary, Simonton raises the question of whether the general framework in the Mumford et al. essay applies to more focused domains of leadership. More specifically, Simonton discusses his own research on leadership styles in the U.S. presidency – interpersonal, charismatic, deliberative, creative, and neurotic – and then examines how these five styles have some correspondence to the three broad types of extraordinary leadership discussed by Mumford et al. His essay also includes a table that provides scores for all U.S. presidents on two key dimensions: charismatic – creative and interpersonal – deliberative. We believe that this integrative thinking and these applications make an excellent contribution and addition to the area of outstanding leadership. Mumford, Caughron, and Friedrich, in their reply, suggest that the comments by Hunt and Davis and by Simonton raise a variety of questions about how the charismatic, ideological, and pragmatic leadership styles should be measured and how hypotheses should be developed with regard to multi-level influence on leader emergence and performance. They effectively Multi-Level Issues in Organizational Behavior and Leadership 5 respond to such issues but also point out that, although evidence is often available for charismatics, it is rare for studies to have contrasted charismatics with ideological and pragmatic leaders. Thus the authors hope their work and the work of Hunt and Davis and of Simonton will provide the impetus for future studies that address the issues raised by the commentators.

LEADERSHIP AND SOCIAL RELATIONS

In the third essay, Kenny and Livi present their social relations model (SRM), which explicitly proposes that leadership simultaneously operates at three levels of analysis: group, dyad, and individual (perceiver and target). According to the authors, by using their model, researchers can empirically determine the amount of variance at each level as well as those factors that explain variance at these different levels. This essay attempts to show how the SRM can be used to address many theoretically important questions in the study of leadership and can be used to advance both the theory of and research in leadership. For example, based on their analysis of leadership ratings from seven studies, Kenny and Livi suggest that there may be substantial agreement (i.e., target variance) about who in a group is the leader. In a second analysis, the authors examine the effects of gender and gender composition on the perception of leadership. They also explore how self-ratings of leadership may differ from group members’ perceptions of leadership. Finally, they discuss how their model can be used with conventional software. This essay offers an extension of the SRA approach to leadership from a multi-level perspective. In their commentary, Hall, Lord, and Foster critically consider limits to the generalization of the variance components analysis results described in Kenny and Livi’s first example, and briefly summarize results of an additional study that supports the original findings. They also suggest interpretational issues and problems of interest to researchers who may wish to continue to apply the SRM to multi-level issues in the study of leadership. In their response, Livi and Kenny focus on whether the SRM variance partitioning would be the same when groups are long term in duration and have formal leaders. In particular, they speculate about how the variance partitioning might change if these dimensions change. They also consider the design and analysis issues, as well as the estimation of group effects, in natural workgroups using their SRM approach. 6 FRED DANSEREAU AND FRANCIS J. YAMMARINO

LEADERSHIP SIMULATION

In the fourth essay, Dionne and Dionne begin by pointing out that previous literature has compared the effectiveness of different styles of leadership, yet most of this research has not compared different levels of analyses regarding leader styles or behaviors. In their essay, these authors develop and present a computational model and describe a levels-based comparison of four types of leadership that represent three different levels: individual, dyad, and group. When examined across a dynamic group decision-making optimization scenario, group-based leadership is found to produce decisions that are closer to optimal than do dyadic- and individual-based leadership. An alternative computational model, in which individual cognitive and experience-based components vary among group members, also indicates that group-based leadership produces more optimal decisions. The essay offers an introduction that discusses simulation as a theoretical development tool and supplies additional evidence related to the use of simulation methods in leadership research. In their commentary, Black, Oliver, and Paris begin by pointing out that clear specification of leadership efforts spanning levels of analysis has lagged behind leadership research in general. Black et al. use agent-based modeling, along with Dionne and Dionne’s choices of leadership styles, to examine the impact of those styles on the generation of an emergent group resource, context-for-learning (CFL), instead of the specific task outcome (group decision making) studied by Dionne and Dionne. In their work, they find consistent effectiveness across leadership styles for workgroups with both high and slightly lower initial individual levels of a CFL. Using a second agent-based model that includes the ability of agents to forget previous learned skills, these authors then report a reduced effectiveness of all leadership styles. However, the effectiveness of the leadership styles differs between the two outcomes (specific group task model and emergent group resource model). Black et al. note reasons for these differences and describe the implications of the comparisons of the two multi-levels models. In their commentary, Wildman and Salas also begin by highlighting the lack of focus on multi-level issues within leadership research. They suggest that while the work of Dionne and Dionne makes a strong contribution to the sciences of leadership, group decision making, and team complexity, many aspects of these authors’ research demonstrate potential for great expansion and improvement. In their commentary, Wildman and Salas discuss and provide suggestions related to the topics of computer simulation in team research, group decision-making theory, and the modeling of team Multi-Level Issues in Organizational Behavior and Leadership 7 complexity. The authors’ intention is to stimulate continued critical thinking and inspire more innovative, practical, and carefully designed multi-level research efforts. In their reply, Dionne and Dionne focus on the importance of multi-level issues in their simulation. They point out that Wildman and Salas suggest more descriptive decision-making models and more sophisticated simulation techniques would improve the practicality of their work. They also note that Black, Oliver, and Paris employ an agent-based model within an emergent task context to examine a leader’s influence on group context-for-learning and, in doing so, found differences from their own work. Dionne and Dionne demonstrate the practicality of their model and contrast their approach with the suggested additional simulations. Their reply, along with their earlier work and that of the commentators, offer an exciting glimpse into the future of group decision-making research from a multi-level perspective.

ENVIROSCAPES

In the fifth essay, Reeves-Ellington offers a paradigm for understanding organizational leadership realities through a multi-level understanding of the organizational environments of climate, knowledge, ethos, and time. To do so, he presents and discusses five enviroscapes: climate, knowledge, ethos, time, and leadership. The author suggests that each of these enviroscapes has two phenotypes: business and commerce. Each of these enviroscapes, with its concomitant phenotypes, is viewed as being used differently at multiple levels of management and leadership by senior managers, middle managers, and entry-level managers. After reviewing these conceptual and theoretical ideas, Reeves-Ellington applies the theory and model to an extended-time case study of land purchase in Indonesia by a U.S. pharmaceutical firm. This essay is likely to surprise readers who are more accustomed to traditional organizational behavior essays, in that it takes an anthropological approach to understanding leadership. In her commentary, Briody focuses on the case study presented at the end of the essay by Reeves-Ellington. Specifically, she attempts to examine the successful integration of an expanding U.S. pharmaceutical firm into Indonesia’s multicultural environment, by viewing the situation as a context marked by the interweaving of market exchange and reciprocity exchange. Her analysis directs her attention to the interactions occurring both among key leaders within the firm, and between those leaders in the firm and their 8 FRED DANSEREAU AND FRANCIS J. YAMMARINO counterparts in the peasant and governmental communities. By focusing on the cultural processes of partnering, Briody attempts to show the contribution of cooperative, healthy relationships in achieving the firm’s business goals. In his response, Reeves-Ellington responds to Briody’s three major areas of concern related to the original case: methodology, theoretical concepts, and leadership processes. He argues that Briody takes an anthropological approach, whereas the original case took the perspective of a business practitioner/researcher. The response essentially casts Briody’s observations as encouraging an ongoing dialogue among academic, practitioner, and anthropological researchers.

CONCLUSION

The essays, commentaries, and replies in this book illustrate the kind of issues that arise in dealing with multiple levels of analysis in organizational behavior and leadership. The deﬁnitions of concepts (i.e., organizational behavior, outstanding leadership, leadership and social relations, leadership simulation, and enviroscapes) change depending on which combination of levels of analysis is involved and added to them. The nuances of analytical methods (i.e., multi-level quantitative or qualitative in nature) change when one moves from one level of analysis to multiple levels of analysis. Moreover, although different paradigms may guide different scholars’ theories and research methods and techniques, levels of analysis issues must be resolved to have a viable paradigm (i.e., traditional or novel). We believe that the explorations of these issues in organizational behavior and leadership found in this volume show that these insights, applications, and advances will apply to numerous areas of scholarly investigation. The authors in this volume have challenged theorists, researchers, and methodologists to raise and address multi-level issues in all their disciplinary and interdisciplinary work. If you would like to be a part of contributing ideas to this scholarly endeavor, please contact us directly or visit our website at www.levelsofanalysis.com.

REFERENCES

Dansereau, F., & Yammarino, F. J. (Eds). (2003). Multi-level issues in organizational behavior and strategy. Vol. 2 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Multi-Level Issues in Organizational Behavior and Leadership 9

Dansereau, F., & Yammarino, F. J. (Eds). (2005). Multi-level issues in strategy and methods. Vol. 4 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Dansereau, F., & Yammarino, F. J. (Eds). (2007). Multi-level issues in organizations and time. Vol. 6 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Mumford, M. D., Hunter, S. T., & Bedell-Avers, K. E. (Eds). (2008). Multi-level issues in creativity and innovation. Vol. 7 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Yammarino, F. J., & Dansereau, F. (Eds). (2002). The many faces of multi-level issues. Vol. 1 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Yammarino, F. J., & Dansereau, F. (Eds). (2004). Multi-level issues in organizational behavior and processes. Vol. 3 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Yammarino, F. J., & Dansereau, F. (Eds). (2006). Multi-level issues in social systems. Vol. 5 of Research in Multi-Level Issues. Oxford, UK: Elsevier.

This page intentionally left blank

PART I ORGANIZATIONAL BEHAVIOR

This page intentionally left blank

A NEW KIND OF ORGANIZATIONAL BEHAVIOR

Francis J. Yammarino and Fred Dansereau

ABSTRACT

Following from the cutting-edge work of Stephen Wolfram in A New Kind of Science (2002), in this chapter we propose ‘‘a new kind of OB’’ (organizational behavior) based on the varient approach to theory building and testing. In particular, we offer four simple, yet comprehensive theories to account for individual behavior, interpersonal relationships, group dynamics, and collectivized processes in organizations. In each case, two constructs, their association, and the levels of analysis of their operation are proposed. While the four theories proposed here are simple notions, they can explain a variety of complex phenomena and behavior in organizations.

INTRODUCTION

Extraordinary claims require extraordinary evidence. – Carl Sagan, Billions and Billions (1997)

The evidence is crummy. There’s a much simpler explanation. – Carl Sagan, The Demon-Haunted World: Science as a Candle in the Dark (1996)

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 13–60 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008001 13 14 FRANCIS J. YAMMARINO AND FRED DANSEREAU

In 2002, Stephen Wolfram – PhD in theoretical physics (at age 20), MacArthur award recipient, and creator of the Mathematica software system – published his tome A New Kind of Science. Wolfram’s key point, which aimed to revolutionize how we view and conduct ‘‘science’’ (broadly defined), was that simple notions can explain complex phenomena. He asserted and then demonstrated in a variety of disciplines and areas (i.e., mathematics, physics, biology, social sciences, computer science, philosophy, art, technology, artificial intelligence, artificial life, catastrophe theory, chaos theory, complexity theory, computational complexity theory, cybernetics, dynamical systems theory, evolution theory, experimental mathematics, fractal geometry, general systems theory, nanotechnology, nonlinear dynamics, scientific computing, self-organization, and statistical mechanics) that simple rules (ideas, notions, theories) can lead to simple or complex phenomena. Moreover, Wolfram demonstrated that more complex or complicated ‘‘rules’’ do not ultimately lead to more complex behavior; using more complicated rules may be ‘‘convenient’’ but they do not add fundamentally new features (p. 62). In line with Wolfram’s postulations, with which we wholeheartedly agree, in this chapter we propose ‘‘a new kind of OB’’ (organizational behavior) based on the varient approach to theory building and testing (see Dansereau, Alutto, & Yammarino, 1984; Dansereau, Cho, & Yammarino, 2006; Dansereau & Yammarino, 2000, 2006; Yammarino, 1994, 1998; Yammarino & Markham, 1992). Specifically, following from the key assertion made by Wolfram, we propose four simple, yet comprehensive theories to account for both simple and complex organizational behavior. These theories, again analogous to the crisp nomenclature of Wolfram, are called Theory 1, Theory 2, Theory 3, and Theory 4. They attempt to understand and explain (1) individual behavior and decision making, (2) interpersonal relations and leadership, (3) group dynamics/team processes and norms, and (4) collectivized processes and roles, respectively. In each case, a theory consists of two constructs, their association, and the levels of analysis of their operation. The four proposed theories that form a new kind of OB are summarized in Table 1. Before elaborating the details of each ‘‘simple’’ theory, we offer a brief summary of those portions of Wolfram’s work that are relevant for our approach and propositions here. We also review briefly ‘‘an old kind of OB’’ – the current approach to building and testing theories in organizational behavior. We then present an overview of a contrasting approach – the varient approach – to generating and testing theories for a new kind of OB. Next, we develop each of the four theories for understanding behavior in A New Kind of OB 15

Table 1. Four Theories for a New Kind of OB.

Theory Entities/Levels of Variables/Relationship Analysis

þ Theory 1: individual behavior Whole persons Option cutting ! commitment and decision making þ Theory 2: interpersonal Whole dyads Investments ! returns relations and leadership þ Theory 3: group dynamics/ Whole groups Interdependence ! cohesion team processes and norms þ Theory 4: collectivized Whole collectives Titles ! expectations processes and roles Theory 1: level-specific at the person level Theory 2: level-specific and emergent at the dyad level Theory 3: level-specific and emergent at the group level Theory 4: cross-level to the collective level organizations in terms of persons, dyads, groups, and collectives. In each case, we assert and justify a general formulation of a theory and then offer a more specific formulation for empirical testing in organizations. Finally, we discuss the implications of our approach and theories for future work in organizational behavior including multi-level empirical testing and the potential integration of our ideas for developing a new kind of OB.

A NEW KIND OF SCIENCE

Wolfram (2002) presents a comprehensive new paradigm for science in general as well as for its many disciplines, ﬁelds, and subﬁelds in particular. Our purpose here is not to review his work in its entirety, but simply to provide an overview of key points to justify and demonstrate Wolfram’s position and our own position that simple ideas are valuable, are preferred, and explain a great deal of both simple and complex behavior for all of science, including OB.

Simple Rules

Wolfram (2002) uses the notion of rules – ideas or theories – to describe and explain behavior and states that ‘‘in reality even systems with extremely simple rules can give rise to behavior of great complexity’’ (p. 110). Moreover, he notes that ‘‘the simpler a structure is, the more likely it is that it will show 16 FRANCIS J. YAMMARINO AND FRED DANSEREAU up in a wide diversity of different places. And, this means that by studying systems with the simplest possible structure, one will tend to get results that have the broadest and most fundamental significance . . . looking at systems with simpler underlying structures gives us a better chance of being able to tell what is really responsible for any phenomena one sees – for there are fewer features that have been put into the system and that would lead one astray’’ (p. 109). This is a goal we hope to accomplish with the four simple theories for OB proposed here. This approach, according to Wolfram, leads to an interesting possibility: ‘‘to consider completely random initial conditions . . . one might think that starting from such randomness no order would ever emerge. But in fact . . . many systems spontaneously tend to organize themselves, so that even with completely random initial conditions, they end up producing behavior that has many features that are not random at all’’ (p. 223). When starting with a fixed or given state, there are three basic types of behavior/patterns identified by Wolfram: simple or repeating; nested or fractal; and complex or random. In terms of randomness, ‘‘something should be considered random if none of our standard methods of perception and analysis succeed in detecting any regularities in it’’ (p. 556). Regularities are repetitions or nesting patterns. More formally, ‘‘something should be considered random whenever there is essentially no simple program [rule, theory, model] that can succeed in detecting regularities in it’’ (p. 556). In terms of complexity, ‘‘when we say that something seems complex what we typically mean is that we have not managed to find any simple description of it – or at least those features of it in which we happen to be interested’’ (p. 557). The simplest descriptions are repetition and nesting: If we don’t ‘‘see’’ these, then we consider things to be complex. Relevant for our work here in OB, our four proposed theories are simple ones that can detect, predict, and help understand regularities in behavior. Other, more complex and complicated theories in OB appear to do no better in detecting regularities, as evidenced by the oftentimes nonreplicability of studies and/or findings in the field, and so may be merely tapping randomness. In Wolfram’s words, ‘‘From the intuition of traditional science we might think that if the behavior of a system is complex, then any model for the system must also somehow be correspondingly complex. But . . . this is not in fact the case, and . . . even models that are based on extremely simple underlying rules can yield behavior of great complexity’’ (2002, p. 364). Our contention is that our four simple theories proposed here meet these conditions – that is, beyond being ‘‘simple,’’ they can account for or explain complex behavior and actions of various entities. A New Kind of OB 17

Wolfram goes on to state, ‘‘Typically it is not a good sign if the model ends up being almost as complicated as the phenomena it purports to describe. And it is an even worse sign if when new observations are made the model constantly needs to be patched in order to account for them’’ (2002, p. 365). Unfortunately, this pattern is often observed in OB, where new variables are constantly added to models, increasing their complexity, to account for unexplained behavior. In contrast, ‘‘[it] is usually a good sign . . . if a model is simple, yet still manages to reproduce, even quite roughly, a large number of features of a particular system. And it is an even better sign if a fair fraction of these features are ones that were not known, or at least not explicitly considered, when the model was ﬁrst constructed’’ (Wolfram, 2002, p. 365). This goal, in fact, underlies our proposing of four simple theories to account for OB at the person (individual), dyad (interpersonal), group (team), and organizational (collective) levels of analysis.

Computations and Rules

Wolfram (2002) creates simple models for explaining complex ‘‘everyday systems’’ such as the growth of crystals, the breaking of materials, the flow of fluids (e.g., air and water), fundamental issues in biology (e.g., molecular structure and natural selection), the growth of plants and animals, biological pigmentation patterns, and financial systems. He also demonstrates how simple models can explain complex phenomena in fundamental physics (e.g., conservation of energy, equivalence of direction in space, models of the universe, space–time and relativity, elementary particles, gravity, and quantum phenomena) and in processes of perception and analysis (e.g., randomness, complexity, data compression, visual and auditory perception, statistical analysis, cryptography and cryptanalysis, mathematical formulas, and human thinking). Wolfram accomplishes this by focusing on the notion of computation. For him, systems can be viewed as simple computer programs or in terms of the computations they can perform. The initial conditions are the input; the state of the system after some number of steps corresponds to the output. Also, different systems may have very different internal workings but the computations the systems perform may be very similar, such that ‘‘any system whatsoever can be viewed as performing a computation that determines what its future behavior will be’’ (p. 641). A further notion Wolfram discusses is universality: ‘‘if a system is universal, then it must effectively be capable of emulating any other system, and 18 FRANCIS J. YAMMARINO AND FRED DANSEREAU as a result it must be able to reproduce behavior that is as complex as the behavior of any other system’’ (p. 643). He notes that cellular automata, Turing machines, substitution systems, and register machines are examples of systems that, despite the great differences in underlying structures, can be made to emulate each other – that is, ‘‘universals.’’ Also, ‘‘any system whose behavior is not somehow fundamentally repetitive or nesting will in the end turn out to be universal’’ (p. 698) and ‘‘universality is in a sense just associated with general complex behavior’’ (p. 713). This behavior results from simple rules and from altering the initial conditions. More specifically, the general underlying hypothesis for Wolfram’s whole paradigm is the principle of computational equivalence (PCE). It applies to any kind of process, whether natural or artificial. The key underlying idea that leads to PCE is the notion that ‘‘all processes, whether they are produced by human effort or occur spontaneously in nature, can be viewed as computations’’ (p. 715). PCE asserts that ‘‘when viewed in computational terms there is a fundamental equivalence between many different kinds of processes . . . almost all processes that are not obviously simple can be viewed as computations of equivalent sophistication’’ (pp. 716–717) and that ‘‘even extremely simple rules can be universal’’ (p. 718). While we are not dealing with computations per se here, our four proposed theories of OB are simple rules that are universal in Wolfram’s sense. Moreover, PCE ‘‘introduces a new law of nature to the effect that no system can ever carry out explicit computations that are more sophisticated than those carried out by systems like cellular automata and Turing machines’’ (Wolfram, 2002, p. 720). PCE suggests ‘‘that beyond systems with obvious regularities like repetition and nesting most systems are universal, and are equivalent in their computational sophistication’’ (p. 735). Again, this idea also applies to the four theories of OB presented in this chapter. Lastly, ‘‘even though a system may follow definite underlying laws its overall behavior can still have aspects that fundamentally cannot be described by reasonable laws’’ (Wolfram, 2002, p. 750). According to Wolfram, this idea explains the phenomenon of free will. In other words, PCE explains and helps us understand why persons, dyads, groups, and collectives can follow option cutting/commitment, investments/returns, interdependence/cohesion, and titles/expectations, respectively, yet still show variation in behavior that is not accounted for by the theories (rules) per se. In short, systems/entities have ‘‘free will’’; thus, in OB as in science in general, despite our simple theories’ attempts to account for a variety of complex behavior, we can still expect some variability in behavior of persons, dyads, group, and collectives. A New Kind of OB 19

AN OLD KIND OF OB

Complex Rules

In contrast to the approach of Wolfram (2002), the ﬁeld of OB is replete with complicated and complex ‘‘rules’’ (ideas, notions, theories) that do not appear to add new explanations for the behavior and actions of systems/ entities. While not wishing to single out any particular publication outlet (because we view this issue as a general problem for OB journals), a typical article published in the Academy of Management Review proposes a theory with ‘‘multiple boxes and arrows,’’ all of which are interconnected to form a complex web of relationships. These complex explanations, which often are never tested or at least never fully tested or replicated in subsequent publications, strain to account for behaviors that may be explained by simpler ‘‘rules,’’ such as those offered in this chapter. These ideas are reﬂected in the quotations at the outset of this chapter from Carl Sagan, who some have suggested may be the second greatest scientist of the twentieth century (after Albert Einstein) based on not only his discoveries but also his role in making science readily accessible to the general public. In brief, the quotes from Sagan’s work indicate that there are too many theories and not enough data, too many untested or weakly tested theories, and weak evidence for complicated theories, especially when simpler explanations are available. These ideas seem to readily apply to OB, where they are further supported by weak or at best marginal results in meta-analyses on most complex (rather than simple) theories.

Levels of Management versus Levels of Analysis

Another problem with current approaches in OB is that they often ignore entities (or systems) per se, instead focusing and building upon on levels of management or assuming levels of analysis (without making them explicit) or both. Even now, despite more than two decades of levels work, there is still confusion and misunderstanding about levels of analysis versus levels of management issues. These issues have been discussed in detail for a variety of areas in OB in general (see Dansereau et al., 1984, 2006; Dansereau, Yammarino, & Kohles, 1999; Dansereau & Yammarino, 2000, 2003, 2005, 2007; Yammarino & Dansereau, 2002a, 2004, 2006; Yammarino, 1994; Yammarino & Markham, 1992) as well as for speciﬁc areas within OB such as leadership research (see Dansereau & Yammarino, 1998a, 1998b; Yammarino, Dionne, Chun, & Dansereau, 2005; Yammarino, 1998). 20 FRANCIS J. YAMMARINO AND FRED DANSEREAU

Levels of Management: Organizational Chart (4 Levels)

------

Levels of Analysis: 15 Persons, 14 Dyads, 7 Groups, 2 Collectives, 1 Organization

Fig. 1. Levels of Management and Levels of Analysis.

While levels of analysis issues are more fully explicated later in this chapter, Fig. 1 begins this clarification and highlights the differences between levels of management and levels of analysis. The same people within the same organization are shown in the upper (levels of management) and lower (levels of analysis) portions of the figure, but a different configuration or view occurs by rearranging the nodes in the figures (adapted from Wolfram, 2002, on network systems). The nodes are individuals who, in the upper portion of the figure, are placed in an organizational chart of four levels of management (with the CEO, e.g., at the top). In the lower portion of the figure, these same individuals (with the CEO in the middle) account for various levels of analysis – that is, 15 persons, 14 dyads (one-to-one relationships), 7 groups (or teams), 2 collectives (e.g., departments), and 1 organization – and can be A New Kind of OB 21 viewed in terms of any of these ‘‘lenses.’’ Note that specifying and/or testing levels of management is not the same thing as specifying and/or testing levels of analysis.

Summary

Given the previously identiﬁed issues, it is time to dispense with ‘‘an old kind of OB’’ and begin to focus on a new kind of OB that values simple ‘‘rules’’ (Wolfram, 2002) or what we might call ‘‘little ideas that are well tested’’ rather than complex ‘‘rules’’ or what we might refer to as ‘‘big ideas that are poorly tested.’’ These simple theories should ideally include the smallest number of constructs and variables possible and should explicitly incorporate entities, systems, or levels of analysis. As noted by Chaitin, ‘‘a theory has to be simpler than the data it explains, otherwise it does not explain anything . . . The simpler the theory, the better you understand something’’ (2006, pp. 76–77). In this sense, we are well served by William of Occam and Occam’s razor – a rule of thumb that states when faced with two hypotheses (theories or ‘‘rules’’) that explain the data or evidence equally well, one should choose the simpler one. So, following the tradition of Occam, Sagan, and Wolfram, we offer four simple theories and a comprehensive approach for testing and gathering better evidence about them for a new kind of OB.

A NEW KIND OF OB: THE VARIENT APPROACH

Endorsing the notion that simple rules, ideas, and theories are better and explain as much or more than complex ones (e.g., Einstein’s famous equation E ¼ mc2), we wish to formulate theories with a minimal number of constructs that explain the behaviors and actions of entities or systems. In fact, the absolute minimum (simple) specification for a ‘‘theory’’ is two constructs, with a relationship specified between them, about something (i.e., entities or objects of study), as exemplified by Einstein’s equation for energy. Our four proposed theories have these characteristics and are notions that have ‘‘deep roots,’’ as they are linked to traditional or classical ideas. We begin this discussion with a focus on the entities or levels of analysis.

Levels of Analysis

Levels of analysis issues and multiple-level approaches are becoming increasingly important in many areas of OB and closely related research (see 22 FRANCIS J. YAMMARINO AND FRED DANSEREAU

Dansereau & Yammarino, 1998a, 1998b, 2003, 2005, 2007; Yammarino & Dansereau, 2002a, 2004, 2006). Various scholars (Dansereau et al., 1984, 1999; House, Rousseau, & Thomas-Hunt, 1995; Klein, Dansereau, & Hall, 1994; Rousseau, 1985; Yammarino et al., 2005) have noted the importance of clearly specifying the levels of analysis at which phenomena are expected to exist theoretically, and have stated that it is critical to ensure that the measurement of constructs and data-analytic techniques correspond to the asserted levels of analysis, so that inference drawing is neither misleading nor artifactual. Levels of analysis are inherent in theoretical formulations. In some formulations, the levels of analysis are implicit or assumed. In other cases, levels of analysis are used to formulate the boundary conditions under which a theory is expected to hold. In still other instances, theories, propositions, and hypotheses explicitly incorporate levels of analysis as an integral component of the formulation. Understanding how and if levels are specified permits an examination of the potential for or degree of prevalence of theoretical misspecification. Moreover, identification of relevant levels of analysis issues may help account for mixed, inconsistent, and contradictory findings in prior research. Without explicit incorporation of levels of analysis issues, incomplete understanding of a construct or phenomenon may lead to faulty measures, inappropriate data-analytic techniques, and erroneous conclusions. Theoretical revolutions in science often emerge when other levels of analysis are considered. For example, a revolution in biology occurred when some theorists suggested, and subsequently demonstrated, that evolution can occur at a level of analysis higher than the organism level. Likewise, a well-known revolution in physics arose when some theorists asserted, and subsequently demonstrated, that quantum mechanics operate at a level of analysis lower than the atomic level. In this same way, OB theory building can advance when we include lower and higher levels of analysis in theory development and hypothesis generation. Levels of analysis are the entities or objects of study. In the current work, we are interested in human beings in work organizations. Entities are typically arranged in hierarchical order such that higher levels (e.g., groups) include lower levels (e.g., persons), and lower levels are embedded in higher levels (see Dansereau et al., 1984; Yammarino, 1996; Yammarino & Bass, 1991). In the various areas of OB research, four key levels of analysis of human beings are relevant: individuals or persons (independent human beings), dyads (two-person groups and interpersonal relationships), groups (workgroups and teams), and organizations (collectives larger than groups A New Kind of OB 23 and groups of groups) (see Dansereau et al., 1984; Yammarino, 1996; Yammarino & Bass, 1991). First, human beings in organizations can be viewed as individuals or persons, independent of one another. In this case, we can focus on an employee, a manager, a leader, or a follower/subordinate, or how these individuals differ from one another. Individual differences are of interest here. Second, human beings in organizations can be viewed as dyads,ortwo individuals who are interdependent on a one-to-one basis. A dyad is a special case of groups – that is, a two-person group. In this case, we can focus on superior–subordinate dyads, leader–follower dyads, peer–peer dyads, coworker–coworker dyads, or interpersonal relationships, independent of the formal workgroup. Third, human beings in organizations can be viewed as groups or teams. While there are some potential differences between groups and teams, we view them similarly here – as a collection of individuals who are interdependent and interact on a face-to-face or ‘‘virtual’’ (non-colocated) basis with one another. Formal workgroups or teams generally consist of a leader or a manager and his or her immediate followers or direct reports. Fourth, human beings in organizations can be viewed as collectives.In this case, the focus is on clusterings of individuals that are larger than groups and that are interdependent based on a hierarchical structuring or a set of common or shared expectations. Collectives include groups of groups, departments, functional areas, strategic business units, and organizations. They often do not involve direct interaction among people (as in groups), but rather are held together by echelons or hierarchies. These four levels of analysis – person, dyad, group, and collective – represent different perspectives on the human beings who make up organizations. In this sense, they can be thought of as different lenses through which human beings can be observed. A key characteristic of these levels is their embeddedness; for example, two persons make up a dyad, multiple persons make up a group, multiple dyads make up a larger group, and multiple groups make up a collective. In other words, as one views human being from increasingly higher levels of analysis, the number of entities decreases (e.g., there are fewer collectives than groups in an organization), and the size of the entities increases (e.g., collectives include a larger number of human beings than do groups).

Wholes and Parts

In our approach, there are four alternatives to consider for each level of analysis (see Table 2). Two of these alternatives are plausible views of the 24 FRANCIS J. YAMMARINO AND FRED DANSEREAU

Table 2. Summary of Single-Level Formulations.

Alternative Members of Associations among Between-Entities Within-Entities Views of Units Unit Members Differences Differences Entities

Wholes Homogeneous Positive Systematic Error Parts Heterogeneous Negative Error Systematic Equivocal Independent Independent Systematic Systematic Inexplicable Not relevant Not relevant Error Error

focal entities (parts and wholes as units of analysis), and two of them indicate that focal entities are not relevant but that other entities may be plausible (equivocal and inexplicable). We distinguish conceptually between two different views of any level of analysis (also see Lerner, 1963). A wholes view is defined as a focus between entities but not within them; differences between entities are viewed as valid, and differences within entities are viewed as error (random). This perspective can be described as a between-units case (Glick & Roberts, 1984; Pedhazur, 1982). In this instance, members of a unit are homogeneous, the whole unit is of importance, and relationships among members of units with respect to constructs of a theory are positive. Relationships among theoretical constructs are a function of differences between units. A parts view is defined as a focus within entities but not between them; differences within entities are valid, and differences between entities are erroneous. This perspective can be described as a within-units case (Glick & Roberts, 1984; Pedhazur, 1982)orafrog pond effect (Firebaugh, 1980). In this instance, members of a unit are heterogeneous, a member’s position relative to other members is of importance, and relationships among members of units with respect to constructs of a theory are negative. Relationships among theoretical constructs are a function of differences within units. These two views – wholes and parts – are conceptually different ways to indicate that a particular level of analysis is relevant for understanding constructs and variables of interest. In addition to permitting effects at a particular level of analysis, various authors indicate that effects may not be evidenced at that level (Lerner, 1963; Miller, 1978; Pedhazur, 1982). Thus the focal level is considered not relevant, and other levels must be considered. In one case, there is a focus both between and within entities at a focal level. Determining whether a wholes or parts view is occurring is difficult because both between- and within-entities differences are valid. Thus the A New Kind of OB 25 focal level of analysis does not clarify our understanding of the constructs and variables of interest. Consequently, other levels must be considered. If the assumption is made that only one level of analysis can be considered, then seemingly both conditions (wholes and parts) are occurring. Because there are always other levels of analysis to consider (Miller, 1978), however, this condition must be viewed as equivocal – between- and within-entities differences are equally likely. The more parsimonious conclusion is that neither wholes nor parts views at the focal level are appropriate (Dansereau et al., 1984; Yammarino & Markham, 1992). In this instance, members of a unit are independent, members are free of the unit’s influence, and relationships among members of units with respect to constructs of a theory are independent. In short, relationships among theoretical constructs are a function of differences between members (e.g., persons) independent of higher-level units (e.g., groups). Another possibility – namely, error or lack of focus between and within entities – is an inexplicable or traditional null view of a focal level. In this case, the focal level also is not relevant for understanding the theoretical constructs of interest; instead, other levels of analysis should be specified conceptually. In summary: Wholes are homogeneous entities that display similarity among members, where between-entities differences are systematic and within-entities differences are error. Parts are heterogeneous entities that display complementarity among members, where within-entities differences are systematic and between- entities differences are error. Equivocal reflects independence among members, where between- and within-entities differences are systematic and other entities should be considered. Inexplicable indicates a null case where between- and within-entities differences are error and other entities should be explored. Given these alternatives, it becomes a matter of selecting among them based on theory and data analysis.

Multiple Levels

Beyond these single levels of analysis (i.e., individuals, dyads, groups, or collectives viewed separately), a key issue is that of multiple levels of analysis. In other words, levels can be viewed in combination or simultaneously 26 FRANCIS J. YAMMARINO AND FRED DANSEREAU

Table 3. Summary of Multiple-Level Formulations.

Multiple-Level Formulation Lower-Level View Higher-Level View

Cross-level wholes Wholes Wholes Cross-level parts Wholes Parts Level-speciﬁc wholes Wholes Equivocal Level-speciﬁc parts Parts Inexplicable Emergent wholes Equivocal Wholes Emergent parts Equivocal Parts Equivocal Equivocal Equivocal Inexplicable Inexplicable Inexplicable

Source: See Dansereau et al. (1984, p. 186) for eight additional (null) alternatives and their interpretation.

(see Table 3). In these cases, we are concerned with multi-level or cross-level effects, as well as with mixed determinants and mixed-level effects (for details and a review, see Dansereau et al., 1984; Dansereau & Yammarino, 2000; Klein et al., 1994; Rousseau, 1985). For us, multi-level or meso formulations (theories, propositions, and hypotheses) are explanations linking variables, which operate at different levels of analysis (e.g., person-level X is positively related to group-level Y) (see Behling, 1978). Rousseau states that such theories specify ‘‘relationships between independent and dependent variables at different levels’’ (1985, p. 20). (Rousseau calls these cross-level – not multi-level or meso – formulations.) Models of this type provide among-level explanations because they link variables in terms of multiple levels of analysis. Included here are mixed-effects models, in which a single variable of interest may have effects at multiple levels with multiple criteria of interest, as well as mixed-determinants models, in which multiple predictor variables at various levels of analysis affect a single criterion at a single level of analysis. For us, cross-level formulations (theories, propositions, and hypotheses) are statements about relationships among variables that are likely to hold equally well at a number of levels of analysis (e.g., X and Y are positively related for individuals and for groups) (see Behling, 1978; Dansereau et al., 1984; Miller, 1978). Rousseau notes that such cross-level formulations ‘‘specify patterns of relationships replicated across levels of analysis’’ (1985, p. 22). (Rousseau, however, calls these multi-level – not cross-level – formulations.) Models of this type are uniquely powerful and parsimonious (simple) because the same effect is manifested at more than one level of analysis (e.g., E ¼ mc2, which holds at multiple levels of analysis). A New Kind of OB 27

Assuming only one level of analysis in a study, or choosing only one level without consideration of other levels, can either mask effects or indicate an effect when none truly exists (Lerner, 1963; Miller, 1978; Pedhazur, 1982; Roberts, Hulin, & Rousseau, 1978). These issues are especially important when individuals are embedded within larger units such as dyads, groups, and collectives in organizations. Thus considering only one level of analysis is insufficient. Instead, multiple levels should be identified in combination. Regarding the particular formulations in Table 3, relationships among constructs may be hypothesized to hold at a lower (e.g., person) level but not at a higher (e.g., group) level. These relationships are discussed as a discontinuity thesis (Miller, 1978), as level-specific formulations (Dansereau et al., 1984; Miller, 1978), or empirically as disaggregated, individual,or level-specific effects (Pedhazur, 1982; Robinson, 1950). In these cases, the higher level of analysis is not relevant for understanding the theoretical constructs. In contrast, relationships among constructs may not be asserted at a lower level but may be hypothesized to manifest themselves at a higher level of analysis. These relationships also are discussed as a type of discontinuity thesis (Miller, 1978), as emergent formulations that hold at a higher (e.g., group) level after not being asserted or found to hold at a lower (e.g., person) level (Dansereau et al., 1984; Miller, 1978), empirically as a higher-level effects that do not disaggregate, or as emergent effects (Miller, 1978; Robinson, 1950). In these cases, the lower level of analysis is not relevant for understanding the theoretical constructs. Thus, in the case of level-specific and emergent formulations, even though a single level of analysis is of primary concern, other levels are considered but defined as not relevant. Alternatively, relationships among constructs may be hypothesized to hold at higher (e.g., collective) and lower (e.g., group) levels of analysis. These relationships are discussed as a homology thesis (Miller, 1978) or empirically as aggregated or ecological effects (Glick & Roberts, 1984; Pedhazur, 1982; Robinson, 1950). As noted previously, they are of two types:

Cross-level explanations (Behling, 1978; Miller, 1978; Dansereau et al., 1984) specify relationships among theoretical constructs that hold equally well, or are replicated across, higher and lower levels. Multi-level or meso explanations (Behling, 1978) assert relationships among theoretical constructs (independent and dependent variables) that operate at different levels of analysis (e.g., individual-level X is related to group-level Y). 28 FRANCIS J. YAMMARINO AND FRED DANSEREAU

We focus on multiple levels of analysis simultaneously and in conjunction with wholes and parts views of levels. Using such specifications for identifying multi-level views, eight cases can be developed. First, wholes at a lower level can aggregate or manifest themselves as wholes at a higher level. This cross-level wholes formulation means that members are homogeneous with respect to the constructs of interest in all entities (e.g., groups and collectives) at both levels of analysis, but the entities (e.g., groups and collectives) differ from one another. Second, wholes at a lower level can aggregate or manifest themselves as parts at a higher level. This cross-level parts formulation means that members are homogeneous with respect to the constructs of interest in all the lower-level entities (e.g., groups), and that these differ from one another. In contrast, in all higher-level entities (e.g., collectives), there is heterogeneity as members within the entities differ from one another. Third, wholes at a lower level may not aggregate or manifest themselves at a higher level (equivocal). This level-specific wholes formulation means that members are homogeneous with respect to the constructs of interest in all lower-level entities (e.g., groups), but that higher-level entities (e.g., collectives) are not relevant. Fourth, parts at a lower level may not aggregate or manifest themselves at a higher level (inexplicable). This level-specific parts formulation means that members are heterogeneous with respect to the constructs of interest in all lower-level entities (e.g., groups), but that higher-level entities (e.g., collectives) are not relevant. Fifth, for an emergent wholes formulation, constructs are expected to hold at a higher (e.g., group) level where members are homogeneous with respect to the constructs after not having been expected or observed at a lower level (equivocal). Sixth, for an emergent parts formulation, constructs are expected to hold at a higher (e.g., group) level where members are heterogeneous with respect to the constructs after not having been expected or observed at a lower level (equivocal). Finally, multi-level formulations can be specified as equivocal (case 7) or inexplicable (case 8) where both higher and lower levels are asserted to be equivocal or inexplicable, respectively, and not relevant for the constructs of interest. In these two cases, additional (other) levels of analysis must be considered for understanding the phenomena of interest. These eight formulations are presented in Table 3 for sake of comple- teness, but only some are relevant here to specify our four simple theories. A New Kind of OB 29

Having outlined an approach for understanding entities in terms of single and multiple levels of analysis, it is now possible to consider variables and constructs and their relationships in conjunction with the entities to specify four simple, yet comprehensive theories for a new kind of OB. As noted earlier, each theory consists of two variables, an association between them, and a speciﬁcation of the levels of analysis at which they are expected to operate (i.e., entities for which they are asserted to hold). These ‘‘little ideas that can be well tested’’ are simple ‘‘rules’’ [in Wolfram’s (2002) terms] that can account for or explain a variety of behaviors and actions in OB.

A THEORY OF INDIVIDUAL BEHAVIOR AND DECISION MAKING

Theory 1. Option cutting and commitment are positively related based on between-person differences (interindividual differences or whole persons); a level-speciﬁc theory at the person level of analysis.

General Formulation

Individuals form psychological attachments to all kinds of entities – family, friends, supervisors, coworkers, and organizations, among others. Such a psychological binding or attachment is typically called commitment (see Becker, 1960; Mathieu & Zajac, 1990; Meyer, Stanley, Herscovitch, & Topolnytsky, 2002; Meyer & Allen, 1984, 1991, 1997; Meyer, Becker, & Vandenberghe, 2004; Meyer & Herscovitch, 2001; Naughton, 1982; Price & Mueller, 1986; Staw, 1976). In OB, commitment to an organization, a superior, and a role are particularly critical, and the basis of this psychological attachment can be cognitive, emotional, or both. How such psychological attachments arise, whether and how they develop over time, and whether certain individuals are more prone to make or form commitments have been the subject of much theorizing, empirical research, and speculation. A traditional view, first espoused by March and Simon (1958), is that organizations or other entities provide ‘‘inducements’’ to individuals to be committed. Thus, the more investments or ‘‘side bets’’ individuals make in various entities, the more likely they are to become committed to them (Becker, 1960). In Becker’s (1960) side-bet theory, as an employee accumulates more significant investments in an organization, the 30 FRANCIS J. YAMMARINO AND FRED DANSEREAU more important it becomes to maintain a relationship with that organization; as the number of side bets increases, so does the commitment. Staw (1976) describes this escalation of commitment to a chosen course of action. The psychological mechanism seems to be powerful enough that individuals continue to commit to various entities or courses of action even when facing continual negative consequences (Staw, 1976). For Meyer and Allen (1984, 1991), continuance commitment occurs when the cost of leaving (an organization) is perceived as prohibitively high, causing the employee to remain at the organization. In brief, individuals, when faced with various alternative courses of action – options or behaviors – appear to become more committed to certain entities as options decrease (i.e., continuance commitment; Meyer & Allen, 1984), as side bets increase (i.e., other options are cut off or viewed as not feasible; Becker, 1960), and as their calculations regarding such costs, investments over time, and lack of alternatives become more salient (‘‘calculative commitment’’; Hrebiniak & Alutto, 1972; Mathieu & Zajac, 1990). As such, for individuals, things persist over time and involve a rejection of feasible alternatives that, presumably, are less desirable. We refer to this process simply as option cutting (also see Naughton, 1982). An interesting issue here is (1) whether these individual differences are a reflection of relatively stable individual traits or characteristics such that we can observe interindividual differences where commitment and option cutting vary from person to person or (2) whether these individual differences shift or develop over time such that we can observe intraindividual differences where commitment and option cutting change within persons. The former view is a whole-person perspective; the latter is a person-parts perspective. Both are plausible levels of analysis alternatives for defining and explaining the option cutting–commitment connection (see Naughton, 1982). Given these points, it seems useful to elaborate some notions about commitment (especially to organization, roles, and superiors), option cutting (and its various forms or types), and person level of analysis issues to develop a testable theory of these ideas.

Speciﬁc Formulation

In terms of commitment, psychological attachments to organizations, roles (or jobs), and one’s superior/supervisor have all been the subject of prior work (e.g., Mathieu & Zajac, 1990; Meyer et al., 2002, 2004). Organizations in general, and work roles (or occupations) and supervisors in particular, are A New Kind of OB 31 an important part of our lives. Oftentimes, beyond spending a large portion of their working hours in organizations and in work roles as well as interacting with superiors, individuals identify with their organizations and jobs and endorse or adopt the values espoused by their superiors. As such, our interest in commitment per se is more specifically focused on an individual’s psychological attachment to his or her (1) organization, (2) role, and (3) superior. In terms of option cutting, several different mechanisms or types may be involved (see Naughton, 1982). First, individuals can display an option-cutting orientation. In other words, perhaps owing to personality traits or characteristics, some individuals may have a greater propensity to seek out and evaluate information on options relative to one another. They might then be more inclined to cut off options as compared to other individuals. Second, this process may be behavioral based and evidenced in terms of option-cutting actions. Some individuals may cut options by engaging in behaviors or actions that preclude or prevent them from engaging in other forms of behavior. Physical activities, emotional energy, and psychological efforts are all limited in some way, so engaging in some activities means not partaking in other ones. Finally, environmental option-cutting is setting- or context-based. In this case, because of various situational constraints or contingencies, individuals may find themselves in environments or contexts where they do not know about options or alternatives or cannot gather information to evaluate them. Their location in such situations may result either from their own choices or from factors beyond their control. When considered together, it seems plausible that option-cutting orientation, option-cutting actions, and environmental option cutting are associated with commitment to an organization, a work role, and a superior based on individual differences (i.e., the person level of analysis). A variety of OB and related literatures appear to support this assertion at the person level of analysis as well as in terms of both stable trait differences (between- persons view; i.e., wholes) and changes in individuals over time (within- persons view; i.e., parts). At what is often recognized as the beginning of OB, March and Simon (1958) extended the classical economics view and discussed the notions of ‘‘bounded rationality’’ and ‘‘satisficing.’’ Individuals are not able to pursue multiple courses of action simultaneously because of physical (space–time) and biological restrictions. Moreover, they often do not know or cannot even conceive of the consequences of their actions because information is not 100% complete or accurate, nor is unlimited information processing 32 FRANCIS J. YAMMARINO AND FRED DANSEREAU time available to assess costs and benefits of alternatives. As such, individuals make rational decisions within limits (bounded rationality) that are satisfactory rather than optimal in nature (satisficing). Individuals are, therefore, viewed as restricting their alternative courses of action or cutting off options that are unknown, viewed as not feasible, perceived as not available, and so on. The more options that are cut or eliminated, the more individuals become committed to the options that remain. Choices or investments that individuals make in various courses of action can explain their attitudes such as commitment to an organization, job, or superior. Becker’s (1960) discussion of side bets aligns with this view. As a function of time, the cost of leaving a system (e.g., organization) increases because an individual has made more and more bets or developed extraneous interests in the system, and as such remains committed to staying. This ‘‘consistent behavior’’ – that is, consistent lines of activity – implies a rejection by the individual of feasible alternatives, resulting in a ‘‘commitment by default’’ (see Becker, 1960). An underlying issue for side bets is that by placing these bets, an individual gives up other alternative courses of actions or options. In a related development, Staw (1976) noted that even in light of negative consequences, individuals tend to hold to a decision and continue investing in a failing course of action, rather than reversing the decision. Individuals appear to continue to commit to these decisions because their behavioral actions prevent them from entertaining or engaging in other options. Individuals cannot optimize their decision, so they use satisficing criteria to justify remaining with the current option and cut off other options (also see March & Simon, 1958). This strategy then yields an escalation of commitment to the current course of action (see Staw, 1976). The preceding discussion suggests two different possibilities to account for how individuals cut options and become committed (also see Naughton, 1982). First, at the person level of analysis, the option cutting–commitment relationship may be based on stable traits or enduring characteristics of individuals. In this case of interindividual differences, option cutting and commitment are properties of individuals and vary from person to person (between-person differences or a wholes perspective). For example, workaholics, by their very nature, may be more committed and cut off more options than non-workaholics. Even so, if option cutting and commitment are viewed as individuals’ stable traits, then organizations might make selection decisions based on these constructs and their specific dimensions. Second, and also at the person level of analysis, the option cutting– commitment relationship may evolve over time as shifts or changes occur A New Kind of OB 33 concomitantly for the two main constructs and their specific dimensions. In this case of intraindividual differences, option cutting and commitment constitute an internal process that varies within each person (within-person differences or a parts perspective). For example, a state of cognitive dissonance or tension might motivate a person to change his or her option cutting, and therefore commitment, so as to restore harmony or balance in the individual’s life. If option cutting and commitment are viewed as changeable internal processes of individuals, then organizations might make training and development decisions based on them. A third possibility is that a shift in levels of analysis – rather than a shift in variables – occurs over time [see Dansereau et al. (1999) for a detailed explanation of this phenomenon] for the option cutting and commitment relationship. In this case, a parts (or wholes) view of option cutting and commitment at a focal time period might ‘‘transform’’ up (or down) to a wholes (or parts) view at a subsequent time period. While this process is plausible, as are the other two previously mentioned views at the person level of analysis, in the spirit of simplicity endorsed here, as well as based on the more traditional view in the literature (see Becker, 1960; Mathieu & Zajac, 1990; Meyer et al., 2002, 2004), we offer below what we believe is the more likely alternative as an initial assertion for empirical testing in future work. In summary, these notions result in the following specific testable formulation:

Option-cutting orientation, option-cutting actions, and environmental option cutting are positively related to commitment to an organization, commitment to a role, and commitment to a superior based on between-employee differences (interindividual differences or whole employees).

Etiology of the Effect Why do we expect this person-level hypothesized effect to hold? Or, stated differently, from an evolutionary theory perspective (Gould, 2002), does the effect make a difference in terms of the human condition and survival? We believe that this effect may matter because without it, or something like this effect, it would likely be impossible for individuals to make decisions and stay with or attach to them. The hypothesis means that an individual can commit to a course of action and pursue it. Without this capability, or something similar to it, when faced with a threat, there would be too many (virtually unlimited) options to pursue, and the individual would likely experience analysis paralysis. In such a situation, if a predator was starting 34 FRANCIS J. YAMMARINO AND FRED DANSEREAU to attack, this paralysis would have resulted in extinction. Thus this effect, or something like it, may be quite important for the survival of humans. Those individuals with capabilities to cut options and become committed may have been likely to pass that characteristic on in an evolutionary sense. As an additional issue, if a person is committed, how can that commitment change? For a change to occur requires that a new set of behaviors or choices arise, such that then the commitment changes. But these behaviors and choices will be resistant to change until the person perceives that the current behavior or option does not satisfy him or her any longer, and that it no longer warrants commitment. In other words, the person would become dissatisﬁed with the current choice and no longer remain committed to it. Barring such a situation, the original option remains the viable one that the person follows.

A THEORY OF INTERPERSONAL RELATIONS AND LEADERSHIP

Theory 2. Investments and returns are positively related based on between- dyad differences (whole dyads); a level-speciﬁc theory that emerges at the dyad level of analysis.

General Formulation

Our theory of interpersonal relationships and leadership is a formulation derived from a general theory of social exchange (see Blau, 1964; Emerson, 1976; Homans, 1958, 1961), which links equity, inequity, and law of effect notions (see Dansereau et al., 1984). As a general theory, social exchange theory can be viewed as a positive relationship between investments and returns. Investments are what one party gives to another party; returns are what one party receives from another party. The ideas of giving and receiving are linked via the principle of reciprocal reinforcement: investments trigger returns and returns trigger investments. A key point of these ideas is that two interdependent parties are involved. Moreover, the dyadic partners are in a balanced relationship where the amount of giving and receiving for each party is similar – that is, whole dyads. Some dyads are rich, with a lot of giving and receiving by both partners; other dyads are poor, with little giving and receiving by both partners. If such dyadic relationships are not balanced, they tend to be short-lived. A New Kind of OB 35

A key assumption of exchange theory is that people are rational, and perhaps hedonistic, seeking pleasure and avoiding pain – or, more realistically, seeking rewards and avoiding punishments. This assumption underlies the law of effect and permits mutual reinforcement – the norm of reciprocity – based on common needs and interests, to hold for exchange formulations (see Gouldner, 1960). This traditional theoretical work from psychology, sociology, and anthropology underlies the giving and receiving (investments and returns) between partners upon which this theory is based.

Speciﬁc Formulation

In work organizations, the key partners involved in social exchange relationships of investments and returns are superiors and subordinates. Superiors make investments (e.g., salary, office space) in and receive returns (e.g., performance) from subordinates. Subordinates, in turn, make investments in and receive returns from superiors. These investments and returns occur on a one-to-one basis in each superior–subordinate dyad (e.g., Yammarino & Dubinsky, 1992; Yammarino, 1995). According to the individualized leadership approach (see Dansereau, 1995; Dansereau et al., 1995; Mumford, Dansereau, & Yammarino, 2000), a critical superior investment is providing support for feelings of self-worth to a subordinate. This can be accomplished by a superior providing attention, support, and assurance to a subordinate. Conversely, a key subordinate investment is providing exceptional performance to a superior. This can be accomplished by a subordinate providing performance at or beyond standards, particularly, in terms of quality, and doing a job in line with a superior’s preferences. These same sets of variables then also serve as subordinate returns (receiving support for his or her self-worth from a superior) and superior returns (receiving satisfying performance from a subordinate), respectively. Again, whole superior–subordinate dyads are involved; that is, some relationships are rich while others are poor, and the relationships tend to be balanced. A key element of such dyadic leadership is the notion of support for self- worth that one individual provides another. This concept is defined as (1) supporting another individual’s actions and ideas; (2) assuring the other individual of confidence and his or her integrity, ability, and motivation; and (3) paying attention to the individual’s feelings and needs. The objective of support for self-worth is not to get people working together (traditional leadership genre) or to establish a vision (new leadership genre), but rather 36 FRANCIS J. YAMMARINO AND FRED DANSEREAU to develop the individual’s competencies. Individualized leadership is a dyadic theory that involves empowerment. One can have empowerment without goals or visions and without getting people to work together. As such, the latter concepts are different aspects of leadership than support for self-worth and individualized leadership (see Dansereau & Yammarino, 1998b; Yammarino, Dansereau, & Kennedy, 2001). Of course, not all ‘‘designated leaders’’ become ‘‘real leaders’’ for all subordinates, as the approach is individualized for each unique dyadic relationship. In other words, individuals (superiors) become leaders for (some, but not necessarily all) subordinates by providing a subordinate with a sense of support for the subordinate’s self-worth as an individual. In turn, (some but not necessarily all) subordinates become followers of a leader by perceiving and receiving a sense of self-worth from the leader and performing in ways that satisfy both the leader and themselves. Their performance is valued by the leader, and in the process these subordinates validate their own sense of self-worth. From a superior’s perspective, subordinates are viewed as independent, unique individuals who can be assessed in terms of specific variables. For example, a superior can assess the degree to which a subordinate’s performance satisfies that superior. Based on exchange theory, one would expect that those subordinates viewed by superiors as performing in ways that satisfy the superior would, in turn, receive reinforcement of their sense of self-worth from the superior. Homans (1961) notes that people give social approval (a sense of self-worth) as a generalized reinforcer to others who have given them commodities they value (satisfying performance), which makes it more likely that the other parties will continue giving valued commodities. Homans (1961) also hypothesizes that individuals who can give rare and highly valued services or commodities as rewards can command in return a great deal of esteem from others. In terms of leadership, then, the key variables are (1) the performance of a subordinate that satisfies a superior and (2) support by a superior for a subordinate’s sense of self-worth. Leadership researchers traditionally focus on subordinate performance as perceived by a superior (see Bass, 1990). A focus on performance that satisfies a superior allows for the possibility that a subordinate’s performance may satisfy one superior but not others. As such, a leader is defined, in part, as one who obtains satisfying performance from another person. Therefore, the ability to satisfy a superior reflects differences between individual subordinates, independent of their formal hierarchical arrangement with a superior. A New Kind of OB 37

Support for self-worth is a judgment or perception that a superior supports a subordinate’s individual self-worth in a variety of ways. Our focus is on the degree to which a superior perceives himself or herself as supporting a subordinate’s feelings of self-worth rather than on the self- esteem of a subordinate independent of the superior. A superior’s perception of one assigned subordinate is independent of the same superior’s perception of other subordinates assigned to him or her. From the perspective of subordinates, subordinates perform in a satisfying manner for superiors based on their perceptions of the superiors. Given social exchange theory, when a subordinate perceives that a superior supports his or her sense of self-worth, the subordinate takes more personal responsibility for the support received because this support focuses on the subordinate’s individual differences in terms of elements such as needs, feelings, and integrity. Based on exchange theory, subordinates reciprocate by adjusting the degree to which they satisfy a superior. Moreover, the perceptions of any one subordinate are independent of the perceptions of the other subordinates who are formally assigned to the same superior. Thus a subordinate’s perceptions of self-worth support and satisfying performance are independent of the perceptions of other subordinates who are formally assigned to the same superior. Combining the subordinate and superior perspectives, subordinates are viewed by themselves and by their superiors as unique individuals, regardless of their formal group assignments. This is the essence of an individualized view of superiors and subordinates. Overlaying leadership concerns on this individualized view requires the notion of reciprocal interdependence between superiors and subordinates (see Bass, 1990). Although there may be a lack of interdependence when superiors and subordinates interact initially, interdependence develops over time in more well-established superior–subordinate (dyadic) relationships. For less well-established superior–subordinate interactions, support for self-worth and satisfying performance, as viewed by both superiors and subordinates, are not reciprocally interdependent. When superiors and subordinates ﬁrst enter into a social exchange relationship, their relationship can be characterized as pooled interdependence. Over time, however, reciprocal interdependence will increase and superiors’ and subordinates’ perceptions will become related. This evolution will occur because subordinates begin to reciprocate with satisfying performance when they perceive that superiors provide them with support for their sense of self-worth. Superiors then complete the cycle by actually reinforcing perceptions of self-worth in exchange for satisfying performance. 38 FRANCIS J. YAMMARINO AND FRED DANSEREAU

For more well-established superior–subordinate interactions, support for self-worth and satisfying performance are reciprocally interdependent. This dyadic notion – the dyad level of analysis – means that there is interdependence between superiors and subordinates on the dimensions of support for self-worth and satisfying performance. Variation or disagreement within dyads (i.e., differences between superiors and subordinates) is viewed as ‘‘error’’ and occurs rarely. If it does happen, usually one party terminates the relationship by transferring to another department or leaving the organization. In contrast, variation between dyads occurs more often. Thus support for feelings of self-worth and satisfying performance from a superior’s perspective depend on the specific subordinate with whom a superior interacts. Feelings of support for self-worth on the part of subordinates vary from one subordinate to another (who report to the same superior). This is a very individualized view, in which the individual is the subordinate and the individuality of the subordinate is enhanced by the superior. The leadership component of individualized leadership is evident because a superior and a specific subordinate (leader and follower) are interdependent. This linkage occurs between the follower as a person and a ‘‘part’’ of the leader (superior) that views the subordinate as a unique person. Such a dyadic perspective depends on the leader’s view of the follower and the follower’s view of the leader – that is, a superior–subordinate dyad. The most typical source of the initiation of this individualized leadership process is superior investments as perceived by the subordinate. A key pivotal event – the initial investment – usually initiates a rich interpersonal connection if it is reciprocated. Beyond the general investments and returns noted previously, specific dimensions of investments and returns can include showing respect, cooperativeness, openness, being task-centered, endorsing values, benevolence, acceptance, authenticity, having fun, providing learning opportunities, communication, and allowing self-responsibility and mistakes. Specific behavioral indicators of investments and returns can include head nods (i.e., agreement), direct eye contact (i.e., attention), praise, questions (by one party), and answers (by the other party). In these behavioral ways, the dyadic bonds can be enhanced and reinforced to richer levels. (For an extended discussion of these notions, see Dansereau et al., 1984, 1995; Mumford et al., 2000; Wallis, 1999; Yammarino & Dansereau, 2002b). In summary, these notions result in the following specific testable formulation:

Superior investments (support for self-worth) and returns (satisfying performance) are positively related to subordinate investments (satisfying performance) and returns A New Kind of OB 39

(support for self-worth) based on between-superior–subordinate dyad differences (whole superior–subordinate dyads).

Etiology of the Effect

Why do we expect this dyad-level hypothesized effect to hold? Stated differently, from an evolutionary theory perspective (Gould, 2002), does the effect make a difference in terms of the human condition and survival? We believe that this effect may matter because without it, or an effect like this one, it seems unlikely that individuals would have seen benefits from interacting with others (making investments), even if only for a very short time. For example, those individuals who are likely to see returns (e.g., of a sexual nature) from interacting with others are those who are more likely to have children. Thus the capability implicit in the effect would be passed on in an evolutionary sense. Moreover, without this effect, it would be difficult for a child to survive because a child cannot fend for himself or herself. Thus, if providing for the child did not provide some benefit to the parent or other caregiver, the child would be unable to survive. An additional issue regarding this effect relates to change. The basic idea is that if a focal individual’s investments in another individual do not relate to that focal individual’s returns from the other individual, the dyadic relationship may collapse because it will be perceived as unfair. This effect can be viewed as underlying equity theory. For example, various changes may be made by members of the dyad to equalize or balance the investments and returns. In addition, the manner in which people learn what is equitable is expected, in part, to be a function of the previous reinforcement contingencies.

A THEORY OF GROUP DYNAMICS/TEAM PROCESSES AND NORMS

Theory 3. Interdependence and cohesion are positively related based on between-group differences (whole groups); a level-speciﬁc theory that emerges at the group level of analysis.

General Formulation

Norms are standards shared by group/team members and, when highly agreed upon, permit the group/team to regulate individual behavior. 40 FRANCIS J. YAMMARINO AND FRED DANSEREAU

(Aligned with our theme of ‘‘simplicity,’’ we use the terms ‘‘group’’ and ‘‘team’’ interchangeably.) Group/team norms are the shared beliefs about members’ expected social behaviors. These norms can guide behavior, exert social control, and inﬂuence performance, either positively or negatively, at both individual and team/group levels of analysis. Group/team norms emerge formally or informally, are transmitted, and then persist through team members’ actions and beliefs. They lead to greater conformity in behavior, beliefs, and attitudes among group/team members, tend to be established early in the group/team formation and development cycle, and are based on interpersonal agreement that arises from members’ interactions, often resulting in cooperation. (For an extended discussion of these issues, see Blau, 1964; Gouldner, 1960; Homans, 1958, 1961; Katz & Kahn, 1978; Kelley & Thibault, 1978; Thibault & Kelley, 1959.) From a social exchange theory perspective (Blau, 1964; Homans, 1958, 1961; Katz & Kahn, 1978), norms create obligations that form from relationships among team members and from each member with the whole team. As a result, team members decrease the variability in their individual behaviors, and their behaviors converge on the agreed-upon norm. The norms then enhance group/team functioning as they facilitate team survival (keeping the team together and protecting it), provide regularity and predictability to team members’ expected behaviors, help the team avoid interpersonal problems, and express the team’s central values and identity (see Taggar & Ellis, 2007; Feldman, 1984). In short, group/team norms facilitate the alignment of individual needs, goals, and expected outcomes across the team so that it operates as a homogeneous entity (i.e., whole team or group). In an interdisciplinary review of social exchange theory, Cropanzano and Mitchell (2005) note that theorists from multiple disciplines agree on the following points: Social exchange involves a series of interactions that generate obligations. These interactions, or transactions, are governed by rules or norms of exchange to which the parties must abide. These interactions are viewed as interdependent and contingent on the actions or behaviors of the parties involved. The resources or ‘‘commodities of exchange’’ may be economic or socioemotional in nature and involve short- or long-term time horizons. Importantly, the interdependent transactions have the potential to generate high-quality relationships involving trust, loyalty, and mutual commitment that evolve over time. A New Kind of OB 41

Collectively, these points suggest the importance of interdependence and cohesion, and their association, which can build to ‘‘cohesive relationships’’ and cooperation over time in groups and teams. Following from the earlier social exchange discussion, interdependence comprises the interconnections among parties that result from their contingent transactions. Cohesion is the motivation to remain in the group or team and the mutual commitment to the other members of that group or team (see Shaw, 1976). In particular, cohesion seems to occur based on the key exchange rule of reciprocity (repayment in kind), which depends on interdependence or outcomes that are based on the parties’ combined efforts (see Cropanzano & Mitchell, 2005; Gouldner, 1960; Homans, 1958, 1961; Blau, 1964). As such, a bidirectional transaction (e.g., investments and returns at the dyad level), involving mutual arrangements, is a basis for reciprocal exchange. When such contingent interpersonal transactions occur across various individuals and dyads within a group or team, the potential arises for (whole) group- or team-level interdependence and cohesion.

Speciﬁc Formulation

Team or group dynamics can focus on the entire team or unit and influence how they operate collectively. In this approach – and especially in high- performing teams – there is a key reliance on shared mental models and shared knowledge and cognition within the team or unit (Baba, Gluesing, Ratner, & Wagner, 2004; Cannon-Bowers & Salas, 2001; Gibson, 2001; Klimoski & Mohammed, 1994; Kozlowski & Ilgen, 2006; Mohammed & Dumville, 2001). No one member of a team stands out always and everywhere; rather, roles and responsibilities are shared and distributed throughout the team depending on the issue, circumstance, expertise needed, time constraints, and other factors. In brief, through team processes of communication, face-to-face interaction, collaboration, and cooperation, there is considerable knowledge acquisition and sharing in teams. Knowledge and information sharing lead to cognitive elaboration, wherein new knowledge structures are created and old structures are modified. Over time, cognitive convergence occurs such that team members gradually acquire enhanced overlap among their cognitive structures. Shared cognitive structures and knowledge, or shared mental models, then can reduce variance in team performance, enhance cohesiveness, build a positive team climate, and promote successful goal accomplishment. These shared mental models, or shared cognitions, are 42 FRANCIS J. YAMMARINO AND FRED DANSEREAU similar, overlapping, compatible, or complementary knowledge or belief structures that represent features of the context such as task-specific knowledge, task-related knowledge, knowledge of team members, and attitudes and beliefs (Klimoski & Mohammed, 1994; Kozlowski & Ilgen, 2006; Cannon-Bowers & Salas, 2001). As noted by these authors and others (e.g., Mohammed & Dumville, 2001), shared and overlapping knowledge and belief structures support the alignment of expectations that permit rapid and smooth coordination of behavior (via shared mental models), lead to greater availability of and access to a larger pool of information for problem solving and task completion, enhance the creation of new knowledge, enable team innovation and learning, and foster faster agreement on problem definition and strategic decisions. In turn, all of these elements result in enhanced team effectiveness and efficiency. As a process-based view, this approach involves ‘‘teamwork’’ with high levels of expended collective effort and high-quality interpersonal relationships. Cohesion, communication, cooperation, and conflict management are both drivers of and outcomes from this process. In this approach, each team member is viewed as being capable of influencing and being influenced by every other member and the team as a whole. Thus team members have a collective influence on one another, which permits members to identify with and be motivated by the team. As such, the focus is on the team level, including how team members evaluate and perceive the influence of the team and the collective responsibility of the team. In this case, there is a shift in the level of analysis (or referent) from the individual to the entire team as a whole. According to this perspective, highly effective teams are characterized as having a clear focus or vision regarding when members are willing to sacrifice individual goals and accomplishments for the team fulfillment of the mission. Team members identify so closely with the team purpose and mission that they are willing to make individual sacrifices for the team and to enhance other team members’ potential and capabilities. This approach results in the building of team/shared mental models, as mentioned previously. Through the evolution of a collective belief structure that is shared among members, teams develop a shared understanding of their behaviors, values, and ideas; develop a sense of cohesiveness; and set expectations of one another that facilitate learning and the ability of the team to lead itself collectively. These notions can be summarized simply by considering the ideas of interdependence and cohesion, and their association, at the whole group/ team level of analysis. Here, interdependence may be viewed as a form of ‘‘positive functioning’’ in which members make contributions and receive A New Kind of OB 43 rewards based on cooperation and reciprocity. This notion stands in contrast to dependence, which can have dysfunctional consequences as a result of over-reliance on conformity and sanctions. Interdependence is the degree to which the ‘‘contingent transactions’’ create interconnections or relationships among parties (group/team members): The more transactions, the more members rely on or are bonded to one another, and the more interdependence. Wageman (1995) notes that both task interdependence (from inputs and process; e.g., group/team task or technology requirements), also called structural interdependence, and outcome interdependence (from goals and rewards; e.g., social demands to work together to accomplish joint outcomes), also called psychological interdependence, to some extent determine how individuals experience or perceive interdependence in groups and teams. In particular, positive functioning interdependence could include the following elements: Group potency: the group’s generalized belief about its ability or capacity to perform (Gibson, 1999) Group efficacy: the group’s belief about its capability to perform a specific task (Gibson, Randel, & Earley, 2000) Group effort: the amount of effort that group members are willing to exert to complete their tasks (Price & Mueller, 1986) Cohesion is the degree to which members of groups/teams are motivated to remain in the same group/team (Shaw, 1976). As such, cohesion is an emergent notion that becomes relevant at the group/team level, is typically psychological and affective in nature, and indicates the degree of shared commitment and attraction among group/team members resulting from their transactions. Highly cohesive teams tend to have less absenteeism; high involvement in team activities; high levels of member coordination and cooperation during tasks; high satisfaction, productivity, and member interactions; a shared vision; and overall higher performance. In particular, cohesion is included in Cohen and Bailey’s (1997) discussion of group psychosocial traits as shared understandings, beliefs, or emotional tones. These also include shared (team) mental models, and (group) affect, and are referred to as ‘‘real group-level phenomena’’ by Cohen and Bailey (1997). In summary, these notions result in the following specific testable formulation:

Task interdependence (from inputs and process) and outcome interdependence (from goals and rewards – potency, efﬁcacy, effort) are positively related to shared mental models and shared understandings, beliefs, and emotional tone (affect) based on between-workgroup or team differences (whole workgroups or teams). 44 FRANCIS J. YAMMARINO AND FRED DANSEREAU

Etiology of the Effect

Why do we expect this group-level hypothesized effect to hold? Stated differently, from an evolutionary theory perspective (Gould, 2002), does the effect make a difference in terms of the human condition and survival? We believe that this effect may matter because without it, or one similar to this effect, it seems unlikely that people would have cooperated with or bonded to one another. Speciﬁcally, when faced with the need to hunt or protect themselves, individuals are almost by deﬁnition interdependent with each other because they would likely be unable to eat or survive without one another. As a result, cooperation and cohesion serve as ways for human beings to compete successfully relative to other species. As a result, and in a somewhat interesting way, cooperation provides a way to compete with other species. Because those individuals who cooperate will likely survive, it is their capability for group-level cooperation and cohesion in response to interdependence that would be passed on in an evolutionary sense. An additional issue regarding this effect is: How does change occur? Here, the point is that as the degree of interdependence faced by individuals increases, cooperation and cohesion increase among them. In other words, changes in the degree of cooperation and cohesion depend on changes in the degree of interdependence faced by individuals.

A THEORY OF COLLECTIVIZED PROCESSES AND ROLES

Theory 4. Titles and expectations are positively related based on between- collectivity differences (whole collectives); a cross-level theory from person to dyad to group to collective levels of analysis.

General Formulation

Several collectivized processes in organizations operate at levels of analysis above the group or team level. One way to capture many of these processes is in terms of the notion of roles (see Alutto, 1968; Biddle, 1979; Dansereau et al., 1984; Katz & Kahn, 1978; Yammarino & Naughton, 1988). To formulate a testable theory of roles per se (rather than role outcomes) that accounts for many organizational processes, it is important to focus on A New Kind of OB 45 positions (or identities) and expectations, which provide information about forms, condition of occurrence, causes, and effects of roles (see Biddle, 1979). The notions of position and expectations (e.g., rights and duties) are common themes included in most definitions of roles (see Alutto, 1968; Biddle, 1979; Dansereau et al., 1984). In general, only expectations that are shared by a number of individuals are considered to be relevant to roles. Likewise, the notion of position implies an individual’s rank or status, which is presumably shared by individuals in similar positions. Moreover, these aspects of roles have been described as applicable to multiple levels of analysis. For example, the concept of ‘‘role’’ has been characterized as an interdisciplinary concept of interest in psychology, social psychology, sociology, and anthropology (see Alutto, 1968; Biddle, 1979; Katz & Kahn, 1978), with individuals in each discipline focusing on different levels of analysis. Following this tradition, Katz and Kahn (1978) have suggested that sociological concepts can be translated to lower levels of analysis. This interdisciplinary focus and associated assumptions of multi-level effects leads us to conclude that any formulation of roles should hold across levels rather than being level-specific in nature. At the person level of analysis, individuals are frequently viewed as actors playing a role, which suggests that individual personalities do not determine the nature of a role. Instead, factors external to an individual (at higher levels of analysis) specify how an individual is to play a role. For example, Blau (1974) notes a fundamental difference between the social processes external to individual personalities that are responsible for playing a role and the internal psychological processes that are not responsible for playing a role. Likewise, at the dyad and group levels of analysis, Katz and Kahn (1978) have proposed a role-taking model. Their model implies that the relationship between two individuals is rather impersonal, in that each actor behaves or is socialized to behave in a way compatible with a role rather than with the personal preferences of the individuals taken either separately or together. In contrast, Miller (1978) has asserted that roles and social systems are composed of human beings and should not be viewed as independent of them. For us, roles do not exist independent of people. If human beings did not exist, roles as we know them could not exist. In other words, roles can be viewed as human contrivances. However, this assertion in no way suggests that one individual or human being will find it very easy to change roles. Indeed, from the perspective of a role-player, roles may appear to be nonmalleable – a perception that may arise from the entities that underlie roles. Nevertheless, this focus on roles as embedded in the social fabric of human existence provides the basis for a theoretical formulation of roles. 46 FRANCIS J. YAMMARINO AND FRED DANSEREAU

At the collectivity level of analysis, individuals are expected to view themselves as ‘‘small’’ compared to the collectivity and as dependent on a large number of indirect connections with other actors. Collectivities are based on hierarchical or other types of impersonal and indirect linkages rather than on personal, face-to-face interactions. In a collectivity, an individual should, in a sense, be able to view himself or herself as simply a cog in a large gear. Moreover, the greater number of connections with other individuals at the collectivity level of analysis (as opposed to the group, dyad, and person levels of analysis) may enhance the feeling that roles are not easily changed. With whole collectivities, the connections between individuals can be viewed as impersonal and not based on group or team processes inside a collectivity. Given these assumptions, variables associated with whole collectivities should be homogeneous within collectivities and different between collectivities. Next, we consider two variables – titles and expectations – and their relationship as key for understanding collectivized processes. Titles constitute a labeling process residing within human beings who occupy physical space. (In contrast, the term ‘‘position’’ can be taken to mean a location in some type of social and perhaps nonphysical space.) We define titles as categories or labels that are properties of collectivities and that reside in human beings. This definition is compatible with prior work on roles (Alutto, 1968; Biddle, 1979). Expectations have typically been defined as having an anticipatory and normative quality and as referring to activities or behaviors (Katz & Kahn, 1978). We define expectations as beliefs about what should or will be done by oneself or by others. Our key assertion – that titles are related to expectations – can be examined in a number of ways. For example, a change in a title may be followed in time by a change in expectations. Alternatively, a change in expectations may result in a change in title over time. Another part of this assertion is that titles and expectations are properties of collectivities. Therefore, collectivities are predicted to differ in terms of their titles and expectations, and these differences are hypothesized to be related. Suppose human beings could associate titles and expectations only at a person level and not at a collectivity level of analysis. In this scenario, each individual would construct titles and expectations independent of other human beings. A similar formulation would result if groups or dyads based on face-to-face interactions were viewed as forming titles and expectations. Therefore, as a cross-level formulation, we suggest that titles and expectations are based on large-scale social movements (see Weber, 1924/1947); persons, dyads, and groups remain important but are viewed differently. A New Kind of OB 47

Specifically, in this formulation, face-to-face interactions in groups or dyads are viewed as reflecting the titles and expectations associated with the collectivities within which these other entities are embedded. Likewise, differences between persons are viewed as simply reflecting the different collectivities within which different persons are embedded. When an individual takes a new job and title, for example, his or her expectations are hypothesized to be drawn in line with the titles and expectations associated with a new collectivity. This cross-level view of persons, dyads, and groups is compatible with previous work (see Alutto, 1968; Biddle, 1979, Katz & Kahn, 1978). From a reductionist perspective, our cross-level formulation means that collectivities are dependent on an alignment of persons, dyads, and groups within each collectivity. Titles and expectations are viewed as somewhat fragile because their maintenance depends on a large number of individuals sharing a similar set of titles and expectations. In contrast, titles and expectations can seem quite powerful when viewed from the perspective of an individual, because a large number of persons, dyads, and groups share similar titles and expectations for themselves and for others in similar roles. If human beings were unable to associate titles and expectations at a collectivity level of analysis, the use of titles (e.g., professor, nurse, doctor, production superintendent) would not trigger any expectations about what an individual with such a title would or should do. Moreover, a collectivity- level formulation implies that a number of individuals recognize such titles and that the titles are not the result of each person’s idiosyncratic ideas or face-to-face interactions with others. Therefore, in formal cross-level terminology, titles and expectations are expected to be associated based on differences between whole collectivities, whole groups, whole dyads, and whole persons.

Speciﬁc Formulation

One way to define titles in a more specific fashion is in terms of their function in society (Weber, 1924/1947; Katz & Kahn, 1978). From this perspective, considering shared (collectivized) individual preferences (values) for producing outputs, the earliest titles may well have been production titles, which focused on the production of goods and services. As people attempted to produce goods and services, the law of entropy (which, in oversimplified form, means that things tend to fall apart) likely influenced organizing processes. This tendency may then have been 48 FRANCIS J. YAMMARINO AND FRED DANSEREAU overcome by activities that involved the maintenance of the production function (maintenance titles). Another method of keeping an organization from decomposing and helping it to survive is to control inputs and outputs. As a result, a new set of titles, support titles, may have evolved. An additional way to overcome entropy is by making changes inside an organization and helping it to evolve or adapt based on adaptive titles. These four functional categories have been suggested by numerous theorists (see Katz & Kahn, 1978) and provide one method for systematically categorizing titles. For example, titles such as production superintendent, foreman, general foreman, and the like are viewed as production titles. Titles such as maintenance engineer, mechanical engineer, and personnel supervisor are viewed as maintenance titles because their function is to provide service to production titles. Production control, quality control, and purchasing can be viewed as support titles because their focus is on controlling inputs and outputs. Titles such as research engineer and chemist can be viewed as adaptive titles because of their focus on change or research and development. In work settings, titles also are linked within a hierarchy in which upper- level titles involve integrating various functions, and lower-level titles focus mainly on the performance of a function. Therefore, eight types of functional titles can be asserted as having theoretical meaning: upper- and lower-level production, maintenance, support, and adaptive titles. Titles can also be categorized in terms of degree of professionalism. The concept of professionalism has been tied to a general (collectivity-level) shift in individuals’ preferences (values) for providing goods and services to society at large. For example, Kornhauser (1962) identified a mutual dependence between organizations and professionals. From this perspective, organizations are viewed as having evolved from small, family-centered economic units, and the criterion for filling jobs became functional knowledge rather than inherited status. In this way, professional associations can be thought of as providing certification or a title, which indicates that an individual has a certain degree of functional knowledge. Because professionalism involves an association or organization of individuals, the collectivity level again seems appropriate for understanding titles. Although conflicts may arise for individuals who are members of both a profession and a work organization, this type of conflict is not inevitable. In work settings, professional titles may complement functional titles. For example, adaptive or research and development titles in a work setting may require greater professionalism on the part of individuals who hold such titles. In contrast, production titles may rely on the specifics of the particular A New Kind of OB 49 situation, rather than on professional titles and expectations. Essentially, collectivities in work settings are hypothesized to have different titles and degrees of professionalism. Different titles may trigger various specific expectations about how much freedom from constraints should be given to an individual. For example, some titles (such as ‘‘production’’) may trigger an expectation of greater attention to rules and regulations, greater interaction with a superior, and greater use of machines, equipment, and technology. Other titles (such as ‘‘adaptive’’) may trigger less attention to rules and regulations, less interaction with superiors, and less reliance on machines, equipment, and technology. Expectations of greater freedom from various constraints might be expected in upper-level adaptive collectivities than in lower-level production collectivities. These specific definitions of expectations are of particular interest because expectations that are shared across levels imply that, within work settings, the behavioral latitude expected of or given to an individual may vary with his or her title. One individual may have greater control over his or her job- related behaviors, whereas another may possess less freedom. Perhaps due to a variety of social and personal values, titles that imply a lack of freedom from constraints are often viewed as objectionable. As a solution to this problem, it is occasionally suggested that collectivities should be enriched to incorporate greater freedom from various constraints, thereby granting increased latitude to all members regardless of title. An individual who holds a title implying a great deal of latitude can experience at least as many problems as one who holds a title signifying little freedom from constraints. The titles and expectations are viewed as determined by collectivity-level processes and not just by individual preferences. Thus differences between whole persons, whole dyads, and whole groups are expected to be cross-level manifestations of membership in different collectivities. In summary, these notions result in the following specific testable formulation:

Functional and professional titles are positively related to expectations about freedom from constraints (such as from machines, equipment, technology, or superiors) at the whole collectivity (department, business unit, or organization), whole workgroup (or team), whole superior–subordinate dyad, and whole employee levels of analysis.

Etiology of the Effect

Why do we expect this collectivity-level hypothesized effect to hold? Stated differently, from an evolutionary theory perspective (Gould, 2002), does the 50 FRANCIS J. YAMMARINO AND FRED DANSEREAU effect make a difference in terms of the human condition and survival? We believe that this effect may matter because without it, or a similar effect, it seems unlikely that language would have developed. Language provides a way to learn from the knowledge of the past about how things occur. Thus the ability to title an event and use this title in the future provides an advantage in the competition of a species for survival. Obviously, without language, each event and approach to it would have to be rediscovered by each generation. With the learning of words (titles) and their meaning (expectations), it becomes possible to use the successes of the past to cope with the present. As such, this capability increases the likelihood of human survival and so would be passed on in an evolutionary sense. An additional issue is: How does change occur in this effect? In this scenario, different titles may be used to label new situations in an attempt to more effectively deal with those situations. Moreover, new situations may occur that are somewhat widespread and may generate new titles and new expectations. Indeed, dictionaries are updated periodically as new words evolve to capture new situations.

IMPLICATIONS

Brief Recapitulation

Our purpose here, following from the work of Wolfram (2002), has been to propose four simple, yet comprehensive theories for a new kind of OB. Each general theory, as summarized in Table 1, asserts a relationship among two constructs and the levels of analysis of operation to explain individual behavior and decision making (Theory 1), interpersonal relationships and leadership (Theory 2), group dynamics/team processes and norms (Theory 3), and collectivized processes and roles (Theory 4) in organizations. We believe these simple notions can account for a variety of both simple and complex behaviors and actions in OB. We have also proposed more specific formulations of these simple ideas by considering specific types of entities (e.g., superior–subordinate dyads rather than dyads in general) and variables (e.g., superior–subordinate investments and returns, including support for self-worth and satisfying performance, rather than investments and returns in general) in work organizations. Table 4 summarizes these specific formulations of our four theories for a new kind of OB. For each theory, the entities/levels of analysis A New Kind of OB 51

Table 4. Speciﬁc Formulations of Four Theories for a New Kind of OB.

Theory Entities/Levels of Constructs/Variables Analysis

Theory 1: individual Whole employees/ Option-cutting orientation behavior and decision managers Option-cutting actions making Environmental option cutting Commitment to organization Commitment to role Commitment to superior Theory 2: interpersonal Whole superior– Superior investments (support for self- relations and subordinate dyads worth) leadership Subordinate investments (satisfying performance) Superior returns (satisfying performance) Subordinate returns (support for self- worth) Theory 3: group Whole workgroups/ Task interdependence (from inputs and dynamics/team teams process) processes and norms Outcome interdependence (from goals and rewards: potency, efﬁcacy, effort) Shared mental models Shared understandings, beliefs, and emotional tone (affect) Theory 4: collectivized Whole departments/ Functional titles processes and roles business units/ Professional titles organizations Freedom from constraints (machines, equipment, technology, superiors)

in work organizations are presented along with the specific variables in work organizations that are manifestations of the general constructs. Given our emphasis here on simplicity, we do not wish to complicate these notions further. Nevertheless, after verification (or disproof) of these ideas, it would be possible to provide theoretical extensions that include multiple variables (e.g., mediators) and multiple relationships (e.g., moderators) involved in the four theories. While beyond the scope of this chapter, as part of the varient approach, both multiple variable analysis (MVA) and multiple relationship analysis (MRA) have been developed to specify theoretically mediators and moderators, respectively, and then test these notions empirically (see Dansereau et al., 1984, 2006; Dansereau & Yammarino, 2000, 2006; Yammarino, 1994, 1998). Also, theoretical specifications and empirical tests of potentially changing variables 52 FRANCIS J. YAMMARINO AND FRED DANSEREAU and shifting levels of analysis over time are plausible (see Dansereau et al., 1999). When considering the four hypothesized effects simultaneously or in combination, the effects can create a variety of frames of reference for viewing organizations and the people in them. Here, for illustrative purposes, we focus on one such frame of reference. Beginning at the individual level of analysis, individuals are viewed as joining an organization: (1) because not working is not an option and (2) by not accepting other jobs or because there were no other jobs (option cutting). These individuals are viewed as, to some degree, committed to the organization, at least initially. Over time, as the individuals interact in dyads, they may receive what they want from others and form relationships that work for them (i.e., provide returns for their investments). At the same time at the group level, they may form cohesive relationships and bond as a way to accomplish various interdependent tasks that require cooperation. All of these processes occur within the context of a structure where labels (titles) provide expectations from the past that guide current activities. Over time, the outcomes from the effects at one level of analysis may actually contradict the outcomes from other levels of analysis. For example, individuals may construct interpersonal agreements at the dyad level of analysis that contradict the titles and expectations of their positions based on the collectivity level of analysis. Consider the corporate scandals at the beginning of this century. Here, despite the (collectivity-level) titles of CEO and Vice President, which carry an expectation of stewardship or watching out for the organization’s welfare and benefit, a number of individuals had interpersonal (dyadic) agreements regarding what would be acceptable to them. These interpersonal agreements contradicted the expectations of their titles. From this multi-level perspective, providing help in and benefit to organizations depends on the ability to understand which outcomes occur at multiple levels of analysis and how to affect them. Clearly, to more fully understand what these outcomes are and how they are affected by the four hypothesized effects will require additional empirical research.

Empirical Testing

Focusing on the four theories as highlighted in Table 4, all the specific formulations are readily testable. In particular, established measures – both quantitative and qualitative in nature – are available for the variables (see A New Kind of OB 53 the previously cited work for each theory). Moreover, it is possible to easily identify all of the specific entities (i.e., employees, managers, superior– subordinate dyads, workgroups, teams, departments, strategic business units, organizations) in the ‘‘real world.’’ As such, our four simple theories can be tested empirically in field studies (as well as laboratory studies if so desired). Because the four theories for a new kind of OB include multiple levels of analysis, a multi-level, data-analytic technique must be used to test empirically for effects and to draw conclusions about them. Within and between analysis (WABA) is such an analytic-inferential technique based on the varient approach employed here (for details on WABA, see Dansereau et al., 1984, 2006; Dansereau & Yammarino, 2000, 2006; Yammarino, 1998, 2003; Yammarino & Dubinsky, 1992; Yammarino & Markham, 1992). WABA is an extension of Robinson’s (1950) seminal work for testing levels of analysis effects when multiple levels (entities) – such as persons embedded within dyads, groups, and collectives – are present. Compatible with the work of Pedhazur (1982), WABA integrates various correlational, ANOVA, and ANCOVA procedures to assess both variation and covariation in variables within and between entities (levels of analysis) and across levels.ThusWABAisideallysuitedfortestingthefourtheoriesproposedhere. Although details about these procedures are beyond the scope of this chapter, an overview of the empirical conditions in WABA and associated inferential conditions is presented in Table 5. For each conceptual view of a single level of analysis (wholes, parts, equivocal, inexplicable), an aligned empirical view, based on between-cells/units and within-cells/units

Table 5. Single-Level Formulations and Inferences.

Alternative Conceptualization Empirical Speciﬁcation Views of Entities WABA I (X and Y WABA II (X–Y covariation) variation)

Between- Within- Between- Within- Between- Within-entities entities entities entities entities entities differences focus focus differences differences differences

Wholes Valid Error Variable Constant Systematic Nonsystematic Parts Error Valid Constant Variable Nonsystematic Systematic Equivocal Valid Valid Variable Variable Systematic Systematic Inexplicable Error Error Constant Constant Nonsystematic Nonsystematic

Source: Adapted from Dansereau et al. (1984, p. 42). 54 FRANCIS J. YAMMARINO AND FRED DANSEREAU differences, is shown in the table. Empirical determinations are made for the variation of each variable separately (called WABA I), the covariation among variables (WABA II), and the consistency of variation and covariation when combined (WABA components) to draw an overall inference. These empirical speciﬁcations, when aligned on a one-to-one basis with the conceptual conditions, permit the drawing of inferences about levels of analysis effects for all cases discussed previously and shown in Tables 2 and 3, and (1) at a single level, (2) for multiple levels, and in terms of (3) multiple variables (e.g., mediators), and (4) multiple relationships (e.g., moderators). (For extensive details about and examples of WABA for all these instances, see Dansereau et al., 1984, 2006; Dansereau & Yammarino, 2000, 2006; Yammarino, 1998, 2003; Yammarino & Dubinsky, 1992; Yammarino & Markham, 1992.)

Additional Considerations

Beyond testing – or more accurately, if empirical tests of the four theories for a new kind of OB are supported and replicated – we might then consider some additional implications of these simple notions. For example, we might investigate the implications for professional and managerial practice. Our four theories suggest a variety of considerations for practice. First, if option cutting and commitment do, in fact, operate in terms of between- person differences and are trait-like and stable over time (i.e., whole persons), then organizations might focus on these issues as part of the hiring and selection process. In this case, it might be easier to select committed or potentially committed individuals for employment rather than try to develop commitment in individuals. In contrast, if option cutting and commitment operate in terms of within-person differences and develop or change over time (i.e., person parts), then organizations might use training and development programs and experiences to enhance employer and manager commitment. Likewise, superior–subordinate investments and returns as well as interdependence and cohesion based on social exchanges in dyads and groups/ teams, respectively, could be used as the basis for assignment and relocation decisions. If leadership and team dynamic processes do not develop as anticipated, then individuals could be reassigned based on potential social exchanges with more ‘‘compatible’’ individuals. Finally, titles and expectations could form the basis of standard operating procedures and policies in various organizational collectives. Clearly, until A New Kind of OB 55 supported through multi-level empirical testing of the four theories, these ideas for professional and managerial practice are pure speculation. Another interesting arena for speculation about and extension to the four theories is future theory building and testing that includes multiple variables and multiple relationships (i.e., MVA and MRA), as noted earlier. Clearly, the previously cited work used for the development of our four theories suggests several directions here. An even more fascinating possibility, and one that is consonant with our theme of ‘‘simplicity,’’ is the possibility of a ‘‘GUT of OB.’’ The search for and development of a grand unified theory (GUT) of OB would be analogous to that endeavor in physics. Physicists are seeking a ‘‘GUT of the four forces’’ (i.e., electromagnetism, gravity, strong nuclear bonds, and weak nuclear bonds) in an attempt to create ‘‘a theory of everything’’ that is a simple unifying description of the world and universe. In OB, we might seek to link and integrate the four simple theories, if empirically supported and replicated, in much the same way to fully and simply account for organizational behavior. Extending this speculation a bit further, it is interesting to note a potential physics–OB ‘‘rules’’ parallel [in Wolfram’s (2002) sense] here for the four forces in physics and our four theories in OB. A potential alignment might be as follows: Electromagnetism with the person level and option cutting and commitment Strong nuclear bonds with the dyad level and investments and returns Weak nuclear bonds with the group level and interdependence and cohesion Gravity with the collective level and titles and expectations While this is an entertaining exercise and analogy for us to ponder, a more useful approach might be to speculate about how a simple ‘‘GUT of OB’’ might begin to be formulated in terms of the variables and entities involved in our four theories. From an entity perspective, a unifying integration seems readily plausible given that, in work organizations, persons are embedded in dyads, groups, and collectives; dyads are embedded in groups and collectives; dyads are a special case of groups (two-person groups); and groups are embedded in collectives. From a variable perspective, cognizant of the entities involved, a GUT of OB also seems possible given the simple theories proposed here. In particular, option cutting and commitment result in individual (person- level) decisions; and one type of decision individuals make is with whom they choose to form dyadic relationships. Thus persons can make choices 56 FRANCIS J. YAMMARINO AND FRED DANSEREAU and commitments about in whom to make investments and from whom to receive returns in dyadic relationships. Moreover, investments and returns, and interdependence and cohesion, as noted earlier, are both social exchange-based formulations, so they have a common theoretical foundation. Also, making investments in and receiving returns from another individual create a type of interdependence that, over time and in multiple dyads, may lead to group cohesion. Likewise, cohesion and interdependence are a way to describe (group) norms; these norms come with and are based on a set of expectations. As such, norms and group functioning may be linked to other expectations and various titles. Again, however, each of our four simple theories would first need to be verified and replicated before venturing into the realm of a GUT of OB.

CONCLUSION

Following from Wolfram (2002), we believe that simple ‘‘rules,’’ or ‘‘little ideas that are well tested,’’ can explain a variety of both simple and complex behavior and actions in organizations. The four simple, yet comprehensive theories proposed here are an attempt to integrate much prior work and form the basis for a new kind of OB.Wedonot believe that (1) option cutting and commitment at the whole person level (for understanding individual behavior and decision making), (2) investments and returns at the whole dyad level (for understanding interpersonal relations and leadership), (3) interdependence and cohesion at the whole group level (for understanding group dynamics/team processes and norms), and (4) titles and expectations at the whole collective level as well as at the whole group, whole dyad, and whole person levels (for understanding collectivized processes and roles) are the explanation for all the behavior in organizations. However, we do believe that they provide one way to account for many behaviors and actions, and that they do so as well as more complex theories in OB. Clearly, multi-level empirical testing and replication of these ideas will demonstrate whether our assertions and beliefs are correct. Whatever the outcome of such investigations, our hope is that we have pushed the science of OB a bit further with Theory 1, Theory 2, Theory 3, and Theory 4 for a new kind of OB.

ACKNOWLEDGMENTS

We thank Joseph Alutto, MacDonald Dumas, Steven Markham, Sidney Nachman, and Thomas Naughton for their contributions to the line of A New Kind of OB 57 research on which this chapter is based. We also thank Marie Iobst for her extraordinary efforts.

REFERENCES

Alutto, J. A. (1968). Role propositions and the analysis of organizations. Unpublished doctoral dissertation, Cornell University. Baba, M. L., Gluesing, J., Ratner, H., & Wagner, K. H. (2004). The contexts of knowing: Natural history of a globally distributed team. Journal of Organizational Behavior, 25(5), 547–587. Bass, B. M. (1990). Bass & Stogdill’s handbook of leadership. New York: Free Press. Becker, H. S. (1960). Notes on the concept of commitment. American Journal of Sociology, 66, 32–40. Behling, O. (1978). Some problems in the philosophy of science of organizations. Academy of Management Review, 3, 193–201. Biddle, B. J. (1979). Role theory: Expectations, identities and behavior. New York: Academic Press. Blau, P. M. (1964). Exchange and power in social life. New York: Wiley. Blau, P. M. (1974). On the nature of organizations. New York: Wiley. Cannon-Bowers, J. A., & Salas, E. (2001). Reflections on shared cognition. Journal of Organizational Behavior, 22, 195–202. Chaitin, G. (2006). The limits of reason. Scientific American, 294(3), 74–81. Cohen, S. G., & Bailey, D. E. (1997). What makes teams work: Group effectiveness research from the shop floor to the executive suite. Journal of Management, 23, 239–290. Cropanzano, R., & Mitchell, M. S. (2005). Social exchange theory: An interdisciplinary review. Journal of Management, 31, 874–900. Dansereau, F. (1995). A dyadic approach to leadership: Creating and nurturing this approach under fire. Leadership Quarterly, 6, 479–490. Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory testing in organizational behavior: The varient approach. Englewood Cliffs, NJ: Prentice Hall. Dansereau, F., Cho, J., & Yammarino, F. J. (2006). Avoiding the ‘‘fallacy of the wrong level’’: A within and between analysis (WABA) approach. Group and Organization Manage- ment, 31, 536–577. Dansereau, F., & Yammarino, F. J. (Eds). (1998a). Leadership: The multiple-level approaches (Part A: Classical and new wave). Vol. 24 of Monographs in Organizational Behavior and Industrial Relations. Stamford, CT: JAI Press. Dansereau, F., & Yammarino, F. J. (Eds). (1998b). Leadership: The multiple-level approaches (Part B: Contemporary and alternative). Vol. 24 of Monographs in Organizational Behavior and Industrial Relations. Stamford, CT: JAI Press. Dansereau, F., & Yammarino, F. J. (2000). Within and between analysis: The varient paradigm as an underlying approach to theory building and testing. In: K. J. Klein & S. W. J. Kozlowski (Eds), Multi-level theory, research, and methods in organizations: Foundations, extensions, and new directions (pp. 425–466). San Francisco, CA: Jossey-Bass. Dansereau, F., & Yammarino, F. J. (Eds). (2003). Multi-level issues in organizational behavior and strategy. Vol. 2 of Research in Multi-Level Issues. Oxford, UK: Elsevier. 58 FRANCIS J. YAMMARINO AND FRED DANSEREAU

Dansereau, F., & Yammarino, F. J. (Eds). (2005). Multi-level issues in strategy and methods. Vol. 4 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Dansereau, F., & Yammarino, F. J. (2006). Is more discussion about levels of analysis really necessary? When is such discussion sufficient? Leadership Quarterly, 17, 537–552. Dansereau, F., & Yammarino, F. J. (Eds). (2007). Multi-level issues in organizations and time. Vol. 6 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Dansereau, F., Yammarino, F. J., & Kohles, J. (1999). Multiple levels of analysis from a longitudinal perspective: Some implications for theory building. Academy of Manage- ment Review, 24, 346–357. Dansereau, F., Yammarino, F. J., Markham, S. E., Alutto, J. A., Newman, J., Dumas, M., et al. (1995). Individualized leadership: A new multiple-level approach. Leadership Quarterly, 6, 413–450. Emerson, R. M. (1976). Social exchange theory. Annual Review of Sociology, 2, 335–362. Feldman, D. C. (1984). The development and enforcement of group norms. Academy of Management Review, 9, 47–53. Firebaugh, G. (1980). Groups as contexts and frog ponds. In: K. H. Roberts & L. Burstein (Eds), New directions for methodology of social and behavioral research: Issues in aggregation (pp. 43–52). San Francisco, CA: Jossey-Bass. Gibson, C. (2001). From knowledge accumulation to accommodation: Cycles of collective cognition in work groups. Journal of Organizational Behavior, 22, 145–160. Gibson, C. B. (1999). Do they do what they believe they can? Group efficacy and group effectiveness across tasks and cultures. Academy of Management Journal, 42(2), 138–152. Gibson, C. B., Randel, A. E., & Earley, P. C. (2000). Understanding group efficacy: An empirical test of multiple assessment methods. Group and Organization Management, 25(1), 67–96. Glick, W. H., & Roberts, K. H. (1984). Hypothesized interdependence, assumed independence. Academy of Management Review, 9, 722–735. Gould, S. J. (2002). The structure of evolutionary theory. Cambridge, MA: Belknap Press of Harvard University Press. Gouldner, A. W. (1960). The norm of reciprocity: A preliminary statement. American Sociological Review, 25, 161–178. Homans, G. C. (1958). Social behavior as exchange. American Journal of Sociology, 63, 597–606. Homans, G. C. (1961). Social behavior: Its elementary forms. New York: Harcourt Brace. House, R., Rousseau, D. M., & Thomas-Hunt, M. (1995). The meso paradigm: A framework for the integration of micro and macro organizational behavior. In: B. M. Staw & L. L. Cummings (Eds), Research in organizational behavior (Vol. 17, pp. 71–114). Oxford, UK: Elsevier. Hrebiniak, L. G., & Alutto, J. A. (1972). Personal and role-related factors in the development of organizational commitment. Administrative Science Quarterly, 17, 555–573. Katz, D., & Kahn, R. L. (1978). The social psychology of organizations. New York: Wiley. Kelley, H. H., & Thibault, J. W. (1978). Interpersonal relationships: A theory of interdependence. New York: Wiley. Klein, K. J., Dansereau, F., & Hall, R. J. (1994). Levels issues in theory development, data collection, and analysis. Academy of Management Review, 19, 195–229. Klimoski, R., & Mohammed, S. (1994). Team mental model: Construct or metaphor? Journal of Management, 20, 403–437. A New Kind of OB 59

Kornhauser, W. (1962). Scientists in industry: Conﬂict and accommodation. Berkeley, CA: University of California Press. Kozlowski, S. W. J., & Ilgen, D. R. (2006). Enhancing the effectiveness of groups and teams. Psychological Science in the Public Interest, 7, 77–124. Lerner, D. (Ed.) (1963). Parts and wholes. New York: Free Parkham. March, J. G., & Simon, H. A. (1958). Organizations. New York: Wiley. Mathieu, J. E., & Zajac, D. M. (1990). A review and meta-analysis of the antecedents, correlates, and consequences of organizational commitment. Psychological Bulletin, 108, 171–194. Meyer, J. P., & Allen, N. J. (1984). Testing the ‘‘side-bet theory’’ of organizational commitment: Some methodological considerations. Journal of Applied Psychology, 69, 372–378. Meyer, J. P., & Allen, N. J. (1991). A three-component conceptualization of organizational commitment. Human Resource Management Review, 1, 61–89. Meyer, J. P., & Allen, N. J. (1997). Commitment in the workplace: Theory, research, and applications. Thousand Oaks, CA: Sage. Meyer, J. P., Becker, T. E., & Vandenberghe, C. (2004). Employee commitment and motivation: A conceptual analysis and integrative model. Journal of Applied Psychology, 89, 991–1007. Meyer, J. P., & Herscovitch, L. (2001). Commitment in the workplace: Toward a general model. Human Resource Management Review, 11, 299–326. Meyer, J. P., Stanley, D., Herscovitch, L., & Topolnytsky, L. (2002). Affective, continuance, and normative commitment to the organization: A meta-analysis of antecedents, correlates, and consequences. Journal of Vocational Behavior, 61, 20–52. Miller, J. G. (1978). Living systems. New York: McGraw-Hill. Mohammed, S., & Dumville, B. C. (2001). Team mental models in a team knowledge framework: Expanding theory and measurement across disciplinary boundaries. Journal of Organizational Behavior, 22, 89–106. Mumford, M. D., Dansereau, F., & Yammarino, F. J. (2000). Followers, motivations, and levels of analysis: The case of individualized leadership. Leadership Quarterly, 11, 313–340. Naughton, T. J. (1982). An empirical examination of a theory of option-cutting and commitment. Unpublished doctoral dissertation, State University of New York at Buffalo. Pedhazur, E. J. (1982). Multiple regression in behavioral research. New York: Holt, Rinehart, & Winston. Price, J. L., & Mueller, C. W. (1986). Handbook of organizational measurement. Marshﬁeld, MA: Pitman. Roberts, K. H., Hulin, C. L., & Rousseau, D. M. (1978). Developing an interdisciplinary science of organizations. San Francisco, CA: Jossey-Bass. Robinson, W. S. (1950). Ecological correlations and the behavior of individuals. American Sociological Review, 15, 351–357. Rousseau, D. M. (1985). Issues of level in organizational research: Multi-level and cross-level perspectives. Research in Organizational Behavior, 7, 1–37. Shaw, M. R. (1976). Group dynamics. New York: McGraw-Hill. Staw, B. M. (1976). Knee-deep in the big muddy: A study of escalating commitment to a chosen course of action. Organizational Behavior and Human Performance, 16, 27–44. Taggar, S., & Ellis, R. (2007). The role of leaders in shaping formal team norms. Leadership Quarterly, 18, 105–120. Thibault, J. W., & Kelley, H. H. (1959). The social psychology of groups. New York: Wiley. 60 FRANCIS J. YAMMARINO AND FRED DANSEREAU

Wageman, R. (1995). Interdependence and group effectiveness. Administrative Science Quarterly, 40, 145–180. Wallis, N. C. (1999). Follow the leader: Understanding the initiation of individualized leadership from follower and leader perspectives. Unpublished doctoral dissertation, The Fielding Institute. Weber, M. (1924/1947). The theory of social and economic organization (T. Parsons, translator). New York: Free Press. Wolfram, S. (2002). A new kind of science. Champaign, IL: Wolfram Media. Yammarino, F. J. (1994). A conceptual–empirical approach for testing meso and multi-level theories. In: H. L. Tosi (Ed.), Extensions of the environment/organization/person model (pp. 20–50). Stamford, CT: JAI Press. Yammarino, F. J. (1995). Dyadic leadership. Journal of Leadership Studies, 2(4), 50–74. Yammarino, F. J. (1996). Group leadership: A levels of analysis perspective. In: M. A. West (Ed.), The handbook of work group psychology (pp. 189–224). Chichester, UK: Wiley. Yammarino, F. J. (1998). Multivariate aspects of the varient/WABA approach: A discussion and leadership illustration. Leadership Quarterly, 9, 203–227. Yammarino, F. J. (2003). Modern data analytic techniques for multisource feedback. Organizational Research Methods, 6(1), 6–14. Yammarino, F. J., & Bass, B. M. (1991). Person and situation views of leadership: A multiple levels of analysis approach. Leadership Quarterly, 2, 121–139. Yammarino, F. J., & Dansereau, F. (Eds). (2002a). The many faces of multi-level issues. Vol. 1 of Research in Multi-Level Issues. Oxford, UK: Elsevier Science. Yammarino, F. J., & Dansereau, F. (2002b). Individualized leadership. Journal of Leadership and Organization Studies, 9, 90–99. Yammarino, F. J., & Dansereau, F. (Eds). (2004). Multi-level issues in organizational behavior and processes. Vol. 3 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Yammarino, F. J., & Dansereau, F. (Eds). (2006). Multi-level issues in social systems. Vol. 5 of Research in Multi-Level Issues. Oxford, UK: Elsevier. Yammarino, F. J., Dansereau, F., & Kennedy, C. (2001). A multiple-level multidimensional approach to leadership: Viewing leadership through an elephant’s eye. Organizational Dynamics, 29, 149–163. Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. Leadership Quarterly, 16, 879–919. Yammarino, F. J., & Dubinsky, A. J. (1992). Superior–subordinate relationships: A multiple levels of analysis approach. Human Relations, 45, 575–600. Yammarino, F. J., & Markham, S. E. (1992). On the application of within and between analysis: Are absence and affect really group-based phenomena? Journal of Applied Psychology, 77, 168–176 (correction, 77, 426). Yammarino, F. J., & Naughton, T. J. (1988). Time spent communicating: A multiple levels of analysis approach. Human Relations, 41, 655–676. THICK OR THIN? A FUNDAMENTAL QUESTION IN ORGANIZATIONAL BEHAVIOR

Neal M. Ashkanasy

ABSTRACT

In their essay, Yammarino and Dansereau (2009) present a new multilevel theory of organizational behavior (OB), based on the idea that simplest solutions are usually the best. This commentary discusses the strengths and shortcomings of their essay and suggests some alternative strategies. In particular, it identiﬁes three core areas for discussion. The ﬁrst is that the authors set out a ‘‘thin’’ theory, at odds with the idea that OB in real organizations is inherently complex and addressable only through ‘‘thick’’ descriptions. Second, while the theory covers four levels of analysis, the authors may have neglected the time dimension. Third, the theory is an example of ‘‘grand’’ theorizing, suggesting it may also share the disappointing fate of such theories in the past.

INTRODUCTION

The imperative for simplicity in theory is an important one, and I applaud Yammarino and Dansereau (2009) for their insights presented in their superb essay. The imperative for simplicity in theory building is often

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 61–67 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008002 61 62 NEAL M. ASHKANASY forgotten. This is especially the case in the field of organizational behavior (OB), where so many theories have come and gone over the past 50 years, and where our understanding of the discipline seems to be mired in complex and competing, even sometimes unfathomable, explanations. The idea of simplicity in theory building is not a new one. As long ago as the fourteenth century, Franciscan friar William of Occam expounded on the ‘‘lex parsimoniae’’ (law of parsimony), more commonly known as ‘‘Occam’s Razor.’’ In Latin, this principle is stated as ‘‘Entia non sunt multiplicanda praeter necessitate,’’ which translates to ‘‘Entities must not be multiplied beyond necessity’’ or, in plain English, ‘‘All things equal, the simplest solution is always best.’’ Indeed, Occam’s Razor is generally regarded as the foundation of modern scientific method. It is why, for example, scientists reject the null hypothesis (the simpler model) only when the evidence to the contrary is overwhelming (see Popper, 1959, especially Chapter 7). Clearly, the framework proposed by Yammarino and Dansereau fits into the mold of Occam’s Razor. I especially appreciate how the authors have formulated their framework in terms of levels of analysis – although this is not surprising, given their outstanding contributions to multilevel theory over the past decades. Moreover, the four theories they describe have an essential elegance that makes them intuitively attractive. I have little doubt that researchers will take up their challenge to begin testing these intriguing theories and to use this framework to develop further relationships along similar lines to those elucidated by Yammarino and Dansereau. Nonetheless, three aspects of the Yammarino and Dansereau theory deserve further attention:

Theirs is inherently a ‘‘thin’’ theory. Is there still a place for ‘‘thick’’ explanations of behavior in organizations? While Yammarino and Dansereau cover four levels of analysis, they may have missed one important level, time. Yammarino and Dansereau’s chapter represents an attempt at ‘‘grand theory’’ building. Such theories do not have a good track record, so why should we go along with this one?

This chapter addresses each of these issues in turn.

THICK VERSUS THIN EXPLANATIONS

One of the fundamental debates in OB concerns the issue of whether explanations of behavior should be thick, detailed and contextualized, or thin, simplified but ostensibly context free. Both Yammarino and Dansereau A Fundamental Question in Organizational Behavior 63 come from an epistemological background of logical positivism, where phenomena are assumed to be accessible via scientific method (Lincoln & Guba, 2000), so their natural inclination is toward thin explanations. But some of the most esteemed publications in the OB discipline are derived from thick explanations, as exemplified by the writings of Karl Weick (1993). Weick’s descriptions of behavioral phenomena are detailed and interpretive, yet the principles derived from his work are widely cited and widely taught in OB classrooms. [At the time of this writing, the Social Sciences Citation Index listed a total of 253 citations to Weick (1993).] Another example that combines qualitative and quantitative approaches is Meyer’s (1982) study of change in San Francisco hospitals, which he characterized as ‘‘adapting to environmental jolts’’ (p. 515). Interestingly, both Weick and Mayer derived theoretical positions that were intuitive yet illuminating. It will be interesting to see if Yammarino and Dansereau’s multilevel theory will ultimately yield explanations with the same level of intuitive elegance. A further downside of strongly theoretical approaches to studying OB is that theories can end up becoming Procrustean (after the Greek mythological giant Procrustes, who cut off the legs and arms of his victims to make them fit his bed). In other words, scholars may try to force naturally occurring phenomena to fit a theoretical position, even if this involves metaphorically cutting off the arms and legs of real phenomena. Advocates of grounded theory (see Corbin & Strauss, 1990), by contrast, argue that data determine theoretical positions, not the other way around. In the specific context of leadership research, for example, Parry (1998) advocates that leadership as a social process is beyond abstract theorizing, and that only a grounded theory approach can ‘‘overcome the deficiencies’’ (p. 85) in the traditional approaches. Of course, as Yammarino and Dansereau point out, there is an inherent risk that such contextualized explanations will become overly conditional and require qualifications that render them opaque. Nonetheless, advocates of context recognition in OB such as Johns (2006) and Rousseau and Fried (2001) would see this narrowing of the explanations as a necessity if real phenomena are to be appropriately dealt with in the OB literature.

IS TIME MISSING?

Since the inception of OB as a discipline, scholars have tended to present phenomena in terms of static relationships. Yammarino and Dansereau’s model appears to continue this tradition. The levels they address encompass individuals, dyads, groups, clusters, and organizations. 64 NEAL M. ASHKANASY

Of course, behavior is not static in real life. People in organizations fluctuate from moment to moment in terms of their performance levels, cognitive function, and emotion. The need for researchers to address within- person temporal fluctuations emerged as a component of the shift to study emotions in organizations (Fisher, 2008; Weiss & Brief, 2001). As Weiss and Cropanzano (1996) point out, it simply does not make sense to study emotions using the traditional (static) between-person models of OB. More recently, Fisher and Noble (2004) have shown how traditional theories of OB that neglect within-person temporal variability lead to erroneous conclusions about the relatedness of job performance and satisfaction (cf. Judge, Thoresen, Bono, & Patton, 2001). Thus, the following question arises here: What would their model look like if Yammarino and Dansereau were to add an additional level to account for within-person variations of behavior and motivation? Would recognition of this level fundamentally alter the model they present? Would it negate the claimed simplicity of their model? Clues to the answers to these questions may be found in the multilevel model of emotion that I published earlier in this book series (Ashkanasy, 2003). In this model, I addressed five levels: (1) within person, (2) between person, (3) dyadic, (4) group, and (5) organization. More importantly, I presented a model integrating the five levels that was based around a within-person perspective [Weiss and Cropanzano’s (1996) affective events theory]. In this sense, temporal variations in behavior, cognition, and affect were envisaged to permeate all levels of the model. For example, groups and even organizations fluctuate in terms of performance, group cognition, and group affect, just as individuals do. As an example, Sy, Coˆte´, and Saavedra (2005) found that the meso-level process of emotional contagion led to temporal fluctuations in group affective states and performance.

GRAND THEORIES

Albert Einstein spent the greater part of his career in a futile search for the Grand Uniﬁcation Theory that would unify time and space (Hawking, 2001), and physicists today continue his search – so far with no luck, although some maintain they are close (e.g., see Mohapatra, 2002). If a unifying theory in the physical science has proved this elusive, why should we expect to be able to identify such a theory in the context of the much more complex and subjective world of human behavior, especially in one isolated corner of OB? Even a brief list of the complexities of behavior – for A Fundamental Question in Organizational Behavior 65 example, bounded rationality (Simon, 1997), bounded emotionality (Mumby & Putnam, 1992), and cross-cultural psychology (House, Hanges, Javidan, Dorfman, & Gupta, 2004) – would suggest the futility of such a quest. Another issue is that so many of the grand theories in OB have failed to withstand the test of time. Maslow’s and Herzberg’s theories of motivation come to mind here. Although still taught to students of OB, most textbooks mention that these theories have been discredited in empirical testing (e.g., see Robbins & Judge, 2008). Major theories of leadership (e.g., path-goal theory; see House, 1971), although again still featured prominently in modern-day textbooks, have met with similar fates. All of these theories purported to model important aspects of OB, yet were found wanting when put to the test. Yammarino and Dansereau respond that theirs is a ‘‘new’’ OB, and is differentiated from traditional theory-building approaches in OB. But their approach depends on the veracity of the propositions set out in Wolfram’s (2002) A New Kind of Science and, in particular, the Principle of Computational Equivalence, which in turn depends on the existence of a set of universal, underlying rules. The answer, of course, is that Yammarino and Dansereau’s theory probably won’t stand up in the long run. This statement is not meant to imply that I am advocating their theory is not a good one; nor do I posit that OB researchers should prematurely consider consigning it to the trash. On the contrary, although essentially a critical realist in terms of my own approach, I am a strong advocate for an elective approach that does not constrain research approaches based on either ontology or epistemology. Each ontological perspective provides a unique view of phenomena. In Yammarino and Dansereau’s model, what is truly unique is the multilevel formulation, where each of the four theories is grounded in the simple universal theory they advocate. At the very least, this approach enables us to appreciate an underlying connection across the theories that is not otherwise obvious. Moreover, each of the theories provides us with a new lens to understand behavior in organizations.

CONCLUSION

While I have expressed some misgivings about the theoretical model proposed by Yammarino and Dansereau, I have no doubt that it will come to be seen as an important and impactful contribution to the OB literature. 66 NEAL M. ASHKANASY

Controversies surrounding the issue of thick versus thin descriptions and explanations of OB phenomena, however, will be with us for some time yet.

REFERENCES

Ashkanasy, N. M. (2003). Emotions in organizations: A multilevel perspective. In: F. Dansereau & F. J. Yammarino (Eds), Multi-level issues in organizational behavior and strategy. Vol. 2 of Research in Multi-Level Issues (pp. 9–54). Oxford, UK: Elsevier/ JAI Press. Corbin, J., & Strauss, A. (1990). Grounded theory research: Procedures, canons, and evaluative criteria. Qualitative Sociology, 13, 3–21. Fisher, C. D. (2008). What if we took within-person performance variability seriously? Industrial and Organizational Psychology: Perspectives on Science and Practice, 1, 185–189. Fisher, C. D., & Noble, C. S. (2004). A within-person examination of correlates of performance and emotions while working. Human Performance, 17, 145–168. Hawking, S. (2001). The illustrated brief history of time, updated and expanded. New York: Bantam. House, R. J. (1971). A path-goal theory of leader effectiveness. Administrative Science Quarterly, 16, 321–339. House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W., & Gupta, V. (2004). Culture, leadership, and organizations: The GLOBE study of 62 societies. Thousand Oaks, CA: Sage. Johns, G. (2006). The essential impact of context on organizational behavior. Academy of Management Review, 31, 386–408. Judge, T. A., Thoresen, C. J., Bono, J. E., & Patton, G. K. (2001). The job satisfaction–job performance relationship: A qualitative and quantitative review. Psychological Bulletin, 127, 376–407. Lincoln, R. S., & Guba, E. G. (2000). Paradigmatic controversies, contradictions, and emerging confluences. In: N. K. Denzin & Y. S. Lincoln (Eds), Handbook of qualitative research (2nd ed., pp. 163–188). Thousand Oaks, CA: Sage. Meyer, A. D. (1982). Adapting to environmental jolts. Administrative Science Quarterly, 27, 515–537. Mohapatra, R. (2002). Unification and supersymmetry. Oxford, UK: Oxford University Press. Mumby, D. K., & Putnam, L. A. (1992). The politics of emotion: A feminist reading of bounded rationality. Academy of Management Review, 17, 465–486. Parry, K. W. (1998). Grounded theory and social process: A new direction for leadership research. Leadership Quarterly, 9, 85–105. Popper, K. (1959). The logic of scientific discovery. New York: Basis Books. Robbins, S. P., & Judge, T. A. (2008). Organizational behavior (13th ed.). Upper Saddle River, NJ: Pearson. Rousseau, D. M., & Fried, Y. (2001). Location, location, location: Contextualizing organizational research. Journal of Organizational Behavior, 22, 1–13. Simon, H. A. (1997). Administrative behavior (4th ed.). New York: Free Press. A Fundamental Question in Organizational Behavior 67

Sy, T., Coˆte´, S., & Saavedra, R. (2005). The contagious leader: Impact of the leader’s mood on the mood of group members, group affective tone, and group processes. Journal of Applied Psychology, 90, 295–305. Weick, K. E. (1993). The collapse of sensemaking in organization: The Mann Gulch disaster. Administrative Science Quarterly, 38, 628–652. Weiss, H., & Cropanzano, R. (1996). Affective events theory: A theoretical discussion of the structure, causes and consequences of affective experiences at work. In: L. L. Cummings & B. M. Staw (Eds), Research in organizational behavior (Vol. 18, pp. 1–74). Westport, CT: JAI Press. Weiss, H. M., & Brief, A. P. (2001). Affect at work: A historical perspective. In: R. L. Payne & C. L. Cooper (Eds), Emotions at work: Theory, research, and applications for management (pp. 133–172). Chichester, UK: Wiley. Wolfram, S. (2002). A new kind of science. Champaign, IL: Wolfram Media. Yammarino, F. J., & Dansereau, F. (2009). A new kind of OB. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald.

This page intentionally left blank

A NEWER ORGANIZATIONAL BEHAVIOR

Francis J. Yammarino and Fred Dansereau

ABSTRACT

This reply to Ashkanasy’s commentary (2009) on our chapter ‘‘A New Kind of OB’’ (Yammarino & Dansereau, 2009) offers some additional thoughts on theory building and theory testing in terms of our four simple theories about individual behavior, interpersonal relationships, group dynamics, and collectivized processes in organizations. We reiterate the key point that these simple notions can explain a variety of complex phenomena and behavior in organizations.

INTRODUCTION

We thank Neal Ashkanasy for his thorough and insightful commentary (2009) on our chapter ‘‘A New Kind of OB’’ (Yammarino & Dansereau, 2009). Neal’s depth of knowledge in so many areas of organizational behavior (OB) has allowed him to offer a series of comments on our work that few others would have been able to provide. In this reply to his commentary, after a brief recapitulation, we address three key issues raised by Neal – namely, formulating (1) ‘‘thin’’ theories (2) without the inclusion of ‘‘time’’ that (3) constitute a ‘‘grand theory.’’ We then offer another

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 69–76 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008003 69 70 FRANCIS J. YAMMARINO AND FRED DANSEREAU perspective on the dialogue to enhance theory building and theory testing in OB, especially in terms of our four simple theories about individual behavior, interpersonal relationships, group dynamics, and collectivized processes in organizations.

BRIEF RECAPITULATION

A key point of our work (Yammarino & Dansereau, 2009) follows directly from the work of Wolfram (2002): Simple notions can explain complex phenomena. Wolfram asserted and then demonstrated in a variety of disciplines and areas (e.g., mathematics, physics, biology, social sciences, computer science, philosophy, art, technology, artificial intelligence) that simple rules (ideas, notions, theories) can lead to simple or complex phenomena. Moreover, Wolfram (2002) demonstrated that more complex or complicated ‘‘rules’’ do not ultimately lead to more complex behavior or add fundamentally new features. This is a point we have tried to make previously (Dansereau, Alutto, & Yammarino, 1984) as well as here (Yammarino & Dansereau, 2009) regarding complex theories of OB as compared to our four simple theories. This trend is consistent with and continues across the sciences. A few examples should suffice. In his comprehensive review of evolutionary theory, Gould (2002) first recounts the history of evolution from simple to complex and then wends his way back to simple explanations of evolution and its various aspects and applications. In physics, Ambjorn, Jurkiewicz, and Loll (2008) have proposed a back-to-basics approach to reconcile quantum theory and Einstein’s general theory of relativity. Unlike complex string theory approaches to reconciliation that require multiple components (both ‘‘real’’ and hypothetical) and n dimensions, their theory of quantum gravity (called casual dynamical triangulations) relies on the simpler Euclidean quantum gravity theory proposed by Stephen Hawking (1988) and four-dimensional space- time approximated as a mosaic of triangles. In a study of human origins and migrations utilizing DNA analysis, Stix (2008) contrasts the current competing theories. This more complex multi- regional theory posits that modern characteristics evolved from archaic hominid populations in Africa, Asia, and Europe that interbred so they remained a single species. The simpler out-of-Africa theory asserts that humans with modern traits left Africa and settled the world. The most recent data and DNA analysis and tracking favor the latter, simpler view. A Newer Organizational Behavior 71

So, for OB, like the other sciences, simple notions or theories can hold great promise for explaining a variety of both simple and complex behaviors in organizations. Complexity, and explaining complex behaviors, does not require complex underlying rule structures (i.e., complex theories or ideas). For us, simple rules or simple theories sufﬁce, and these simple ideas are summarized as follows:

Theory 1. A theory of individual behavior and decision making. Option cutting and commitment are positively related based on between-person differences (inter-individual differences or whole persons); this level- speciﬁc theory emerges at the person level of analysis.

Theory 2. A theory of interpersonal relationships and leadership. Investments and returns are positively related based on between-dyad differences (whole dyads); this level-speciﬁc theory emerges at the dyad level of analysis.

Theory 3. A theory of group dynamics/team processes and norms. Interdependence and cohesion are positively related based on between- group differences (whole groups); this level-speciﬁc theory emerges at the group level of analysis.

Theory 4. A theory of collectivized processes and roles. Titles and expectations are positively related based on between-collectivity differences (whole collectives); this cross-level theory emerges from person to dyad to group to collective levels of analysis.

KEY ISSUES FROM ASHKANASY

Ashkanasy (2009), while complimenting us on the comprehensiveness and elegance of our work (Thank you very much!) and endorsing the notion of simplicity in science (a` la Occam’s Razor), notes some limitations with our approach to theory building and theory testing for a new kind of OB. His position, simply stated, is that formulating (1) ‘‘thin’’ theories (2) without the inclusion of ‘‘time’’ that (3) constitute a ‘‘grand theory’’ has not worked in the past and will not work in the future to advance theory building and theory testing in OB. We respond brieﬂy to each of Ashkanasy’s points. 72 FRANCIS J. YAMMARINO AND FRED DANSEREAU

‘‘Thin’’ Theories

Ashkanasy (2009) indicates that a fundamental debate in OB is whether explanations of behavior should be ‘‘thick’’ (detailed and contextualized) versus ‘‘thin’’ (simpliﬁed but ostensibly context free). He notes our work as an exemplar of the latter approach, but suggests we may have neglected the former approach. He frames his comments in terms of different philosophy of science positions (e.g., phenomenology versus logical positivism, respectively) and research methodologies (i.e., qualitative versus quantitative methods, respectively). For us, this ‘‘debate’’ is illusory, as the best theory building should be able to begin from any philosophical position – or, more accurately, without regard for philosophical position. Similarly, the best theory testing should be able to employ any methodology – or, more accurately, both qualitative and quantitative methods, seeking triangulation between them. Whether one begins with a thick approach (e.g., studying one instance intensely with qualitative methods) or a thin approach (e.g., searching for general laws with quantitative methods) is irrelevant. What does matter is that both approaches are integrated at some point to thoroughly understand the notions of interest. In qualitative work on Theory 2 (investments and returns in individualized leadership) using a grounded theory approach, for example, these same underlying simple notions were revealed (Wallis, 1999; Wallis & Yammarino, 2008). We have also employed unobtrusive and qualitative methods along with quantitative ones in other research on Theory 2 (see Dansereau et al., 1995). Our point is that combining approaches ultimately provides an opportunity to resolve an Aristotelian either-or controversy with a Galilean integration. While our preference is to begin with one, we clearly would not dismiss the other. But, like the obesity problem in the United States, the ﬁeld of OB is overweight and needs to thin down!

Inclusion of ‘‘Time’’

Ashkanasy speciﬁes and gives examples of the importance of ‘‘time’’ in OB theory building and theory testing. He also asks in a subheading whether ‘‘time’’ is missing in our work. We fully agree with him about the importance of the time issue and the inclusion of developmental notions as well as the need to examine changes over time in OB. Moreover, we can answer his question with an unqualiﬁed ‘‘No!’’ A Newer Organizational Behavior 73

In our original work on these matters (Dansereau et al., 1984), we provided an extended treatment on the inclusion of time in theory building and theory testing (see pp. 264–275). Subsequently, we (Dansereau, Yammarino, & Kohles, 1999) elaborated notions regarding stability and changes over time in variables/relationships as well as levels of analysis/ entities. In our work on Theory 2 (investment and returns in individualized leadership), for example, we specifically included tests of developmental notions over time both quantitatively (Dansereau et al., 1995) and qualitatively (Wallis, 1999; Wallis & Yammarino, 2008). Moreover, for each of our four theories presented in the current volume (Yammarino & Dansereau, 2009), we specifically discuss developmental ideas in their formulation and the etiology of the effects from an evolutionary perspective. To Ashkanasy’s specific point, for Theory 1, we discussed within-person variation, differences in persons, and potential changes and development in individuals’ option cutting and commitment over time. These were presented as alternative explanations framed in terms of other levels of analysis effects for testing. So, yes, for our four theories, we wholeheartedly agree with and endorse Ashkanasy’s position that temporal variations and time matter!

A ‘‘Grand Theory’’

Ashkanasy characterizes our work as being in the realm of a ‘‘grand theory,’’ based on the assumption of underlying law-like principles, and ultimately subject to the same fate as prior similar work – scholarly death! Ashkanasy may prove correct, but at this time we are unwilling to accept that fate. Instead, we take a somewhat different view on this matter. Again, in our original work on these ideas (Dansereau et al., 1984), we addressed the notion of grand theories in OB (see pp. 65–66). Our position then and now is that any attempt to make distinctions among grand, middle-range, and simple theories is arbitrary as the differences are of degree rather than kind. We prefer to focus on general theories or law-like propositions ﬁrst (see Yammarino & Dansereau, 2009, Table 1), as we did for each of our four theories (with two variables, a relationship, and the entities of focus), and then present more speciﬁc theories for testing per se (see Yammarino & Dansereau, 2009, Table 4). Moreover, we mentioned the possibilities and inclusion of mediators and moderators for our four theories, preferring to begin with simple notions and test them well with replication before proceeding to (perhaps unnecessary) elaborations. 74 FRANCIS J. YAMMARINO AND FRED DANSEREAU

Even accepting the grand theory label and the failures of such theories in the past does not imply we should stop trying. In physics, for example, although work on the ‘‘four forces’’ separately has gone quite well, work on the GUT or ‘‘theory of everything’’ has been less than fully successful. Nevertheless, physicists have not given up, and new work continues to seek this desirable end state (see Ambjorn et al., 2008). Likewise, in a variety of other areas and ﬁelds, Wolfram (2002) shows how such grand uniﬁcations are possible, plausible, and derived from simple notions or theories, not complex ones. While not in our lifetimes, after fully understanding the ‘‘four forces of OB’’ presented in our chapter, we remain optimistic regarding the emergence of a similar GUT for OB!

ANOTHER PERSPECTIVE ON THE DIALOGUE

We continue to believe there is some value in four ‘‘thin’’ theories that include ‘‘time’’ and that someday may constitute a ‘‘grand theory’’ of OB. In fact, these four simple notions can be used to provide some entertainment regarding our commentary–reply dialogue here about theory building and theory testing in OB. Ashkanasy’s view – which is perhaps the view of the vast majority of the OB ﬁeld – can be summarized as follows using our four theories:

1. At the extreme there are two options: pursue traditional OB approaches or pursue our ‘‘new kind of OB’’ approach. Believing our approach will not work (given that other approaches, perhaps similar to ours, have failed in the past), this option is cut off and the field becomes even more committed to doing ‘‘the same old OB.’’ 2. Investing time and effort in a new approach and others who wish to pursue it will not result in personal gains, so the formation of strong dyadic interpersonal relationships will be with others who share the traditional or older set of interests. 3. Taking a Weickian view in the field (for example) can result in advancements for endorsing the group norms. Essentially, there is cooperation within traditional groups with others holding the Weickian view, so cohesiveness and interdependence within the group of Weickian scholars (and those in allied fields) persist. 4. The new OB approach (title) is logical positivism and quantitative methods, so it is limited and likely to fail (expectations). A Newer Organizational Behavior 75

In contrast, our position – though perhaps a minority view within the OB field – can be summarized as follows using these same four theories: 1. The current options are not very attractive to us, so we have created a new option, cut off traditional ones, and become committed to this new approach to OB. 2. Investing time and effort in our approach and others who wish to pursue it will result in personal gains for us (e.g., the process of scientific discovery is rewarding) as well as the formation of strong dyadic interpersonal relationships based on common interests. 3. Taking a non-Weickian view in the field (for example) can result in sanctions for violating the group norms. Essentially, cooperation between groups with others holding the Weickian view will be limited, but there are cohesiveness and interdependence both within our group (and the physical sciences) with this new approach and within the group of Weickian scholars (and allied fields) regarding their approach. 4. Labeling (titles) individuals, groups, and collectives as logical positivists and/or quantitative methodologists creates expectations about them. Likewise, labeling (titles) individuals, groups, and collectives as grounded theorists and/or qualitative methodologists creates different expectations about them. Similarly, labeling (titles) individuals, groups, and collectives as ‘‘integrators’’ creates still different expectations about them and their pursuits.

CONCLUSION

We again thank Neal Ashkanasy for his insightful and thoughtful commentary on our work. Through this dialogue among scholars about simple notions explaining complex behaviors in organizations, we hope to stimulate others to enhance theory building and theory testing for a newer OB – one that explains and tests individual behavior, interpersonal relations, group dynamics, and collectivized processes in organizations to advance our ﬁeld further faster.

REFERENCES

Ambjorn, J., Jurkiewicz, J., & Loll, R. (2008). The self-organizing quantum universe. Scientiﬁc American, 299(1), 42–49. 76 FRANCIS J. YAMMARINO AND FRED DANSEREAU

Ashkanasy, N. M. (2009). Thick or thin? A fundamental question in OB. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory testing in organizational behavior: The varient approach. Englewood Cliffs, NJ: Prentice Hall. Dansereau, F., Yammarino, F. J., & Kohles, J. (1999). Multiple levels of analysis from a longitudinal perspective: Some implications for theory building. Academy of Manage- ment Review, 24, 346–357. Dansereau, F., Yammarino, F. J., Markham, S. E., et al. (1995). Individualized leadership: A new multiple-level approach. Leadership Quarterly, 6, 413–450. Gould, S. J. (2002). The structure of evolutionary theory. Cambridge, MA: Belknap Press of Harvard University Press. Hawking, S. (1988). A brief history of time. New York: Bantom Books. Stix, G. (2008). Traces of a distant past. Scientiﬁc American, 299(1), 56–63. Wallis, N. C. (1999). Follow the leader: Understanding the initiation of individualized leadership from follower and leader perspectives. Unpublished doctoral dissertation, The Fielding Institute. Wallis, N. C., & Yammarino, F. J. (2008). Individualized and transformational leadership: A qualitative study of senior executive leaders. Manuscript under revise – and – resubmit invitation for Leadership Quarterly. Wolfram, S. (2002). A new kind of science. Champaign, IL: Wolfram Media. Yammarino, F. J., & Dansereau, F. (2009). A new kind of OB. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald.

PART II OUTSTANDING LEADERSHIP

This page intentionally left blank

CHARISMATIC, IDEOLOGICAL, AND PRAGMATIC LEADERSHIP: AN EXAMINATION OF MULTI-LEVEL INFLUENCES ON EMERGENCE AND PERFORMANCE

Michael D. Mumford, Samuel T. Hunter, Tamara L. Friedrich and Jay J. Caughron

ABSTRACT

Theories of outstanding, historically notable, leadership have traditionally emphasized charisma. Recent research, however, suggests that charisma may represent only one pathway to outstanding leadership. Outstanding leadership may also emerge from ideological and pragmatic leadership. In this article, we examine the conditions inﬂuencing the emergence and performance of charismatic, ideological, and pragmatic leaders. It is argued that different conditions operating at the environmental, organizational, group, and individual levels inﬂuence the emergence and performance of each of these three types of leaders. Implications for understanding the origins and impact of charismatic, ideological, and pragmatic leaders are discussed.

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 79–116 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008004 79 80 MICHAEL D. MUMFORD ET AL.

INTRODUCTION

Traditionally, students of leadership have sought to identify the individual and situational variables that influence leader emergence and performance in routine, day-to-day, organizational settings. Thus, the literature has stressed behaviors such as consideration, initiating structure, participation, and change management (e.g., Fleishman, 1953; Hunt, 2004; Marta, Leritz, & Mumford, 2005; Yukl, 2006), and situational variables such as follower expertise (Hersey & Blanchard, 1982), leader–follower relationships (Graen & Uhl-Bien, 1998), and the degree of task structure (House, 1971) that might moderate the impact of these behaviors on leader emergence and performance. Although this research has contributed much to our understanding of leadership in its normative form, it is open to question whether it has told us much about incidents of truly outstanding leadership – incidents where leaders exercise exceptional influence over followers to obtain notable results (Bass, 1985; Mumford, 2006). Recognition of the limitations of normative leadership theory in accounting for incidents of outstanding leadership, for example, Winston Churchill in the Battle of Britain, Michael Collins in the foundation of the Irish Republic, and Thomas Watson in the creation of IBM, has led students of leadership to seek to identify the attributes of leaders that make these incidents of outstanding leadership possible (House, 1977). Theories of charismatic and transformational leadership have become the dominant models applied in attempts to account for incidents of outstanding leadership (Avolio, Howell, & Sosik, 1999; Bass & Avolio, 1990; Conger, 1999; Conger & Kanungo, 1988; House & Howell, 1992). Although theories of charismatic and transformational leadership differ from each other in some notable ways, they share a common core (Hunt, 1999; Yukl, 1999). More specifically, they hold that outstanding leadership is based on effective articulation of a future-oriented vision that motivates and directs others while providing a sense of meaning and affective engagement (Bass, 1990; Shamir, House, & Arthur, 1993). In fact, the available evidence indicates that a leader’s articulation of a viable vision is positively related to various indices of organization performance (Deluga, 2001; Kirkpatrick & Locke, 1996), follower motivation (Sosik, Kahai, Avolio, 1999), effective group interaction (Parry & Proctor-Thomson, 2003), and satisfaction with both the leader and the group (Dumdum, Lowe, & Avolio, 2002). Although there is evidence available indicating that charismatic visioning may contribute to outstanding leadership, charismatic theories have been subject to some noteworthy criticisms (Beyer, 1999). Outstanding Leadership 81

For example, charismatic leadership seems to exert stronger effects in bureaucratic than non-bureaucratic organizations (Lowe, Kroeck, & Sivasubramaniam, 1996) and may prove ineffective in some groups such as research and development professionals (Mumford, Scott, Gaddis, & Strange, 2002). Similarly, de Hoogh, den Hartog, and Koopman (2004) found that charismatic leadership was more strongly related to perceptual outcomes (e.g., subordinate attitudes) than some organizational criteria (e.g., organizational liquidity and solvency). These perceptual relationships, moreover, appeared stronger under conditions of environmental uncertainty. There is also evidence indicating sociocultural context may moderate the relationship of transformational leadership and organizational innovation (Elenkov & Manev, 2005), highlighting the notion that charismatic leadership may not prove equally effective across all conditions. Finally, charismatic leadership theories seem to discount, or ignore, some key functions of leaders such as planning and decision-making and may place too great of an emphasis on dyadic relationships (Yukl, 1999, 2006). These observations led Mumford and his colleagues (Mumford, 2006; Mumford & Van Doorn, 2001; Strange & Mumford, 2002) to argue that it may be necessary to examine alternative pathways to, or alternative types of, outstanding leadership. Drawing from earlier work by Weber (1924), they argued that three distinct types of outstanding leadership may exist – types they have labeled as charismatic leadership (e.g., John F. Kennedy), ideological leadership (e.g., Ronald Regan), and pragmatic leadership (e.g., Dwight Eisenhower). In the present chapter, we will examine the conditions shaping the emergence and performance of charismatic, ideological, and pragmatic leaders at the environmental, organizational, group, and individual levels. Before examining these multi-level inﬂuences on charismatic, ideological, and pragmatic leadership; however, it would seem germane to consider the general model of outstanding leadership giving rise to these three alternative pathways.

OUTSTANDING LEADERSHIP

Crises and Mental Models

Perhaps the most straightforward conclusion one can draw about outstanding leadership is that outstanding leadership requires placing the right person in the right situation. Abraham Lincoln’s greatness is inexorably linked to the U. S. Civil War. Bill Gates’ achievements at Microsoft 82 MICHAEL D. MUMFORD ET AL. depended on the phase of development of the computer industry. These observations about the situations giving rise to outstanding leadership; however, point to a broader conclusion. More specifically, outstanding leadership appears to emerge under conditions of crisis, change, and turbulence (Beyer, 1999). This point is illustrated by John F. Kennedy, who, when discussing Lincoln’s greatness appeared keenly aware of the role crises played in outstanding leadership: ‘‘War, he observed, made it easier for a [Lincoln] to achieve greatness’’ (Schlesinger, 1997, p. 180). Some support for the above conclusion may be found in a study by Hunt, Boal, and Dodge (1999) who asked management students to work on two business tasks under crisis and noncrisis conditions where a crisis involved high-priority goals where little response time was available. They found that visionary charismatic leadership proved especially important to follower perceptions of leadership under crisis conditions. Some further support for this conclusion has been provided by Halverson, Holladay, Kazma, and Quinones (2004) who found that followers were more likely to attribute charisma to a leader under conditions of crisis, as well as Pillai and Meindl (1998) who found that crisis conditions cause people to prefer leaders evidencing charisma. Similarly, Merolla, Ramos, and Zechmeister (2007) found that President George W. Bush was rated higher on charisma in experimentally manipulated crisis conditions, with these perceptions of charisma also impacting perceptions of blame and failure in Iraq. Finally, Chiozza and Goemans (2003) examined 1,505 leaders from the years 1919–1992 and found that as risk of international violence increased, leaders were less likely to lose office. These studies appear to indicate that, crisis, or change, creates conditions where the emergence of charismatic leaders, and perhaps other forms of outstanding leadership, becomes possible. One potential explanation for the impact of crisis, change, and turbulence on outstanding leadership is that these situations, by undermining normative routines, provide leaders with discretion allowing the exercise of exceptional influence (Lowe, 2001). Another potential explanation for these effects; however, may be found in Mumford (2006). He argued that under conditions of crisis, change, and turbulence, the behavior of complex social systems becomes unpredictable. This unpredictability not only allows for the emergence of new threats and opportunities, but the actions needed to effectively respond to these threats and opportunities will be ambiguous, or unclear, because people lack an understanding of the causes and consequences of these change events. Under these conditions, leaders can exercise exceptional influence by engaging in sensemaking activities that clarify goals and paths to goal Outstanding Leadership 83 attainment (Weick, 1995). These sensemaking activities on the part of leaders induce feelings of control, provide a framework for collective action, reduce perceptions of threat, clarify opportunities, and minimize the feeling of anomie and identity diffusion associated with change. Some support for this conclusion may be found in a series of studies by Gioia and his colleagues (Gioia & Thomas, 1996; Gioia, Thomas, Clark, & Chittipeddi, 1994). They examined sensemaking activities on the part of people moving into leadership positions in academic institutions. They found that leader sensemaking provided a basis for guiding organizational change – although this sensemaking may, at times, depend on sense- breaking or the prior reconfiguration of extant interpretive structures being applied by followers (Fiol, Harris, & House, 1999; Pratt, 2000). Some further support for this conclusion may be found in a qualitative study of a new product development effort by Drazin, Glynn, and Kazanjain (1999). They found that successful new product development efforts require leaders to engage in sensemaking concerning the origins of crises occurring during the course of the product development effort which, in turn, provided project teams with a framework for resolving the crisis. Mumford and Strange (2002) have examined how leaders might go about sensemaking. They argued that, ultimately, sensemaking activities on the part of leaders will be based on a mental model of the social system under consideration. These mental models represent an abstract schematic knowledge structure describing the key causes of the behavior of a social system vis-a` -vis select functional goals (Frankwick, Walker, & Ward, 1994; Goldstone & Sakamoto, 2003; Hemlo-Silver & Pfeffer, 2004; Johnson- Laird, 1999). The descriptive mental model people form over time with exposure to a social system provides a basis for construction of a prescriptive mental model that represents an idealized image of this social system. That is, a prescriptive model presents a situation as it could be. The development of this prescriptive mental model is based on analysis and reconfiguration of the descriptive mental model. More specifically, in model reconfiguration people are held to analyze the goals being pursued and the key causal operatives influencing attainment of these goals in relation to personal experiences. This reflective, experiential, appraisal of goals and causes gives rise to a prescriptive mental model. This prescriptive mental model permits sensemaking, and sensemaking behavior, on the part of leaders while allowing formation of a vision as the prescriptive mental model is recast to articulate an image of the future to followers. In a recent experimental study intended to test this theory, Strange and Mumford (2005) asked undergraduates to assume the role of principal of a 84 MICHAEL D. MUMFORD ET AL. new experimental school. They were asked to write a speech describing their vision for the school. Prior to writing this speech; however, they were provided with (1) good and poor case models, (2) asked, or not asked, to analyze goals and causes, and (3) asked, or not asked, to reflect on goals and causes in relation to prior personal experiences in school. It was found that analysis of goals and causes in relation to descriptive models and prior experiences resulted in the production of stronger vision statements as appraised by students, parents, and teachers.

Charismatic, Ideological, and Pragmatic Leadership

Within this general model of outstanding leadership, charismatic, ideological, and pragmatic leadership are held to arise from the strategies leaders apply in the formation of prescriptive mental models. Charismatic leaders, in the formation of their prescriptive mental models, are held to stress goals – typically positive, future-oriented goals (Conger & Kanungo, 1988). Accordingly, charismatic leaders’ prescriptive mental models will stress goals as opposed to causes. The causes applying in available descriptive mental models will be those that are retained by charismatic leaders in formation of their prescriptive mental model. In model formation, however, charismatic leaders will tend to retain causes that are subject to control as a result of peoples’ actions. Thus, charismatic leaders define future goals pointing others to causes they might act on to bring about the attainment of these goals. In other words, charismatic leaders see people as creators of their own destiny through the actions they take on a set of known, relatively unambiguous, causes. This charismatic logic is aptly illustrated in John F. Kennedy’s statement ‘‘Ask not what your country can do for you but what you can do for your country.’’ Charismatic leaders must encourage people to act, and act on relevant causes of goal attainment. One way charismatic leaders may motivate this kind of action is through role modeling and self-sacrifice (Yorges, Weiss, & Strickland, 1999). Another way charismatic leaders can motivate people to act on the causes giving rise to goal attainment is through communication (Fiol et al., 1999). Not only will charismatic leadership require communication, it will require mass appeal intended to build commitment and encourage follower actions with regard to relevant causes of goal attainment. Thus, charismatic leaders exert influence, and maintain control, through goals while allowing followers the freedom to act on causes that will bring about goal attainment. Outstanding Leadership 85

Ideological leaders, like charismatic leaders, apply prescriptive mental models that emphasize goals. Ideological leaders, however, do not construct models around future goals. Instead, the goals emphasized by ideological leaders are those that have proven viable in the past based on the leader’s personal experiences. Causation, in the logic of ideological leaders, is not under people’s control but rather it is controlled by situational forces that are operating to undermine current attempts to attain known, valued, goals (Moghaddam, 2005; Mumford, Espejo et al., in press; Post, Ruby, & Shaw, 2002). This negative framing of causation results in a tendency for ideological leaders to attempt to induce change in the situation that will remove blockages to attainment of goals that will naturally come about under appropriate conditions. Thus, Ronald Regan believed that America must remain a shinning light on the hill – a light that had been dimmed by big government and its’ tax bill. Ideological leaders seek followers who intrinsically believe in the goals and values that provide a basis for the prescriptive mental model being advocated. As a result, ideological leaders attempt to appeal to a cadre of like-minded followers rather than the masses. This commonality in beliefs and values, of course, promotes shared leadership (Mumford, 2006). Debate centers less on goals and values than the actions that must be taken to induce change in the causes undermining attainment of these goals. In contrast to ideological and charismatic leaders, the prescriptive mental models underlying the actions of pragmatic leaders do not stress goals. For pragmatic leaders, goals are created by objective threats and opportunities evident in the situation at hand. Thus, the prescriptive mental models formulated by pragmatic leaders stress the causes giving rise to these threats and opportunities within the local situation (Mumford & Van Doorn, 2001). Accordingly, the prescriptive mental models constructed by pragmatic leaders tend to be situational based rather than global images. Pragmatic leaders, moreover, see causes as involving both people and situational factors, subject to varying degrees of control, with actions being framed in terms of key controllable causes. This mindset is evident in Dwight D. Eisenhower’s patience in waiting for conditions that would allow him to act to halt McCarthyism. The tendency of pragmatic leaders to build prescriptive mental models around causes, those key controllable causes subject to inﬂuence, implies that pragmatic leaders will prefer logical argumentation to emotionally evocative arguments. These arguments, moreover, will not be framed to appeal to people in general but rather knowledgeable elites who understand, and can induce control, over relevant causes and contingencies. 86 MICHAEL D. MUMFORD ET AL.

The commitment of these elites to the leader, however, will be based on mutual interest rather than personal commitment to the leader.

Prior Research

An ongoing series of studies conducted by Mumford and his colleagues (Mumford, 2006; Mumford, Espejo et al., in press; Mumford & Strange, 2002; Mumford & Van Doorn, 2001; Strange & Mumford, 2002, 2005), have sought to provide some support for the distinctions drawn between charismatic, ideological, and pragmatic leaders. In this first study along these lines, Mumford and colleagues (2006) attempted to demonstrate that the behavioral differences suggested by this model were, in fact, observed across outstanding, historically notable, leaders. He obtained biographies for 120 historical leaders where leaders were assigned, on an a priori basis, to a charismatic, ideological, or pragmatic group. Using the historiometric approach (Simonton, 1990), the ‘‘rise to power’’ and ‘‘in power’’ chapters included in these biographies were content analyzed to assess the expression of behaviors relevant to (1) problem-solving, (2) leader–member exchange, (3) communication strategies, and (4) political tactics. Additionally, the prologue and epilogue chapters included in these biographies were analyzed to obtain information about leader performance (e.g., number of contributions, number of institutions established). As expected, charismatic, ideological, and pragmatic leaders did not differ with respect to overall performance – although pragmatic leaders were more likely to build lasting institutions while charismatic leaders were more likely to initiate mass movements. Charismatic, ideological, and pragmatic leaders did, however, differ with regard to their characteristic behavior. For example, pragmatic leaders relied on expertise in problem-solving while, in contrast, expertise was less important to charismatic leaders. Moreover, pragmatics tended to employ logical appeals in communication and rely on rational influence tactics. Charismatic and ideological leaders, on the other hand, tended not to employ rational influence tactics and logical appeals. Instead, they relied on emotional appeals to followers. Ideological and charismatic leaders, as expected, also differed with respect to leader–member relationships with ideological leaders, but not charismatic leaders, who evidenced shared direction of the group. Some further support for this model has been proved by another historiometric study conducted by Ligon, Hunter, and Mumford (2008). Drawing from the proposition that prescriptive mental models are based in Outstanding Leadership 87 part on reflection on prior life events, they content coded biographies to determine whether charismatic, ideological, and pragmatic leaders differed with respect to exposure to different types of life events. They found that ideological leaders were more likely than charismatic and pragmatic leaders to have been exposed to anchoring events giving rise to strongly held personal goals and values. Charismatic leaders, in contrast, were more likely to have been exposed to turning point events than ideological and pragmatic leaders – events that require adaptation, future- oriented adaptation, to change. A final comparative study conducted by Bedell, Hunter, Angie, and Vert (2006) content analyzed the Machiavellian behaviors of 80 historical leaders who were a priori categorized into pragmatic, ideological, and charismatic groupings. Machiavellian behaviors were rated by means of communications between the leader and their respective subordinates. Results indicated that pragmatic leaders evidenced the greatest number of Machiavellian behaviors, followed by charismatic leaders. Ideological leaders, in contrast, evidenced the least amount of Machiavellianism – a finding that is not particularly surprising given the ideological leader’s strong commitment to a core set of beliefs and values; a set of beliefs they are apparently unwilling to compromise. In addition to these comparative studies, other studies have focused on the characteristics of a particular type of leader. For example, Mumford, Espejo et al. (in press) obtained biographies for 80 leaders who were violent, and nonviolent, ideologues and violent, and nonviolent, charismatic and pragmatic leaders. These biographies were content analyzed for expression of common group, organizational, and environmental characteristics. It was found that ideological leaders differed from charismatic and pragmatic leaders with regard to both just world commitments and ideological extremism or expression of strongly held beliefs and values. In another study along these lines, Mumford and Van Doorn (2001) conducted a qualitative study of the social innovations (e.g., subscription libraries, volunteer fire departments, secular collegiate education) attributed to Benjamin Franklin – a pragmatic leader. An ‘‘in-depth’’ analysis of Franklin’s writings bearing on these initiatives indicated that extensive analysis of social causes was integral to Franklin’s development of these innovations. More centrally, when there was not general agreement about goals, as was the case in the Albany Plan of Union, Franklin failed to exercise the influence that characterized his more successful leadership efforts. In addition to these qualitative and historiometric studies, evidence for the plausibility of this model of the origins of charismatic, ideological, 88 MICHAEL D. MUMFORD ET AL. and pragmatic leadership has been provided by a series of experimental studies. For example, Strange and Mumford (2005), in their study of vision formation, found that viable visions could be constructed by using the causes evident in successful prior models (a charismatic strategy) or by rejecting the goals evident in unsuccessful models (an ideological strategy). In another study along these lines, Bedell, Hunter, and Mumford (in press) developed measures to assess people’s proclivity to a charismatic, ideological, or pragmatic leadership style. Subsequently, she assessed forecasting strategies and found that charismatic, ideological, and pragmatic leaders displayed the expected differences with regard to the causes and goals applied in projecting the outcomes of future events. A final experimental study conducted by Hunter, Bedell-Avers, and Mumford (2009) also assessed participant’s proclivity for a leader type. Researchers then asked participants to engage in a computerized leadership simulation that allowed for the manipulation of multiple contextual factors, including environmental complexity and situational framing. Results of the study indicated that the three leader types performed differently under varying contextual conditions. It should be noted that, although moderated performance effects were observed, no main effect differences were found among the three leader types on any of the study’s criteria. This ‘‘non- finding’’ combined with the situational moderator effects highlight the general notion that the three pathways represent distinct, yet equally viable pathways to outstanding leadership (Mumford, 2006).

MULTI-LEVEL INFLUENCES

If it is granted that at least some support is available for the distinction we have drawn between charismatic, ideological, and pragmatic leadership, a new question comes to the fore: What multi-level variables might operate to shape the emergence and performance of charismatic, ideological, and pragmatic leaders? The question of multi-level inﬂuence is critical to understanding leadership for a number of reasons. The most straightforward is that leadership is an inherently multi-level phenomenon (e.g., Yammarino, Dionne, Chun, & Dansereau, 2005; Yammarino & Dansereau, 2008). Going further, failure to correctly specify the unit of analysis can result in misguided and even erroneous conclusions (e.g., Kozlowski & Klein, 2000; Yammarino, 1996). A series of studies illustrate this concept, as well as the idea that correct conclusions may be drawn if appropriate methods and approaches are used (e.g., Kenny, Mannetti, Outstanding Leadership 89

Pierro, Stefano, & Kashy, 2002; Schriesheim, Castro, & Yammarino, 2000; Yammarino, 1996). In addition, multi-level consideration of a phenomenon may also bring to light issues that were previously unimagined. For example, in their investigation of multi-level innovation, Mumford and Hunter (2005) discovered a series of multi-level paradoxes – where what was necessary for innovation at one level of analysis seemingly contradicted what was required at another. With respect to the structure of the chapter, we will use a multi-level framework employed in a number of prior investigations of multi-level phenomena (e.g., Mumford & Hunter, 2005; Mumford, Bedell et al., 2008; Mumford, Espejo et al., in press). More directly stated, we will consider the variables that might operate to shape the emergence and performance of charismatic, ideological, and pragmatic leaders operating at the individual, group, organization, and environment levels.

Individual-Level Inﬂuences

Leadership, at its most basic level, involves contact and interaction between a given leader and a follower. In this sense, there are a number of individual- level variables that may impact the emergence and performance of charismatic, ideological, and pragmatic leaders, including: (1) leader skills, (2) follower contact, (3) follower identity, and (4) leader commitment. One set of skills relevant to leader emergence and performance is implied by our foregoing discussion with regard to the distinctive characteristics of pragmatic leaders. Earlier, we argued that pragmatic leadership involves the extended analysis of causation in complex systems. This rather straightforward observation has two important implications with regard to leader emergence and performance. First, pragmatic leadership will depend on the acquisition of especially complex mental models describing the causes and consequences of system operations (Mumford, Marcy, Eubanks, & Murphy, in press). In other words, pragmatic leadership will require prior opportunities to acquire requisite expertise and, thus, will evidence some domain speciﬁcity. Second, the performance of pragmatic leaders will depend on the skills needed to analyze and resolve organizational problems vis-a` -vis this expertise. Thus, pragmatic leadership will call for intelligence, creativity, critical thinking, judgment, wisdom, and tacit knowledge (Hedlund et al., 2003; Mumford & Connelly, 1991). Some support for this conclusion may be found in a series of studies by Mumford and his colleagues (Connelly et al., 2000; Mumford, Marks, 90 MICHAEL D. MUMFORD ET AL.

Connelly, Zaccaro, & Reiter-Palmon, 2000; Mumford et al., 2000; Vincent, Decker, & Mumford, 2002; Zaccaro, Mumford, Connelly, Marks, & Gilbert, 2000). They administered an extensive battery of cognitive skill tests to 1,818 army officers – leaders working in an organization known to value pragmatic leadership. They found that indices of leader performance, ranging from critical incident performance to medals awarded, were positively related (R ¼ .40) to the leader’s expression of complex social problem-solving skills such as creative thinking skills, intelligence, expertise, and wisdom. These findings, taken in conjunction with our foregoing observations, imply the following proposition: Proposition 1. At the individual level, the emergence and performance of pragmatic leaders will be more strongly influenced by expertise and cognitive skills than charismatic and ideological leaders, and will, vis-a` -vis expertise requirements, evidence greater domain specificity.

Although charismatic and ideological leadership will make more limited demands on complex problem-solving skills and expertise, charismatic and ideological leaders will still need sufficient expertise to permit the formation of viable prescriptive mental models. The basis for the prescriptive mental models being articulated by charismatic and ideological leaders; however, lies in the identification of shared social goals (Prati, Douglas, & Ferris, 2003). The implication of this statement is that charismatic and ideological leaders must be sensitive to, and aware of, the concerns and motivations of followers. This point, in turn, implies that social intelligence will be more important to the emergence and performance of charismatic and ideological leaders than pragmatic leaders (Zaccaro, Gilbert, Thor, & Mumford, 1991). This point is illustrated in a number of studies including Barbuto and Burbach (2006) who, using a sample of 80 elected officials, found that transformational/charismatic leaders exhibited higher levels of social intelligence. More directly, Wesphal and Stern (2006) found that differing leaders received board appointments via alternative skill pathways. Specifically, the researchers observed that stronger interpersonal skills could ameliorate a lack of expertise (i.e., expertise typically associated with pragmatic leaders), and allowed these leaders to receive high-level promotions. The aggregate of these studies, then, leads to our second proposition. Proposition 2. At the individual level, social skills will exert a stronger influence on the emergence and performance of charismatic and ideological leaders than pragmatic leaders. Outstanding Leadership 91

Along related lines, charismatic and ideological leadership requires sensemaking and sensegiving through articulation of relevant goals. The need for effective communication, in turn, implies that charismatic and ideological leaders must know their audience. In other words, they must minimize psychological, if not physical/structural distance, if they are to perform effectively (Antonakis & Atwater, 2002; Howell & Hall-Merenda, 1999; Shamir, 1995). Thus, interpersonal engagement will prove necessary for charismatic and ideological leaders. In contrast, direct engagement, and the affect unveiled in the course of this engagement, may undermine the kind of complex causal analysis on which pragmatic leadership depends (Mumford, Schultz, & Osburn, 2002). As a result, the emergence and performance of pragmatic leaders will depend on structures that allow them to distance themselves from the immediate concerns of followers and manage follower interactions in such a way as to minimize the need for emotional processing. These observations, in turn, led to our next proposition. Proposition 3. At the individual level, the emergence and performance of charismatic and ideological leaders will require low psychological distance and high contact between leaders and followers while the emergence and performance of pragmatic leaders will require high psychological distance and more limited contact with followers. One of the key variables shaping people’s willingness to follow charismatic and ideological leaders lies in the clear articulation of a prescriptive mental model in terms of a global, goal-oriented, vision (Conger & Kanungo, 1998). The attraction of this vision is that it provides followers with a sense of meaning and identity that promotes affective social engagement (Ashkanasy, 2002; Shamir et al., 1993). One implication of these observations is that charismatic and ideological leaders are most likely to emerge and effectively exercise influence when followers have a weak and fragile sense of identity. Although change, conflict, and social loss can induce a weakened sense of social identity (Beyer, 1999), immaturity, inexperience, and identity diffusion may have similar effects (Dvir & Shamir, 2003; Sy, Tram, & O’Hara, 2006). Under these conditions, the performance of charismatic and ideological leaders will be tied to the quality of the vision statement and the strength of the prescriptive mental model that provides a basis for the vision statement. In contrast, pragmatic leadership does not depend on the leader creating a sense of identity in followers. This point is of some importance because it implies that pragmatic leaders are most likely to arise in autonomous, 92 MICHAEL D. MUMFORD ET AL. high achieving, populations where identity is provided by the individual activities of followers and threat is low – a notion illustrated in a study by Sy et al. (2006) who found that stable, socially aware employees did not need a leader that was cognizant of their emotional needs; they could perform well and were satisfied with their jobs independent of a leader who provided a sense of support and identity. These origins of pragmatic leadership are noteworthy, in part, because they suggest that the performance of pragmatic leaders may depend as much on the quality and capabilities of followers as on the prescriptive mental model being articulated (Collinson, 2006). Taken together, these observations suggest the following two propositions.

Proposition 4. At the individual level, charismatic and ideological leaders will emerge in vulnerable populations while pragmatic leaders will emerge in autonomous high achieving populations.

Proposition 5. At the individual level, the performance of charismatic and ideological leaders will depend on the quality of the prescriptive mental model underlying the vision being articulated while the performance of pragmatic leaders will depend on the skills and capabilities of the followers they recruit as much as the prescriptive mental model of the leader. Although charismatic and ideological leaders both rely on followers investing in a vision, it is important to recognize the differences evident in the prescriptive mental model underlying the visions being articulated by charismatic and ideological leaders. Charismatic leader’s prescriptive mental models are based on the feasibility of attaining future goals through collective action. Thus, charismatic leadership depends on the articulation of opportunity. When positive collective action is not possible, however, the removal of threats to traditional patterns of behavior, the logic provided by ideological leaders through their prescriptive mental models, will prove more attractive. Along related lines, one would expect that charismatic leaders will perform effectively in the pursuit of opportunities while ideological leaders will perform more effectively when salient emerging threats loom in the environment.

Proposition 6. At the individual level, charismatic leaders are more likely to emerge and perform effectively when the pursuit of opportunities is crucial while ideological leaders are more likely to emerge and perform effectively when the removal of threats is crucial. Outstanding Leadership 93

Not only do the conditions that give rise to charismatic and ideological leaders differ, vis-a` -vis the key characteristics of their prescriptive mental models, the characteristics leaders must evidence to allow for the effective exercise of influence will also differ. Charismatic leadership will generally benefit from openness, displays of optimism, and manifest confidence on the part of the leader. Although these dispositional characteristics may also contribute to ideological leadership, under conditions of threat peoples’ range of information processing tends to narrow (Jackson & Dutton, 1988)– focusing on the leader and his/her behavior with respect to perceived threats. This observation is noteworthy because it suggests that followers will attend more to the behavior of ideological leaders and respond more negatively to incongruent, non-value consistent, behavior. As a result, ideological leaders, as opposed to charismatic leaders, must evidence greater conformity to group values and a greater willingness to make personal sacrifices for these values (Yorges et al., 1999) – point illustrated by Bedell et al. (2006) who found that ideological leaders, in comparison to charismatic and pragmatic leaders, were the more strongly committed to their values and were unwilling to compromise their beliefs for personal attainment or gain. Accordingly, the following proposition seems indicated. Proposition 7. At the individual level, self-sacrifice and sustained personal commitment to a vision will be more important to the emergence and performance of ideological leaders than charismatic and pragmatic leaders. In sum, we proposed that a number of variables operating at the individual level impact the emergence and performance of charismatic, ideological, and pragmatic leaders. Broadly, we suggest that pragmatic leaders will emerge and largely rely on their expertise to lead, while ideological and charismatic leaders will, in general, emphasize social skills and emotional appeals when interacting with followers. Such observations have implications for other individual-level factors, including physical and psychological distance, where strong social skills would seem to play a larger and more substantive role in leader success. Along related lines, pragmatic leaders appear more likely to excel in populations that respond well to logical, rational appeals. Finally, the differing aspects of the leaders’ mental models would appear to impact the responses of subordinates in varying ways. For example, the ideological leaders’ emphasis on a core value or set of values would seem to suggest that sacrifice for those ideals would be more critical to the emergence and performance of this leader type. The propositions generated highlight the dynamic interplay between leader, follower, and context, with the congruence of all three being necessary for high-level achievement. 94 MICHAEL D. MUMFORD ET AL.

Group-Level Inﬂuences

Our foregoing observations with regard to self-sacrifice and sustained personal commitment on the part of leaders point to a variety of potential group-level influences on the emergence and performance of charismatic, ideological, and pragmatic leaders. Although a variety of group-level variables might influence leader emergence and performance, four variables appear especially noteworthy: (1) trust, (2) cohesion, (3) interdependence, and (4) exchange relationships. Although trust has been conceptualized in different ways by different investigators (e.g., Pillutla, Malhotra, & Murnighan, 2003; Rousseau, Sitkin, Burt, & Camerer, 1998), it can be argued that trust arises from both interpersonal appraisals and the fairness of exchange relationships. In fact, trust appears integral to the emergence and performance of ideological leaders. Ideological leaders appear to emerge in response to perceived unfairness and victimization of a group (Moghaddam, 2005). Thus, Mumford et al. (2008) found that incidents of other’s corruption were often associated with the emergence of ideological leaders. The emergence of ideological groups in reaction to conditions where trust is low; however, requires ideological leaders to evidence unusually high levels of interpersonal trustworthiness if they are to exercise effective influence over followers. Proposition 8. At the group level, ideological leaders will emerge under conditions where trust is low; however, the performance of ideological leaders will require high levels of interpersonal trust in the leader. Interpersonal trust also appears integral to both the emergence and performance of charismatic leaders. Thus, Pillai, Williams, Lowe, and Jung (2003) in a study of presidential voting found that trust mediated the relationship between attributed charisma and perceived voting patterns. Along similar lines, Bommer, Rubin, and Baldwin (2004) found that cynicism within groups tended to undermine transformational leadership behavior. When these findings are considered in light of the fact that future- oriented visions, by virtue of their intangibility, require placing trust in the leader, the following proposition seems indicated. Proposition 9. At the group level, the emergence and performance of charismatic leaders will depend on high levels of trust. In contrast to charismatic and ideological leaders, interpersonal trust does not appear to be integral to the emergence and performance of pragmatic Outstanding Leadership 95 leaders. Pragmatic leadership, however, does require people to work together for shared interests (Mumford & Van Doorn, 2001). The basis of interaction in shared interest, in turn, implies that the emergence and performance of pragmatic leaders will be related more to perceptions of equity and justice than interpersonal trust. Hence, the emergence and performance of pragmatic leaders will be conditioned by perceptions of process, procedural, and distributive justice – with distributive justice proving a particularly important influence on performance. Proposition 10. At the group level, perceptions of process, procedural, and distributive justice will prove more important to the emergence and performance of pragmatic leaders than interpersonal trust. In addition to trust, cohesion also seems to influence the emergence and performance of charismatic, ideological, and pragmatic leaders. The importance of cohesion to the emergence and performance of ideological leaders is illustrated in a recent study by Mumford et al. (2008). They examined the characteristics of the groups giving rise to ideological, charismatic, and pragmatic leaders. Qualitative material describing group processes was content coded for a variety of variables including group exclusivity, strong group affect, and peer group influence. In a subsequent discriminant analysis contrasting groups, it was found that groups led by ideologues differed from groups led by charismatics and pragmatics with respect to group insecurity and oppositional bonding (actions taken to differentiate the group from other groups). Thus, it seems reasonable to conclude that high levels of group cohesion, cohesion that may in fact be created by the leader through techniques such as oppositional bonding, are necessary for the emergence and performance of ideological leaders. In the case of charismatic and pragmatic leaders, however, cohesion appears to operate in a somewhat more complex fashion. Charismatic leaders, by articulating a future-oriented vision that defines the group and provides a basis for action, will require cohesion for the effective exercise of influence through people. This point is well illustrated by Sanders and Schyns (2006) who found positive relationships between charismatic behaviors and group consensus. However, charismatic leaders emerge, in part, because the vision being articulated provides a basis for cohesion – an observation suggesting that low levels of group cohesion may promote the emergence of charismatic leaders. On the contrary, because pragmatic leaders cannot function effectively when there is no consensus about shared goals and interests, the emergence of pragmatic leaders will depend on at 96 MICHAEL D. MUMFORD ET AL. least a minimal level of cohesion in the group. These observations about cohesion, in turn, suggest the following three propositions. Proposition 11. At the group level, the emergence and performance of ideological leaders will depend on high levels of group cohesion. Proposition 12. At the group level, the emergence of charismatic leaders will be linked to low levels of cohesion while the performance of charismatic leaders will depend on the leader creating high levels of cohesion. Proposition 13. At the group level, the emergence of pragmatic leaders will depend on some minimal level of cohesion but the performance of pragmatic leaders will not depend on high levels of cohesion. Cohesion, of course, proves beneficial to group performance when the level of interdependence in the activities of group members is high (Zaccaro, Gualtieri, & Minionis, 1995). Interdependence also calls for shared mental models (Day, Gronn, & Salas, 2004). Thus, charismatic and ideological leaders may prove more likely to emerge under conditions where interdependence is substantial by virtue of their articulation of a common vision. Moreover, the interdependence of group members may promote the spread and acceptance of this model through social reinforcement (Klein & House, 1998) while the institutional frameworks arising from this interdependence will promote effective group action in relation to the vision being articulated by the leader (Jacobsen & House, 2001). Although interdependence will facilitate the emergence and performance of both charismatic and ideological leaders, interdependence will prove less important to pragmatic leaders. In fact, it seems reasonable to suggest that when group members evidence low levels of interdependence and high levels of autonomy, the ability of pragmatic leaders to orchestrate action through planning and the analysis of causation will prove especially beneficial. Proposition 14. At the group level, high levels of interdependence will contribute to the emergence and performance of charismatic and ideological leaders while low levels of interdependence will contribute to the emergence and performance of pragmatic leaders. A final group-level variable that appears of note in understanding the emergence and performance of charismatic, ideological, and pragmatic leaders may be found in the nature and quality of leaders’ exchange relationships with followers. Not only is there reason to suspect that the nature of exchange relationships differ across leader types but also that these Outstanding Leadership 97 exchange relationships will have differential effects on leader emergence and performance. Thus, Strange (2004) in contrasting charismatic, ideological, and pragmatic leaders with respect to variables characterizing interactions with lieutenants, or close followers, found that charismatic, ideological, and pragmatic leaders differed with respect to follower interaction. More specifically, charismatic leaders, but not pragmatic leaders, were likely to have relationships characterized by a mutual exchange of support. This finding suggests that joint commitment to a vision is critical for charismatic leadership. Ideological leaders, but not charismatic leaders, however, were more likely to share direction of the group with key lieutenants. Shared leadership is likely to prove effective, and contribute to the performance of ideological leaders, due to the foundation of ideological groups in shared beliefs and values. For charismatic leaders, however, the central role played by the leader in articulating his/her personal vision may effectively prohibit shared leadership. Proposition 15. At the group level, the emergence and performance of ideological but not charismatic leaders will be influenced by shared leadership. To summarize, there appear to be a number of noteworthy and interesting variables operating at the group level that would appear likely to impact the performance and emergence of the three leader types. Because group-level phenomena typically involve the coordination and interaction among multiple individuals, variables such as cohesion and interdependence are of notable impact to the performance of the differing leader types. For example, pragmatic leaders would appear to perform well in autonomous, high achieving, and independent populations. In contrast, interdependence, cohesion, and shared leadership are more relevant to the success of ideological and charismatic leaders, albeit with a few notable differences. These observations seem to highlight the challenges leaders face when attempting to coordinate and interact with collections of individuals. Moreover, available evidence on the three leader types appears to indicate that charismatic, pragmatic, and ideological leaders approach group-level tasks in qualitatively different ways and may perform differently as the context and follower composition varies.

Organizational-Level Inﬂuences

Not only will group-level variables influence the emergence and performance of charismatic, ideological, and pragmatic leaders, it appears that a 98 MICHAEL D. MUMFORD ET AL. number of variables operating at the organizational-level will also exert some noteworthy effects in this regard. Although a host of variables have been used to describe organizations, five variables appear especially noteworthy in shaping the emergence and performance of charismatic, ideological, and pragmatic leaders: (1) order, (2) complexity, (3) professionalism, (4) politics, and (5) culture. The concept of order, or its inverse chaos, derives from recent applications of complexity theory in attempts to account for organizational behavior (e.g., Bradbury & Lichtenstein, 2000; Colbert, 2004). Chaos may be defined as the amount of uncertainty resulting from actions taken on potential causal operatives within an organizational system. In other words, chaotic environments are characterized by shifting and unstable causation. Instability in causal operatives, of course, limits the feasibility and effectiveness of the kind of causal analysis that provides a basis for formation of the prescriptive models commonly applied by pragmatic leaders. Chaotic environments, as a result, will tend to inhibit pragmatic leadership. Under chaotic conditions, however, the future, goal-oriented models constructed by charismatic leaders may prove more effective. A future- oriented vision allows for emergent causes, which are subject to potential exploitation in the service of goal attainment. Charismatic models, moreover, can be readily adjusted to take into account shifts in causation provided that the opportunities being pursued remain stable. In this regard, however, it is important to bear in mind a caveat. In highly chaotic settings, uncertainty may become so pronounced that it becomes effectively impossible to envision a future – any future. These conditions will, of course, limit the effectiveness of charismatic leaders. By the same token, however, in highly chaotic organizational settings, where the future is clouded in uncertainty, shared beliefs and values arising from the past might provide a framework for organizational action. As a result, ideological leadership may prove particularly compelling, and particularly effective, in a highly chaotic setting. These points are highlighted by Hunter (2007), who found that charismatic leaders demonstrated decrements in performance in highly chaotic, vision-framed situations. Ideological leaders, on the other hand, seemed to generally excel in the majority chaotic conditions. Taken as a whole, these observations about chaotic organizations imply the following proposition.

Proposition 16. At the organizational level, emergence and performance will vary as a function of the amount of chaos in organizational Outstanding Leadership 99

operations with pragmatic leaders emerging and performing well in stable settings, charismatic leaders emerging and performing well in ordered settings, and ideological leaders emerging and performing well in highly chaotic settings. The amount of order, or chaos, in organizational operations represents only one variable that might be used to understand the opportunities giving rise to different leader types. Another variable likely to prove of some importance in this regard is the complexity of the organization’s operations. Complexity, as it operates at the organizational level, is a multifaceted construct. Complexity is a function of technology and the nature of the production process, the number and nature of shareholders, and the structure of roles and responsibilities within the organization. Broadly speaking, the available evidence indicates that as the complexity of organizational operations increases, planning becomes a progressively more important inﬂuence on organizational performance (Dean & Sharfman, 1996; Miller & Cardinal, 2004; Mumford et al., 2002). The need for planning, in turn, suggests that complexity will call for pragmatic leaders. However, as complexity increases past a certain point, the effectiveness of planning will decrease. High levels of complexity require overarching directive structures such as those found in the future-oriented visions of charismatic leaders (Jacobsen & House, 2001). Although the future-oriented prescriptive models underlying the vision of charismatic leaders will prove useful in guiding planning, by virtue of their focus on past shared beliefs and values, the visions articulated by ideological leaders will prove of substantially less value in this regard.

Proposition 17. At the organizational level, organizational complexity will facilitate the emergence and performance of pragmatic and charismatic leaders but not ideological leaders. Complex organizational systems, as a result of their dependence on technology and functional specialization, are characterized by a high degree of professionalization (Damanpour, 1991). In professionalized organizations, expertise, vis-a` -vis role demands, provides a basis for decision-making. Professionalized organizations, moreover, typically grant substantial autonomy, or discretion, to those people who possess requisite expertise encouraging development of a professional identity that goes beyond the organization. The focus of experts on their professional identity and the value placed on autonomy with regard to professional decisions suggests that visionary leadership will not garner widespread support in professional 100 MICHAEL D. MUMFORD ET AL. organizations thereby inhibiting the emergence and performance of charismatic and ideological leaders. In keeping with this observation, studies of innovation (Jaussi & Dionne, 2003; Kahai, Sosik, & Avolio, 2003) and leadership in research and development organizations indicates that charismatic and transformational leadership often prove ineffective in these settings. Accordingly, these findings suggest the following proposition. Proposition 18. At the organizational level, professionalism will facilitate the emergence and performance of pragmatic leaders and inhibit the emergence and performance of charismatic and ideological leaders. In addition to complexity and professionalism, organizations also differ with respect to the nature of the political behavior that characterizes organizational decision-making and the intensity of people’s involvement in political behavior (Ammeter, Douglas, Gardner, Hochwarter, & Ferris, 2002; Ferris, Adams, Kolodinsky, Hochwarter, & Ammeter 2002). Moreover, there is reason to suspect that the nature and integrity of organizational political behavior will influence leader emergence and performance. Thus, Mumford (2006) found that pragmatic leaders are more likely to apply rational influence tactics, such as resource control and appeals to expertise, than control-based influence tactics such as aggression, coalition building, appeals to status, and personal appeals. In fact, by virtue of others’ commitment to their vision and greater group cohesion, these control-based influence tactics may prove more effective for charismatic and ideological leaders. Charismatic and ideological leaders, moreover, may find these tactics appealing due to the conflict that often surrounds the goals upon which the prescriptive mental models underlying vision-based leadership are founded. These observations imply the following. Proposition 19. At the organizational level, charismatic and ideological leaders will emerge and prove more effective in organizations evidencing high levels of political conflict while pragmatic leaders will emerge and prove more effective in organizations evidencing low levels of political conflict. Political behavior, of course, is influenced by the culture of the organization. Although culture may have a diverse set of effects on the emergence and performance of different leader types, one key aspect of organizational culture permits us to draw some general conclusions about leader emergence and performance. Organizations differ with respect to the strength of their cultures (Ployhart & Schneider, 2002). Variations in cultural strength are noteworthy because in strong cultures it will prove difficult to acquire Outstanding Leadership 101 support for a future-oriented vision inconsistent with current organizational culture. This point is explicitly noted by Tsui, Zhang, Wang, Xin, and Wu (2006) who, in their investigation of CEO leadership in the Republic of China discovered two distinct leader types which may, in this context, be referred to as charismatic and ideological leaders. The researchers noted that the charismatic leaders, termed ‘‘performance builders’’ were ‘‘found in weak cultures because of their exclusive focus on external adaptation priorities’’ (p. 120). In contrast, the ideological leaders, referred to as ‘‘institution builders,’’ were ‘‘associated with strong cultures because of their attention to internal integration (institutionalizing the values and systems of the processes)’’ (p. 120). Thus, it appears that strong culture may act to undermine charismatic leadership, and by the same token, however, the tendency of ideological leaders to anchor their vision in culture should allow ideological leaders to exercise unusual influence. Proposition 20. At the organizational level, strong organizational culture will promote the emergence and performance of ideological leaders and inhibit the emergence and performance of charismatic leaders advocating culturally inconsistent visions. In sum, the available evidence on organization-level variables seems to indicate that there are a number of variables that may impact the emergence and influence of charismatic, ideological, and pragmatic leaders. The variables that appear most noteworthy, at this level, are those that directly impact the functioning and operations of the organization. For example, as chaos increases in an organization, it may prove difficult for leaders emphasizing a future-oriented vision to develop and get others to commit to that vision. Thus, under such organizational conditions, pragmatic leaders rather than charismatic leaders may exhibit greater performance. Other influences include organizational culture, where ideological leaders, given their ability to construct their vision anchored to and existing within a given culture may perform particularly well. In general, the organizational variables considered seem to underscore the importance of congruence between a given leader and their operating environment. Failure to do so, it seems, may lead to potentially negative outcomes.

Environmental-Level Inﬂuences

Our foregoing observations concerning organizational culture and politics, of course, suggest that broader environmental variables will also influence 102 MICHAEL D. MUMFORD ET AL. the emergence and performance of charismatic, ideological, and pragmatic leaders. Although a number of environmental variables might exert some effects in this regard, four environmental-level variables appear especially important: (1) culture, (2) social disruption, (3) turbulence, and (4) elites. Organizations, of course, exist within a broader cultural context. As one might suspect, the existence of a strong, tradition-oriented culture appears to promote the emergence and acceptance of ideological leadership (Post et al., 2002). Culture, however, is a multifaceted phenomenon. Thus, it is possible, and indeed likely, that certain cultural values will promote the emergence of certain types of leaders. In fact, Hofstede’s (1983) distinction between individualistic and collectivist cultures suggests a relatively straightforward hypothesis bearing on cultural conditions and leader emergence. More specifically, by virtue of their emphasis on individual autonomy and the individual’s pursuit of their own future, individualistic cultures may be especially prone to produce pragmatic leaders. Illustrating this point, Choi and Yoon (2005) found that participants in an individualistic culture were more responsive to a leader’s individual expertise and competence than those in the collectivist culture. In contrast, collectivist cultures, due to their emphasis on shared social traditions, may be especially prone to produce ideological leaders (Kuhnen, Hannover, & Shubert, 2001). Proposition 21. At the environmental level, the emergence of ideological leaders will be facilitated by collectivist cultures while the emergence of charismatic and pragmatic leaders will be facilitated by individualistic cultures. Not only will broader cultural trends influence leader emergence, and at least potentially leader performance, conditions where little agreement exists about the future course of society, and extant social systems have failed, will prove difficult for charismatic and pragmatic leaders. Charis- matic leadership will be undermined by disagreement about goals and a negative appraisal of the future while pragmatic leadership will be undermined by disorder. However, one might expect these conditions to promote ideological leadership as people look to shared traditional values as a way of rebuilding society. Some support for this conclusion has been provided by Mumford et al. (2008). They obtained book length qualitative descriptions of ideological and nonideological groups that emerged across the 20th century in Western and non-Western societies. The environmental conditions characterizing the society during the development of these groups were assessed with respect to variables such as globalization, loss of institutions, and economic Outstanding Leadership 103 displacement. A discriminant analysis indicated that ideological groups could be distinguished from nonideological groups by three variables: (1) social disruption, (2) social conflict, and (3) institutional replacement. These findings, in turn, suggest. Proposition 22. At the environmental level, social disruption and the failure of extant institutions will promote the emergence of ideological leaders. Of course, turbulence and change in the environment need not always undermine society and extant social institutions. Turbulence and change may also arise from the introduction of new technologies and the emergence of new patterns of social relationships. Although changes in technology and social interaction patterns are more ordered than the conditions giving rise to ideological leaders, they are also associated with the emergence of new opportunities as well as the need to adapt extant institutions. Keck and Tushman (1993), for example, found that turbulence caused by such events as technological discontinuities led to greater team change and heterogeneity; conditions that tend to promote charismatic leadership (Mumford, 2006). In keeping with this observation, Tushman and O’Reilly (1997) found that the survival of organizations under conditions of radical technological change was often associated with the emergence of charismatic leaders articulating a positive vision for adaptive change through their prescriptive mental model. Proposition 23. At the environmental level, conditions of social and technological change will contribute to the emergence and performance of charismatic leaders. Societies, of course, are commonly held to be composed of elites controlling access to resources and key institutions. Elite control, however, is likely to be threatened by turbulence and social disruption. Thus, when societies have strong, stable elites the potential influence of charismatic and ideological leaders will be limited. Elites, moreover, by virtue of their power and autonomy will tend to pursue their own interests. These observations imply that negotiation and the ability of leaders to work through multiple elites will be a critical influence on leader emergence and performance in stable societies – especially stable meritocratic societies. These observations led Mumford and Van Doorn (2001) to argue that the operation of viable elites will give rise to a preference for pragmatic leadership; at least under conditions where there is agreement among elites about relevant goals. When this goal consensus does not exist; however, charismatic leaders, 104 MICHAEL D. MUMFORD ET AL. by bringing about a new consensus with regard to relevant goals, may be preferred. These observations, in turn, led to our final proposition. Proposition 24. At the environmental level, when institutions are subject to viable elite control, pragmatic leaders will emerge and perform effectively under conditions of goal consensus while charismatic leaders will emerge and perform effectively under conditions of goal conflict. To summarize, there appear to be a number of critical environment-level variables that may impact the emergence and performance of the three leader types: including turbulence, social disruption, a broader social culture and the existence of an elite control group. Generally speaking, as turbulence and social disruption increase, followers are likely to seek the guidance of charismatic and ideological leaders, respectively. With respect to cultural issues, available evidence seems to indicate that ideological leaders will emerge in more collectivist cultures, while pragmatic and charismatic leaders will be preferred by those in more individualistic societies. Finally, pragmatic leaders would seem particularly suited to dealing with an elite social group – particularly when those individuals agree on the general pursuit of a common goal. Clearly, leaders do not operate within a vacuum and these propositions, broadly speaking, highlight the importance of considering leadership within the larger environmental context.

DISCUSSION

Before turning to the broader conclusions flowing from the present effort, certain limitations should be noted. To begin, no attempt was made in the present effort to provide a comprehensive description of charismatic, ideological, and pragmatic leadership. In part, the limitation reflects the fact that more complex descriptions of charismatic (e.g., Conger & Kanungo, 1998), ideological (e.g., Strange & Mumford, 2002), and pragmatic (e.g., Mumford & Van Doorn, 2001) leaders are available. More directly, however, this limitation arises from the unique focus of the present effort on the individual, group, organizational, and environmental variables influencing the emergence and performance of charismatic, ideological, and pragmatic leaders. Along related lines, it should be recognized that the present study could not, and in fact has not, examined all of the individual, group, organizational, and environmental variables that might conceivably influence the emergence and performance of charismatic, ideological, and pragmatic Outstanding Leadership 105 leaders. Instead, in the present effort we have focused on variables that account for differential emergence and performance on the part of charismatic, ideological, and pragmatic leaders. Thus, less attention was given to communication skills (e.g., Fiol et al., 1999) than cognitive skill requirements for pragmatic leaders (e.g., Marcy & Mumford, in press; Mumford & Van Doorn, 2001) and social skill requirements for charismatic and ideological leaders (e.g., Zaccaro et al., 1991). Similarly, we chose to engage in a somewhat narrowed multi-level discussion of leadership. For example, we did not discuss issues of aggregation, expressly specify the level at which phenomena should be explored or address issues of data analysis and investigation. Moreover, our discussion of multi-level influences was directed at the emergence and performance of the three leader types, with less discussion on the dynamic interplay between leader and follower as well as other constituencies. Along related lines, we focused our commentary on broad issues impacting leader emergence and performance and did not, with regard to the chapter, consider issues of time frame and shifts in relationships over time (Kozlowski & Klein, 2000). This somewhat narrowed investigation was intentional, however, to allow for a reasonably concise multi-level consideration of the new leadership model. Such focused multi-level discussion, moreover, is somewhat beyond the scope of this largely theoretical piece. These issues stand as important, however, and we hope they will be addressed as more empirical studies are conducted that investigate multi-level influences of charismatic, ideological, and pragmatic leaders. Finally, observations with regard to individual, group, organizational, and environmental influences on differential leader emergence and performance were formulated with respect to a given level of analysis. This within-level formulation is, of course, consistent with the current stage of development of research on charismatic, ideological, and pragmatic leadership – despite the existence of many studies on charismatic leadership systematic studies of ideological and pragmatic leadership have appeared only in the last few years. This approach, while perhaps dictated by the current status of the literature should not, however, be taken to imply that significant cross-level effects do not exist. Indeed, these effects should be examined in future studies exploring charismatic, ideological, and pragmatic leadership. Nonetheless, our observations with regard to multi-level influences on the emergence and performance of charismatic, ideological, and pragmatic leaders do suggest some cross-level phenomena that might be of interest in future work along these lines. For example, it is possible that charismatic 106 MICHAEL D. MUMFORD ET AL. and ideological leaders may at times induce political conflict to create conditions that favor their emergence and performance. In contrast, pragmatic leaders may seek to minimize political conflict due to the detrimental effects of intense conflict on the effective application of complex problem-solving skills. Shared leadership, and the cohesion of ideological groups, may serve to minimize the tendency of followers to scrutinize ideological leaders thereby limiting the potential detrimental effects of follower evaluations on ideological leadership. Ideological leaders, moreover, may create perceptions of victimization and unfairness to engender the threat perceptions that make their prescriptive mental models more attractive to followers. Although other examples of this sort may be cited, the foregoing examples seem sufficient to illustrate the potential value of cross-level studies examining the emergence and performance of charismatic, ideological, and pragmatic leaders. Even bearing these caveats in mind, we believe that our examination of individual, group, organizational, and environmental influences on the emergence and performance of charismatic, ideological, and pragmatic leaders has some noteworthy implications. These implications are summarized in Table 1. To begin, prior studies contrasting charismatic, ideological, and pragmatic leaders have focused primarily on demonstrating the existence of different patterns of behavior across these three types of outstanding leaders (e.g., Mumford, 2006; Strange & Mumford, 2002). In this regard, our observations in the present effort are noteworthy because they suggest not only that charismatic, ideological, and pragmatic leaders differ in their behavior but also in (a) the conditions promoting emergence of charismatic, ideological, and pragmatic leaders, and (b) the conditions that facilitate effective exercise of influence on the part of charismatic, ideological, and pragmatic leaders. Thus, our observations in the present effort provide some further support for the substantive meaningfulness of these three alternative pathways to outstanding leadership. In addition to providing some further evidence pointing to the distinction we have drawn between charismatic, ideological, and pragmatic leadership; our observations provide some support for the general model of outstanding leadership held to give rise to these alternative pathways. More specifically, this model holds that outstanding leadership ultimately arises from sensemaking activities in relation to crises (Hunt et al., 1999). In sensemaking, however, leaders make different assumptions about causation and the relative emphasis to be placed on goals and causes in the formation of the prescriptive mental models that will be applied in sensemaking and vision formation (Mumford, 2006). Outstanding Leadership 107

Table 1. Summary of Hypotheses about Multi-Level Inﬂuences.

Charismatic Ideological Pragmatic

General characteristics of outstanding leaders Crisis, change events present þþþ Sensemaking by leader þþþ Formation of viable mental models þþþ Strength of follower commitment to leader þþ Goals stressed in mental models þþ Causes stressed in mental models þþ Support based on mass reactions þþ Support based on elite reactions þ Individual-level Cognitive skills þ Social skills þþ Amount of follower contact þþ Importance of leader mental model to followers þþ Pursuit of opportunities þ Removal of threats þ Self-sacriﬁce by leader þ Group-level Trust required for emergence þ Trust required for performance þþ Actions taken to maintain perceptions of þ fairness Group cohesion and emergence þþ Group cohesion and performance þþ Level of group interdependence for emergence þþ and performance Amount of leadership shared with group þ Organizational-level Lack of order þþ Amount of complexity þþ Amount of professionalism þ Amount of political conﬂict þþ Strength of organizational culture þ

Environmental-level Collectivist cultures þ Individualistic cultures þþ Social disruption þ Conditions of sociotechnical change þ Consensus among elites þ Lack of elite consensus þ

Note: þ, indicates positive relationship. 108 MICHAEL D. MUMFORD ET AL.

These differential frames for sensemaking are, in fact, consistent with the findings obtained in recent studies examining individual, group, organizational, and environmental influences on the emergence and performance of charismatic, ideological, and pragmatic leaders. For example, social disruption, social conflict, and institutional failure should lead to an emphasis on situational causes, a focus on threats, and a tendency to stress traditional goals and values – all characteristics of the ideological prescriptive mental model. And, consistent with these observations, Mumford et al. (2008) found that these conditions promote the emergence of ideological leaders. Not only does this theory produce hypotheses that have been confirmed in recent studies examining individual, group, organizational, and environmental influences, it also generates hypotheses that seem consistent with much of what we know about the emergence and performance of charismatic, ideological, and pragmatic leaders. For example, the proposition that pragmatic leaders rely on cognitive skills is not only consistent with the prior observations of Mumford and Van Doorn (2001), but it is in keeping with the findings of Marcy and Mumford (in press) indicating that causal analysis contributes to social innovation – a hallmark of pragmatic leaders. Along similar lines, prior studies of charismatic leadership indicate that the formation of cohesive groups dedicated to the future-oriented goals being articulated by charismatic leaders may represent a critical determinant of the ability of charismatic leaders to exercise influence (Klein & House, 1998). Another piece of evidence pointing to the viability of this theory is that it leads to some new hypotheses about the conditions giving rise to the emergence and performance of charismatic, ideological, and pragmatic leaders. For example, chaos and complexity appear to exert a rather complex set of effects on the emergence and performance of these three types of leaders. Chaotic conditions favor ideological leaders while ordered conditions favor pragmatic leaders. Pragmatic leaders, however, along with charismatic leaders, appear more likely to emerge and perform effectively in complex environments. Alternatively, it appears plausible to argue that the viability of the leader’s prescriptive mental model may be more important for charismatic and ideological leaders than pragmatic leaders – a hypothesis that warrants some attention in future studies. A final piece of evidence bearing on the validity of this model arises from its ability to permit differential predictions with regard to leader emergence and performance. Thus, while social disruption and conflict may promote the emergence of ideological leaders, this conflict, when coupled with Outstanding Leadership 109 oppositional bonding at the group level, may act to undermine performance. Similarly, trust may not be necessary for the emergence of ideological leaders, but trust building does appear central to long-term performance. For charismatic leaders, however, trust appears integral to both emergence and performance. Taken as a whole, these observations not only provide support for the validity of the model of outstanding leadership under consideration but also the existence of charismatic, ideological, and pragmatic pathways to outstanding leadership – pathways that emerge from a dynamic interaction of leaders and their environments as they seek to make sense of crises. Although the potential existence of multiple alternative pathways to outstanding leadership is not new (Weber, 1924), it is also true that leadership theory has for some time now focused almost exclusively on one form of outstanding leadership – charismatic leadership. Our focus on charismatic leaders, however, may have resulted in a simplistic, indeed an overly simplistic, image of outstanding leadership. This image results in an assumption that an emotionally evocative future-oriented vision is the way to lead despite the fact that a variety of conditions operating at the individual, group, organizational, and environmental levels condition the emergence and performance of charismatic leaders and result in differential opportunities for charismatic, ideological, and pragmatic leadership. Hopefully, the present effort by illustrating how various conditions operating at different levels influence the differential emergence and performance of charismatic, ideological, and pragmatic leader will set the stage for a new, more sophisticated, set of studies that recognize the complex nature of the different pathways to outstanding leadership.

ACKNOWLEDGMENTS

We would like to thank Allison Antes, Katrina Bedell-Avers, Jill Strange, and Ginamarie Scott-Ligon, for their contributions to the present effort. Parts of this effort were supported by a series of grants from the U. S. Department of Defense, Michael D. Mumford Principal Investigator. Signiﬁcant portions of this article were published in Leadership Quarterly, Vol. 19, Michael D. Mumford, Alison L. Antes, Jay J. Caughron, and Tamara L. Friedrich, Charismatic, ideological, and pragmatic leadership: Multi-level inﬂuences on emergence and performance, pp. 114–160, Copy- right Elsevier 2008. 110 MICHAEL D. MUMFORD ET AL.

REFERENCES

Ammeter, A. P., Douglas, C., Gardner, W. L., Hochwarter, W. A., & Ferris, G. R. (2002). Toward a political theory of leadership. Leadership Quarterly, 13, 751–796. Antonakis, J., & Atwater, L. (2002). Leader distance: A review and a proposed theory. Leadership Quarterly, 13, 673–705. Ashkanasy, N. M. (2002). Studies of cognition and emotion in organizations: Attribution, affective events, emotional intelligence and perception of emotion. Australian Journal of Management, 27, 11–20. Avolio, B. J., Howell, J. M., & Sosik, J. J. (1999). A funny thing happened on the way to the bottom line: Humor as a moderator of leadership style effects. Academy of Management Journal, 42, 219–227. Barbuto, J. E., & Burbach, M. E. (2006). The emotional intelligence of transformational leaders: A field study of elected officials. Journal of Social Psychology, 146, 51–64. Bass, B. M. (1985). Leadership beyond expectations. New York: Free Press. Bass, B. M. (1990). Bass and Stogdill’s handbook of leadership: Theory, research, and managerial applications. New York: Free Press. Bass, B. M., & Avolio, B. J. (1990). The implications of transactional and transformational leadership for individual, team, and organizational development. In: W. Pasmore & R. W. Woodman (Eds), Research in organizational change and development (pp. 231–272). Greenwich, CT: JAI Press. Bedell, K. E., Hunter, S. T., Angie, A. D., & Vert, A. (2006). A historiometric examination of Machiavellianism and a new taxonomy of leadership. Journal of Leadership and Organizational Studies, 12, 50–72. Bedell, K. E., Hunter, S. T., & Mumford, M. D. (2009). Problem-solving tactics of charismatic, ideological and pragmatic leaders: A comparative experimental study. Leadership Quarterly, 19, 89–106. Beyer, J. M. (1999). Taming and promoting charisma to change organizations. Leadership Quarterly, 10, 307–330. Bommer, W. H., Rubin, R. S., & Baldwin, T. T. (2004). Setting the stage for effective leadership: Antecedents of transformational leadership behavior. Leadership Quarterly, 15, 195–210. Bradbury, H., & Lichtenstein, B. M. B. (2000). Relationality in organizational research: Exploring the space between. Organization Science, 11, 551–564. Chiozza, G., & Goemans, H. E. (2003). Peace through insecurity: Tenure and international conflict. Journal of Conflict Resolution, 47, 443–467. Choi, Y., & Yoon, J. (2005). Effects of leaders’ self-sacrificial behavior and competency on followers’ attribution of charismatic leadership among Americans and Koreans. Current Research in Social Psychology, 11, 51–69. Colbert, B. A. (2004). The complex resource-based view: Implications for theory and practice in strategic human resource management. Academy of Management Review, 29, 341–358. Collinson, D. (2006). Rethinking followership: A post-structuralist analysis of follower identities. Leadership Quarterly, 17, 179–189. Conger, J. A. (1999). Charismatic and transformational leadership in organizations: An insider’s perspective on developing streams of research. Leadership Quarterly, 10, 145–180. Outstanding Leadership 111

Conger, J. A., & Kanungo, R. S. (1988). Toward a behavioral theory of charismatic leadership in organizational settings. Academy of Management Review, 12, 637–647. Conger, J. A., & Kanungo, R. S. (1998). Charismatic leadership in organizations. Thousand Oaks, CA: Sage. Connelly, M. S., Gilbert, J. A., Zaccaro, S. J., Threlfall, K. V., Marks, M. A., & Mumford, M. D. (2000). Exploring the relationship of leadership skills and knowledge to leader performance. Leadership Quarterly, 11, 65–86. Damanpour, F. (1991). Organizational innovation: A meta-analysis of effects of determinants and moderators. Academy of Management Journal, 34, 555–590. Day, D. V., Gronn, P., & Salas, E. (2004). Leadership capacity in teams. Leadership Quarterly, 15, 857–880. Dean, J. W., Jr., & Sharfman, M. P. (1996). Does decision process matter? A study of strategic decision-making effectiveness. Academy of Management Journal, 39, 368–396. de Hoogh, A. H. B., den Hartog, D. N., & Koopman, P. L. (2004). Charismatic leadership, environmental dynamism, and performance. European Journal of Work and Organiza- tional Psychology, 13, 447–471. Deluga, R. J. (2001). American president Machiavellianism: Implications for charismatic leadership and rated performance. Leadership Quarterly, 12, 334–363. Drazin, R., Glynn, M. A., & Kazanjain, R. K. (1999). Multi-level theorizing about creativity in organizations: A sensemaking perspective. Academy of Management Review, 24, 286–329. Dumdum, U. R., Lowe, K. B., & Avolio, B. J. (2002). A meta-analysis of transformational and transactional leadership correlates of effectiveness and satisfaction: An update and extension. In: B. J. Avolio & F. J. Yammarino (Eds), Transformational and charismatic leadership: The road ahead (Vol. 2, pp. 35–66). Oxford, UK: Elsevier. Dvir, T., & Shamir, B. (2003). Follower developmental characteristics as predicting transformational leadership: A longitudinal field study. Leadership Quarterly, 14, 327–344. Elenkov, D. S., & Manev, I. M. (2005). Top management leadership and influence on innovation: The role of sociocultural context. Journal of Management, 31, 381–402. Ferris, G. R., Adams, G., Kolodinsky, R. W., Hochwarter, W. A., & Ammeter, A. P. (2002). Perceptions of organizational politics: Theory and research directions. In: F. J. Yammarino & F. Dansereau (Eds), The many faces of multi-level issues (pp. 179–254). Oxford, UK: Elsevier. Fiol, C. M., Harris, D., & House, R. J. (1999). Charismatic leadership: Strategies for effecting social change. Leadership Quarterly, 10, 449–482. Fleishman, E. A. (1953). The description of supervisory behavior. Journal of Applied Psychology, 37, 1–6. Frankwick, G. L., Walker, B. A., & Ward, J. C. (1994). Belief structures in conflict: Mapping a strategic marketing decision. Journal of Business Research, 31, 183–195. Gioia, D. A., & Thomas, J. B. (1996). Institutional identity, image, and issue interpretation: Sensemaking during strategic change in academia. Administrative Science Quarterly, 41, 370–403. Gioia, D. A., Thomas, J. B., Clark, S. M., & Chittipeddi, K. (1994). Symbolism and strategic change in academia: The dynamics of sensemaking and influence. Organization Science, 5, 363–383. Goldstone, R. L., & Sakamoto, Y. (2003). The transfer of abstract principles governing complex adaptive systems. Cognitive Psychology, 46, 414–466. 112 MICHAEL D. MUMFORD ET AL.

Graen, G. B., & Uhl-Bien, M. (1998). Relationship-based approach to leadership: Development of leader-member exchange (LMX) theory of leadership over 25 years: Applying a multi- level multi-domain perspective. In: F. Dansereau & F. J. Yammarino (Eds), Leadership: The multi-level approaches (pp. 103–134). Stamford, CT: JAI Press. Halverson, S. E., Holladay, C. C., Kazma, S. M., & Quinones, M. A. (2004). Self-sacriﬁcial behavior in crisis situations: The competing roles of behavioral and situational factors. Leadership Quarterly, 15, 211–240. Hedlund, J., Forsythe, G. B., Horvath, J. A., Williams, W. M., Snook, S., & Sternberg, R. (2003). Identifying and assessing tacit knowledge: Understanding the practical intelligence of military leaders. Leadership Quarterly, 14, 117–140. Hemlo-Silver, C. E., & Pfeffer, M. G. (2004). Comparing expert and novice understanding of a complex system from the perspective of structures, behaviors, and functions. Cognitive Science, 28, 127–138. Hersey, P., & Blanchard, K. A. (1982). Leadership style: Attitudes and behaviors. Training and Development Journal, 36, 50–52. Hofstede, G. (1983). National cultures revisited. Behavior Science Research, 18, 285–305. House, R. J. (1971). A path-goal theory of leader effectiveness. Administrative Science Quarterly, 16, 321–338. House, R. J. (1977). A 1976 theory of charismatic leadership. In: J. G. Hunt & L. L. Larson (Eds), Leadership: The cutting edge (pp. 189–207). Carbondale, IL: Southern Illinois University Press. House, R. J., & Howell, J. M. (1992). Personality and charismatic leadership. Leadership Quarterly, 3, 81–108. Howell, J. M., & Hall-Merenda, K. E. (1999). The ties that bind: The impact of leader-member exchange, transformational and transactional leadership, and distance on predicting follower performance. Journal of Applied Psychology, 84, 680–694. Hunt, J. G. (1999). Transformation/charismatic leadership’s transformation of the ﬁeld: An historical essay. Leadership Quarterly, 10, 129–144. Hunt, J. G. (2004). Consideration and structure. In: J. M. Burns, G. R. Goethals & L. Sorenson (Eds), Encyclopedia of leadership (pp. 196–204). Great Barrington, MA: Berkshire/Sage. Hunt, J. G., Boal, K. B., & Dodge, G. E. (1999). The effects of visionary and crisis responsive charisma on followers: An experimental examination of two kinds of charismatic leadership. Leadership Quarterly, 10, 423–448. Hunter, S. T. (2007). The impact of situational framing and complexity on charismatic, ideological and pragmatic leaders: Investigation using a Computer Simulation. Dissertation Abstracts International, The University of Oklahoma (Vol. 52, p. 417). Hunter, S. T., Bedell-Avers, K. E., & Mumford, M. D. (2009). Impact of situational framing and complexity on charismatic, ideological and pragmatic leaders: Investigation using a computer simulation. The Leadership Quarterly, 20, 21–45. Jackson, S. E., & Dutton, J. E. (1988). Discerning threats and opportunities. Administrative Science Quarterly, 33, 370–387. Jacobsen, C., & House, R. J. (2001). Dynamics of charismatic leadership: A process theory, simulation model, and tests. Leadership Quarterly, 12, 75–112. Jaussi, K. S., & Dionne, S. D. (2003). Leading for creativity: The role of unconventional behavior. Leadership Quarterly, 14, 351–368. Johnson-Laird, P. N. (1999). Deductive reasoning. Annual Review of Psychology, 50, 109–135. Outstanding Leadership 113

Kahai, S. S., Sosik, J. J., & Avolio, B. J. (2003). Effects of leadership style, anonymity, and rewards on creativity-relevant processes and outcomes in an electronic meeting system context. Leadership Quarterly, 14, 369–376. Keck, S. L., & Tushman, M. L. (1993). Environmental and organizational context and executive team structure. Academy of Management Journal, 36, 1314–1344. Kenny, D. A., Mannetti, L., Pierro, A., Stefano, L., & Kashy, D. A. (2002). The statistical analysis of data from small groups. Journal of Personality and Social Psychology, 83, 126–137. Kirkpatrick, S., & Locke, E. A. (1996). Direct and indirect effects of three core charismatic leadership components on performance and attitudes. Journal of Applied Psychology, 81, 36–51. Klein, K. J., & House, R. J. (1998). On fire: Charismatic leadership and levels of analysis. In: F. Dansereau & F. J. Yammarino (Eds), Leadership: The multiple level approach (pp. 2–33). Stamford, CT: JAI Press. Kozlowski, S. W., & Klein, K. J. (2000). A multilevel approach to theory and research in organizations: Contextual, temporal and emergent processes. In: K. Klein & S. Kozlowski’s (Eds), Multilevel theory, research and methods in organizations. San Francisco, CA: Jossey-Bass. Kuhnen, U., Hannover, B., & Shubert, B. (2001). The semantic-procedural interface model of the self: The role of self-knowledge for context-dependent versus context-independent modes of thinking. Journal of Personality and Social Psychology, 80, 397–409. Ligon, G. S., Hunter, S. T., & Mumford, M. D. (2008). Development of outstanding leadership: A life narrative approach. Leadership Quarterly, 19, 312–334. Lowe, K. B., Kroeck, K. G., & Sivasubramaniam, N. (1996). Effectiveness correlates of transformational and transactional leadership: A meta-analytic review of the MLQ literature. Leadership Quarterly, 7, 385–425. Lowe, P. (2001). The leadership strategies of top CEOs. Success, 48, 22. Marcy, R. T., & Mumford, M. D. (in press). Social innovation: Enhancing creative performance through causal analysis. Creativity Research Journal. Marta, S., Leritz, L. E., & Mumford, M. D. (2005). Leadership skills and group performance: Situational demands, behavioral requirements, and planning. Leadership Quarterly, 16, 97–120. Merolla, J. L., Ramos, J. M., & Zechmeister, E. J. (2007). Crisis, charisma and consequences: Evidence from the 2004 Presidential election. Journal of Politics, 69, 1–30. Miller, C. C., & Cardinal, L. B. (2004). Strategic planning an firm performance: A synthesis of more than two decades of research. Academy of Management Journal, 37, 1649–1666. Moghaddam, F. M. (2005). The staircase to terrorism: A psychological exploration. American Psychologist, 60, 161–169. Mumford, M. D. (2006). Pathways to outstanding leadership: A comparative analysis of charismatic, ideological, and pragmatic leaders. Mahwah, NJ: Erlbaum. Mumford, M. D., Bedell, K. E., Hunter, S. T., Espejo, J., Eubanks, D., & Connelly, M. S. (2008). Violence in ideological and non-ideological groups: A quantitative analysis of qualitative data. Journal of Applied Social Psychology, 38, 1521–1561. Mumford, M. D., & Connelly, M. S. (1991). Leaders as creators: Leader performance and problem solving in ill-defined domains. Leadership Quarterly, 2, 289–315. Mumford, M.D., Espejo, J., Hunter, S. T., Bedell, K. E., Eubanks, D. L. & Connelly, S. (in press). The sources of leader violence: A multi-level comparison of ideological and non-ideological leaders. The Leadership Quarterly. 114 MICHAEL D. MUMFORD ET AL.

Mumford, M. D., & Hunter, S. T. (2005). Innovation in organizations: A multi-level perspective on creativity. In: F. J. Yammarino & F. Dansereau (Eds), Research in multi-level issues (Vol. IV). Oxford, England: Elsevier. Mumford, M. D., Marcy, R. T., Eubanks, D. L., & Murphy, S. T. (in press). Leader cognition in complex systems: The identification and manipulation of causes. Leadership Quarterly. Mumford, M. D., Marks, M. A., Connelly, M. S., Zaccaro, S. J., & Reiter-Palmon, R. (2000). Development of leadership skills: Experience and timing. Leadership Quarterly, 11, 87–114. Mumford, M. D., Schultz, R. A., & Osburn, H. K. (2002). Planning in organizations: Performance as a multi-level phenomenon. In: F. J. Yammarino & F. Dansereau (Eds), Research in multi-level issues: The many faces of multi-level issues (pp. 3–36). Oxford, UK: Elsevier. Mumford, M. D., Scott, G. M., Gaddis, B., & Strange, J. M. (2002). Leading creative people: Orchestrating expertise and relationships. Leadership Quarterly, 13, 705–750. Mumford, M. D., & Strange, J. M. (2002). Vision and mental models: The case of charismatic and ideological leadership. In: B. J. Avolio & F. J. Yammarino (Eds), Charismatic and transformational leadership: The road ahead (pp. 109–142). Oxford, UK: Elsevier. Mumford, M. D., & Van Doorn, J. R. (2001). The leadership of pragmatism: Reconsidering Franklin in the age of charisma. Leadership Quarterly, 12, 274–309. Mumford, M. D., Zaccaro, S. J., Johnson, J. F., Diana, M., Gilbert, J. A., & Threlfall, K. V. (2000). Patterns of leader characteristics: Implications for performance and development. Leadership Quarterly, 11, 115–133. Parry, K. W., & Proctor-Thomson, S. B. (2003). Leadership, culture and performance: The case of the New Zealand public sector. Journal of Change Management, 3, 376–399. Pillai, R., & Meindl, J. R. (1998). Context and charisma: A ‘meso’ level examination of the relationship of organic structure, collectivism, and crisis to charismatic leadership. Journal of Management, 24, 643–671. Pillai, R., Williams, E. A., Lowe, K. B., & Jung, D. L. (2003). Personality, transformational leadership, trust, and the 2000 US presidential vote. Leadership Quarterly, 14, 161–192. Pillutla, M. M., Malhotra, D., & Murnighan, J. K. (2003). Attributions of trust and the calculus of reciprocity. Journal of Experimental Social Psychology, 39, 448–456. Ployhart, R. E., & Schneider, B. (2002). A multi-level perspective on personnel selection research and practice: Implications for selection system design, assessment, and construct validation. In: F. J. Yammarino & F. Dansereau (Eds), The many faces of multi-level issues (pp. 95–140). Oxford, UK: Elsevier. Post, J. M., Ruby, K. G., & Shaw, E. D. (2002). The radical group in context: I. An integrated framework for the analysis of group risk for terrorism. Studies in Conflict and Terrorism, 25, 73–100. Prati, L. M., Douglas, C., & Ferris, G. R. (2003). Emotional intelligence, leadership effectiveness and team outcomes. International Journal of Organizational Analysis, 11, 21–40. Pratt, M. G. (2000). The good, the bad, and the ambivalent: Managing identification among Amway distributors. Administrative Science Quarterly, 45, 456–493. Rousseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. (1998). Not so different after all: A cross-discipline view of trust. Academy of Management Review, 23, 393–404. Sanders, K., & Schyns, B. (2006). Leadership and solidarity behavior: Consensus in perception of employees within teams. Personnel Review, 35, 538–556. Outstanding Leadership 115

Schriesheim, C. A., Castro, S. L., & Yammarino, F. J. (2000). Investigating contingencies: An examination of the impact of span of supervision and upward controllingness on leader-member exchange using traditional multivariate within and between entities analysis. Journal of Applied Psychology, 85, 659–677. Shamir, B. (1995). Social distance and charisma: Theoretical notes and an exploratory study. Leadership Quarterly, 6, 19–47. Shamir, B., House, R. J., & Arthur, M. B. (1993). The motivational effects of charismatic leadership: A self-concept based theory. Organizational Science, 4, 577–594. Schlesinger, A. M. (1997). Rating the presidents: Washington to Clinton. Political Science Quarterly, 112, 179–190. Simonton, D. K. (1990). Psychology, science, and history: An introduction to historiometry. New Haven, CT: Yale University Press. Sosik, J. J., Kahai, S. S., & Avolio, B. J. (1999). Leadership style, anonymity, and creativity in group decision support systems: The mediating role of optimal flow. Journal of Creative Behavior, 33, 227–256. Strange, J. M. (2004). Exceptional leaders and their followers: An in-depth look at leader-follower interactions. Dissertation Abstracts International (Vol. 65, Issue 1-B, p. 470) (UMI No. AAT 3120032). Strange, J. M., & Mumford, M. D. (2002). The origins of vision: Charismatic versus ideological leadership. Leadership Quarterly, 13, 343–377. Strange, J. M., & Mumford, M. D. (2005). The origins of vision: Effects of reflection, models and analysis. Leadership Quarterly, 16, 121–148. Sy, T., Tram, S., & O’Hara, L. A. (2006). Relation of employee and manager emotional intelligence to job satisfaction and performance. Journal of Vocational Behavior, 68, 461–473. Tsui, A. S., Zhang, Z., Wang, H., Xin, K. R., & Wu, J. B. (2006). Unpacking the relationship between CEO leadership behavior and organizational culture. Leadership Quarterly, 17, 113–137. Tushman, M. L., & O’Reilly, C. A. (1997). Winning through innovation: A practical guide to leading organizational change and renewal. Cambridge, MA: Harvard Business School Press. Vincent, A. S., Decker, B. P., & Mumford, M. D. (2002). Divergent thinking, intelligence, and expertise: A test of alternative models. Creativity Research Journal, 14, 163–178. Weber, M. (1924). The theory of social and economic organizations. New York: Free Press. Weick, K. E. (1995). Sensemaking in organizations. Thousand Oaks, CA: Sage. Wesphal, J. D., & Stern, I. (2006). The other pathway to the boardroom: Interpersonal influence behavior as a substitute for elite credentials and majority status in obtaining board appointments. Administrative Science Quarterly, 51, 169–204. Yammarino, F. J. (1996). Group leadership: A levels of analysis perspective. In: M. A. West (Ed.), The handbook of work group psychology (pp. 189–224). Chichester, England: Wiley. Yammarino, F. J., & Dansereau, F. (2008). Multi-level nature of and multi-level approaches to leadership. Leadership Quarterly, 19, 135–141. Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state of science review. Leadership Quarterly, 16, 879–919. Yorges, S. L., Weiss, H. M., & Strickland, O. J. (1999). The effects of leader outcomes on influence, attributions, and perceptions of charisma. Journal of Applied Psychology, 84, 428–436. 116 MICHAEL D. MUMFORD ET AL.

Yukl, G. (1999). An evaluation of conceptual weaknesses in transformational and charismatic leadership theories. Leadership Quarterly, 10, 285–305. Yukl, G. (2006). Leadership in organizations. Upper Saddle River, NJ: Prentice-Hall. Zaccaro, S. J., Gilbert, J., Thor, K. K., & Mumford, M. D. (1991). Leadership and social intelligence: Linking social perceptiveness and behavioral ﬂexibility to leader effectiveness. Leadership Quarterly, 2, 317–331. Zaccaro, S. J., Gualtieri, J., & Minionis, D. (1995). Task cohesion as a facilitator of team decision making under temporal urgency. Military Psychology, 7, 77–93. Zaccaro, S. J., Mumford, M. D., Connelly, M. S., Marks, M. A., & Gilbert, J. A. (2000). Assessment of leader problem-solving capabilities. Leadership Quarterly, 11, 37–64.

LEVELS OF PERFORMANCE: MULTI-LEVEL PERSPECTIVES ON OUTSTANDING LEADERSHIP

James G. (Jerry) Huntw and John N. Davis

ABSTRACT

Mumford, Hunter, Friedrich, and Caughron (2009) consider and evaluate conditions that may inﬂuence how charismatic, ideological, and pragmatic leaders emerge and perform. In particular, they look at differing conditions that operate at the individual, group, organizational, and environmental levels, and suggest how conditions at these multiple levels of observation may drive how each of the three types of outstanding leaders emerges and performs in society. This commentary considers how scholars might use their work to make predictions about outstanding leadership, and which conditions might be ideal for the emergence of each of the three types of outstanding leadership.

INTRODUCTION Mumford, Hunter, Friedrich, and Caughron (2009) consider and evaluate conditions that may inﬂuence how charismatic, ideological, and pragmatic leaders emerge and perform. They note that recent theories of outstanding

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 117–122 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008005 117 118 JAMES G. (JERRY) HUNT AND JOHN N. DAVIS leadership have emphasized charisma, but, continuing a stream of research by Mumford and colleagues (Mumford, 2006), they also consider how outstanding leadership may emerge from ideological and pragmatic types of leadership. In particular, they look at differing conditions that operate at the individual, group, organizational, and environmental levels. Mumford et al. then suggest how conditions at these multiple levels of observation may drive how each of the three types of outstanding leaders emerges and performs in society. Prior to their consideration of the multi-level variables that might shape the emergence and performance of the three types of leaders, Mumford et al. describe outstanding leadership. In this section of their chapter, they consider the impact of crisis and mental models; they define the characteristics of charismatic, ideological, and pragmatic leadership; and they review prior research in support of the distinctions among these three types of leadership. These authors claim that the emergence of all three types of outstanding leadership requires some sort of crisis or major change, some amount of sense making by the leader, and the formation of viable mental models by the leader. They also discuss other general characteristics of outstanding leaders. In a section on multi-level influences, Mumford et al. consider ‘‘which multi-level variables might operate to shape the emergence and performance of charismatic, ideological, and pragmatic leaders.’’ In this, the heart of their multi-level analysis of outstanding leadership, they review the relevant literature and develop 24 propositions. Table 1 in their chapter summarizes these influences, dividing them into general characteristics of outstanding leaders as well as influences at the individual, group, organization, and environmental levels. We were particularly intrigued by the summary of information in Table 1 in Mumford et al.’s chapter and want to consider how one might use it to make predictions about the three types of outstanding leadership, including how they operate at each level. The remainder of this commentary is organized as follows. We first consider the general characteristics of outstanding leadership. We then discuss, in turn, predictions that might flow from conditions that might be ideal for the emergence of each of the three types of leadership. We conclude, as is usual, with suggestions for future research.

GENERAL CHARACTERISTICS OF OUTSTANDING LEADERSHIP

Our look at Table 1 of Mumford et al. (2009) first impressed us with the fact that every type of outstanding leadership requires, as noted earlier, some sort Multi-Level Perspectives on Outstanding Leadership 119 of crisis or major change, some amount of sense making by the leader, and the formation of viable mental models by the leader. We think these requirements are significant in that Mumford et al. have specified necessary, but not sufficient, conditions for the emergence of all three types of outstanding leadership. Likewise, we note some interesting parallels. Specifically, these three hypotheses parallel the traditional perspectives on leadership of traits, behaviors, and contingencies. That is, outstanding leadership would require (1) the trait (or capability) of sense making, (2) the behavior of forming viable mental models, and (3) a situation of crisis or radical change. Much like the poor, these three perspectives seem to be with us always. In addition, Table 1 of Mumford et al. (2009) summarizes 34 hypotheses about influences on the three types of leadership at the individual, group, organization, and environmental levels of analysis. Twelve of these hypotheses are proposed to have positive relationships with only one of the three types of outstanding leadership. We found this linkage to be of interest, and we wondered what might be predicted by stacking hypotheses that apply to only one type of outstanding leadership, thereby creating what one might call an ideal context in which each type of leadership might be most expected to emerge and perform.

UNIQUE CHARACTERISTICS OF OUTSTANDING CHARISMATIC LEADERSHIP

Three of Mumford et al.’s 34 hypotheses are proposed to have a positive relationship only with charismatic leadership. These are (1) at the group level, trust required for performance; (2) at the environmental level, conditions of sociotechnical change; and (3) at the environmental level, lack of elite consensus. No hypotheses that are proposed to have a positive relationship only with charismatic leadership are provided: (1) about general characteristics of outstanding leaders, (2) at the individual level, or (3) at the organizational level. Thus, all other things being equal, the strongest case for the emergence and performance of outstanding charismatic leadership would seem to be when trust is present at the group level (at the time of emergence), and under conditions of sociotechnical change and lack of elite consensus at the environmental level. When might these conditions prevail? Suppose that, in a crisis, a leader emerges who has the skill of sense making and has the trust of group members based on previously demonstrated 120 JAMES G. (JERRY) HUNT AND JOHN N. DAVIS successes. In an environment of social and technological change when elites present conflicting goals, that leader’s vision of a consistent, desirable end state could, indeed, prove compelling to others in the group. We can easily imagine such a leader being described as charismatic both by followers and by researchers. Nevertheless, it seems counterintuitive that there are no general characteristics of only outstanding charismatic leaders. We would have thought that the more than two decades of research that have transformed the field of leadership studies (Hunt, 1999) would have produced at least one. Similarly, we would have expected that studies of both close and distant charisma (e.g., Shamir, 1995) would have identified conditions unique to outstanding charismatic leadership at either the individual or organizational level of analysis.

UNIQUE CHARACTERISTICS OF OUTSTANDING IDEOLOGICAL LEADERSHIP

Four of the 34 hypotheses are proposed to have a positive relationship only with ideological leadership. These are (1) at the group level, the amount of leadership shared with the group; (2) at the organizational level, strength of the organizational culture; (3) at the environmental level, collectivist cultures; and (4) at the environmental level, social disruption. No hypotheses that are proposed to have a positive relationship only with ideological leadership are provided: (1) about general characteristics of outstanding leaders or (2) at the individual level. Thus, all other things being equal, the strongest case for the emergence and performance of ideological leadership would seem to be when leadership can be shared at the group level, when culture is strong at the organizational level, and in collectivist cultures facing social disruption at the environmental level. When might these conditions prevail? Suppose that, in a crisis, a leader emerges who has the skill of sense making, is willing to share leadership within his or her group, and is a member of an organization with a strong culture in an environment of social disruption in a collectivist culture. Suppose further that this leader presents a vision of a return to solutions that have proven themselves valuable to the leader and his or her followers in the past. We can easily imagine these followers responding to this vision. Indeed, we can imagine like-minded followers coming out of the woodwork in response to what they see as responsible, commonsense goals that can (in their minds) resolve the crisis. Multi-Level Perspectives on Outstanding Leadership 121

While these followers might not necessarily refer to the leader as an ideological leader, their view of the leader as bringing good sense to the situation certainly ﬁts the ideological leader as described by Mumford et al. We again ﬁnd it somewhat counterintuitive that there are no general characteristics of outstanding ideological leaders alone. Similarly, the lack of a hypothesis unique to the individual level surprises us. We might have expected one or more unique responses at the individual level to a strong ideological appeal.

UNIQUE CHARACTERISTICS OF OUTSTANDING PRAGMATIC LEADERSHIP

Five of Mumford et al.’s 34 hypotheses are proposed to have a positive relationship only with pragmatic leadership. These are (1) as a general characteristic of outstanding leadership, support based on elite reactions; (2) at the individual level, cognitive skills; (3) at the group level, actions taken to maintain perceptions of fairness; (4) at the organizational level, amount of professionalism; and (5) at the environmental level, consensus among elites. Thus, hypotheses outlining a positive relationship only with pragmatic leadership are proposed as a general characteristic of outstanding leaders and at each of the four levels of analysis under consideration. Thus, all other things being equal, the strongest case for the emergence and performance of pragmatic leadership would seem to be when support is based on elite reactions, when the leader possesses strong cognitive skills at the individual level, and when the leader takes action to maintain perceptions of fairness at the group level; the amount of follower professionalism at the organization level, and consensus among elites at the environmental level. When might these conditions prevail? Suppose that, in a crisis, a leader of high individual intelligence, with the support of elites, promotes perceptions of fairness in an organization of professionals in an environment of elite consensus. We can easily imagine this as fertile ground for outstanding pragmatic leadership. We were intrigued by the fact that hypotheses unique to pragmatic leadership were presented as general characteristics of outstanding pragmatic leadership as well as at the individual, group, organization, and environmental levels. We also wonder if there is something inherent in outstanding pragmatic leadership that requires unique multi-level perspectives, whereas outstanding charismatic and ideological leadership apparently do not. 122 JAMES G. (JERRY) HUNT AND JOHN N. DAVIS

CONCLUSION

The specification of necessary, but not sufficient, conditions for the emergence and performance of outstanding leadership strikes us as a valuable contribution to the literature. Our first suggestion would be to continue this stream of research to see how well this claim is supported. Similarly, we wonder if this line of research could eventually lead to the specification of conditions sufficient for outstanding leadership. This would be a significant contribution indeed. In addition, we would be intrigued to see studies that compared any of the theoretically ideal situations for the emergence and performance of outstanding charismatic, ideological, or pragmatic leadership to less than ideal conditions. Findings that supported the emergence and performance of outstanding leadership in the ideal conditions would lend support to Mumford and his colleagues’ (2006, 2009) research.

REFERENCES

Hunt, J. G. (1999). Transformation/charismatic leadership’s transformation of the ﬁeld: An historical essay. Leadership Quarterly, 10, 129–144. Mumford, M. D. (2006). Pathways to outstanding leadership: A comparative analysis of charismatic, ideological, and pragmatic leaders. Mahwah, NJ: Lawrence Erlbaum. Mumford, M. D., Hunter, S. T., Friedrich, T. L., & Caughron, J. J. (2009). Charismatic, ideological, and pragmatic leadership: An examination of multi-level inﬂuences on emergence and performance. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Shamir, B. (1995). Social distance and charisma: Theoretical notes and an exploratory study. Leadership Quarterly, 6, 19–47.

PRESIDENTIAL LEADERSHIP STYLES: HOW DO THEY MAP ONTO CHARISMATIC, IDEOLOGICAL, AND PRAGMATIC LEADERSHIP?

Dean Keith Simonton

ABSTRACT

Mumford, Hunter, Friedrich, and Caughron (2009) discuss at length three generic types of extraordinary leadership: charismatic, ideological, and pragmatic. This commentary raises the question of whether this general framework applies to more focused domains of leadership. More speciﬁcally, the author discusses his own research on leadership styles in the U.S. presidency – interpersonal, charismatic, deliberative, creative, and neurotic – and then examines whether these ﬁve styles have some correspondence to the three broad types of extraordinary leadership.

INTRODUCTION Mumford, Hunter, Friedrich, and Caughron’s (2009) work represents an impressive scholarly achievement. These authors provide a comprehensive treatment of three major forms of outstanding leadership. Besides deﬁning

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 123–133 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008006 123 124 DEAN KEITH SIMONTON the nature of these three forms, the authors articulate how leader emergence and performance are differentially contingent on variables operating at individual, group, organizational, and environmental levels. Mumford et al. also present their integration at the highest level of abstraction so that it would presumably be applicable to any domain of historically notable leadership, such as political, military, religious, and entrepreneurial leadership. This inclusiveness sets their treatment apart from other typologies. For instance, Ludwig’s (2002) sixfold typology – monarchs, tyrants, visionaries, authoritarians, traditionalists, and democrats – is applicable only to political leadership. Of course, it is useful to inquire about the extent to which abstract models apply well to concrete domains of exceptional leadership. This is where my own work comes in. For more than 30 years, I have been conducting empirical and theoretical research on the determinants of history-making leadership. Some of these investigations have concerned military leaders (e.g., Simonton, 1980) and others hereditary monarchs (for a review, see Simonton, 2001). Nevertheless, the vast majority of these studies have focused on presidents of the United States (for a review, see Simonton, 2008). As part of this line of research, I have examined hundreds of individual and situational factors that influence presidential performance as gauged by a diversity of subjective and objective criteria (Simonton, 1987, 1993). Of the tremendous wealth of empirical findings, one subset has special relevance to the target article – namely, my work on styles of presidential leadership. This commentary first provides an overview of these results and then discusses how these styles might fit with the three forms of leadership proposed in Mumford et al.’s chapter.

PRESIDENTIAL STYLE

In 1986, I published an article on presidential personalities in which I extracted quantitative information from character sketches compiled from biographical sources (Simonton, 1986). These sketches had all identifying material removed and were placed in random order before they were subjected to analyses by several independent raters. Further analyses yielded factor scores on 14 personality dimensions for 39 presidents from George Washington to Ronald Reagan. Two years later, I used these same sketches to derive measures of presidential style (Simonton, 1988), in this case taking advantage of an inventory that had been developed at the Institute for Personality Assessment and Research (IPAR) at the Berkeley campus of the University Presidential Leadership Styles 125 of California (Historical Figures Assessment Collaborative, 1977). After seven blind and independent raters applied this style scale to the sketches, a factor analysis was executed for those items that displayed the highest reliability coefficients (49 items out of the initial 82). The result was five stylistic factors: interpersonal, charismatic, deliberative, creative, and neurotic (listed in order of variance explained in the items). The remainder of this section describes these five styles and indicates some of the personality traits and performance criteria associated with each (for the biographical experiences associated with each style, see Simonton, 1988).

Interpersonal Style

Interpersonal style is evident when the president ‘‘allows Cabinet members considerable independence,’’ ‘‘encourages the exercise of independent judgment by aides,’’ ‘‘gives credit to others for work done,’’ ‘‘endears himself to staff through his courtesy and consideration,’’ ‘‘is flexible,’’ ‘‘emphasizes teamwork,’’ ‘‘is frequently in contact with his advisers and Cabinet,’’ ‘‘maintains close relationships with wide circle of associates,’’ is ‘‘willing to make compromises,’’ ‘‘relies on working in a staff system, deciding among options formulated by advisers,’’ ‘‘keeps members of his staff informed on matters concerning other departments,’’ ‘‘knows his limitations,’’ and sometimes even ‘‘permits himself to be outflanked.’’ In contrast, such a chief executive seldom ‘‘accepts recommendations of others only under protest,’’ rarely ‘‘believes he knows what is best for the people,’’ and is much less likely to be ‘‘emphatic in asserting his judgments,’’ to be ‘‘suspicious of reformers,’’ or to be ‘‘impatient, abrupt in conference.’’ He is also less prone to base ‘‘decisions on willfulness, nervousness, and egotism’’ or to ‘‘force decisions to be made prematurely.’’ In short, such presidents ‘‘work well with others.’’ Interestingly, Woodrow Wilson and Richard Nixon scored especially low on this style. In terms of personality traits, presidents with an interpersonal style tend to be pleasant, good-natured, easy-going, friendly, flexible, moderate, conservative, and nonbelligerent, but not forceful, petty, distrustful, or Machiavellian (i.e., sly, deceitful, unscrupulous, evasive, shrewd, greedy but not sincere or honest). Interpersonal leaders in the White House also tend to be more physically attractive. On the ‘‘Big Five’’ personality dimensions, the interpersonal style is positively correlated with agreeableness but negatively correlated with neuroticism (see Table 1). The interpersonal style is also negatively associated with interpersonal dominance – the desire to exert 126 DEAN KEITH SIMONTON

Table 1. Correlations between the Big Five Personality Factors and the Five Presidential Styles.

Factor Presidential Style

Interpersonal Charismatic Deliberate Creative Neurotic

Extraversion 0.11 0.55** 0.47** 0.12 0.10 Agreeableness 0.68*** 0.40* 0.20 0.49** 0.17 Conscientiousness 0.07 0.26 0.40* 0.12 0.08 Openness 0.10 0.06 0.39* 0.21 0.03 Neuroticism 0.42* 0.20 0.07 0.16 0.12

Notes: Based on presidential style scores published in Simonton (1988) and Big Five factor scores published in Rubenzer and Faschingbauer (2004) for the 29 presidents that were shared between the two studies. These correlations are published here for the ﬁrst time. *po0.05, **po0.01, ***po0.002. control over others in personal relationships. All in all, this style is the most amiable from the perspective of face-to-face interaction. Turning our attention to the criteria of leader performance, presidents exhibiting the interpersonal style are likely to experience fewer Cabinet resignations, indicating an ability to get along with their appointees. Such presidents also tend to get along better with Congress: Interpersonal presidents use the veto power less often, have fewer of their vetoes over- turned, and have fewer of their nominees to the Cabinet or Supreme Court rejected by the Senate. Lastly, although interpersonal presidents are not more or less likely to go down in history as great presidents, their standing with posterity is much less controversial. Such presidents neither awe nor alienate.

Charismatic Style

A president operating under the charismatic style is a leader who ‘‘finds dealing with the press challenging and enjoyable,’’ ‘‘enjoys the ceremonial aspects of the office,’’ ‘‘is charismatic,’’ ‘‘consciously refines his own public image,’’ ‘‘has a flair for the dramatic,’’ ‘‘conveys clear-cut, highly visible personality,’’ is a ‘‘skilled and self-confident negotiator,’’ ‘‘uses rhetoric effectively,’’ is a ‘‘dynamo of energy and determination,’’ is ‘‘characterized as a world figure,’’ ‘‘keeps in contact with the American public and its moods,’’ ‘‘has ability to maintain popularity,’’ ‘‘exhibits artistry in Presidential Leadership Styles 127 manipulation,’’ and ‘‘views the presidency as a vehicle for self-expression.’’ By comparison, he is rarely ‘‘shy, awkward in public.’’ The highest scorers on this style were Andrew Jackson and Franklin Roosevelt. It is worth pointing out that presidents who score high on this style have been shown to display a distinctive rhetoric in their speeches (Emrich, Brower, Feldman, & Garland, 2001). The rhetoric is actually comparable to what is found in great poetry! With respect to personality traits, charismatic presidents are prone to be outgoing, natural, and witty, rather than shy or withdrawn. They are also inclined to be intellectually brilliant, Machiavellian, forceful, immoderate, and progressive rather than conservative. In addition, the charismatic style is linked with elevated scores on two basic motives: the need for power and the need for achievement. Charisma also correlates with two Big Five factors: positively with extraversion and negatively with agreeableness (see Table 1). Hence, charismatic presidents are more dominant than friendly. In fact, high charisma tends to go with high interpersonal dominance: These are pushy people. Turning to performance criteria, charisma is positively correlated with the number of significant acts passed, specific legislative victories, and special messages sent to Congress. The charismatic chief executive also likes to promulgate more executive orders to exert his will without needing the approval of Congress. Most significantly, charismatic presidents are great presidents, at least according to posthumous assessments. Besides tending to receive higher overall ratings, charisma has a positive correlation with ratings of prestige, strength, activeness, and accomplishment.

Deliberative Style

Deliberative style describes that chief executive who ‘‘understands implications of his decisions; exhibits depth of comprehension,’’ is ‘‘able to visualize alternatives and weigh long-term consequences,’’ ‘‘keeps himself thoroughly informed; reads brieﬁngs, background reports,’’ and is ‘‘cautious, conservative in action.’’ At the same time, such a president infrequently – if ever – ‘‘indulges in emotional outbursts.’’ This style was most characteristic of George Washington. The deliberative style has a positive correlation with the interpersonal style. An example of a president who was low on both is Andrew Jackson. Deliberative presidents are organized, insightful, polished, methodical, intelligent, sophisticated, ﬂexible, and moderate to an almost passive degree. 128 DEAN KEITH SIMONTON

With regard to the Big Five personality factors, these chief executives score high on conscientiousness and openness but low on extraversion (see Table 1). Similar to interpersonal presidents, deliberative chief executives are less likely to suffer numerous Cabinet resignations. In this case, the low turnover may indicate superior wisdom in making the appointments in the first place. The deliberative style is also similar to the interpersonal style in that such presidents tend to veto fewer bills sent to their desk. Otherwise, this style appears much less successful. The deliberative president has fewer bills passed by Congress, including fewer really important bills, and tends to have a shorter administration – he is more like a transient caretaker. The only plus is that such leaders have lower odds of having major scandals dishonor their administrations. Clearly, in this area of performance their deliberative nature pays off; they do not appoint anyone to high office who has no business being there. Strangely, when we look at posthumous greatness assessments, the leadership performance of deliberative presidents is no better or no worse than the norm. Perhaps they are so risk adverse that they are unwilling to take the risks necessary for bona fide greatness – yet they avoid out-and-out failure en route to leader mediocrity.

Creative Style

A creative orientation is associated with a leader who regularly ‘‘initiates new legislation and programs’’ and who ‘‘is innovative in his role as an executive.’’ Yet he is very seldom a ‘‘middle-of-the-roader.’’ Thomas Jefferson, Andrew Jackson, and Franklin Roosevelt are counted among the best exemplars of this style. The creative style often coexists with the charismatic style; many presidents are high (or low) on both together. The creative style is correlated with being inventive and artistic – with pervasive intellectual brilliance – as well as being Machiavellian, inflexible, immoderate, forceful, aggressive, and progressive rather than conservative. Furthermore, the creative style is associated with higher scores on the need for power and the need for achievement. Although the charismatic style is also connected with these two motives, charisma is more strongly correlated with power motivation while creativity is more strongly correlated with achievement motivation. Nonetheless, the creative style is even more conspicuously linked with interpersonal dominance than is the charismatic style. Perhaps as a consequence, the creative style has a somewhat more negative correlation with the Big Five agreeableness dimension than does Presidential Leadership Styles 129 the charismatic style (see Table 1). It would seem that creative presidents are so concerned with getting where they want to go that they are more willing to step on a lot of toes than are charismatic presidents. The latter have much less interest in alienating people to get the job done. Like the charismatic style, the creative style is positively correlated with the number of significant acts passed, specific legislative victories, and special messages sent to Congress. However, creativity – unlike charisma – is more predictive of reelection to a second term, and hence serving more years in office. Even more important are the relations with expert evaluations of ultimate presidential performance. Like charismatic chief executives, creative leaders earn higher posthumous ratings, including higher assessments on specific criteria of prestige, strength, activeness, and accomplishment. Nevertheless, the correlations are almost always higher for the creative style relative to the charismatic style. Moreover, charismatic chief executives have a higher likelihood of becoming controversial, of provoking disagreements about the president’s global success. Thus, if one were a voter deciding between a creative candidate and a charismatic candidate, the wiser choice would be the first. Not only are creative presidents better, on average, but they are less likely to be seen by some as worse.

Neurotic Style

The last of the five styles is the most peculiar. The neurotic style applies to that president who most frequently ‘‘places political success over effective policy’’ and who ‘‘suffers health problems [that] tend to parallel difficult and critical periods in office.’’ Making matters even worse, presidents with this style are almost never exhibit a ‘‘direct, uncomplicated approach’’ to political leadership. The neurotic style is best exemplified by James Polk, whereas Ronald Reagan’s style was the least neurotic. Presidents featuring the neurotic style tend to exhibit Machiavellian and petty (i.e., greedy and self-pitying) personalities, and to display higher than average levels of achievement (but not power) motivation. These leaders have a ‘‘chip on their shoulder’’ and, therefore, manipulate politics to meet their ego needs. Although the neurotic style is positively correlated with the Big Five factor neuroticism, this correlation is far from statistically significant (see Table 1). This null result may partly reflect the poor reliability of this particular factor. The inferior reliability may also explain why the neurotic style is not a conspicuous predictor of leadership performance. It correlates with only one criterion – the number of failures in getting the 130 DEAN KEITH SIMONTON president’s legislative proposals through Congress. Evidently, neuroticism is not very persuasive on Capitol Hill.

DISCUSSION

Now how do these five styles line up with the three forms of outstanding leadership advanced by Mumford and his coauthors? Let us simplify the analysis by dismissing the neurotic style. High scores on this factor are very rare, and it has very few correlates, especially regarding performance criteria. That leaves the interpersonal, charismatic, deliberative, and creative styles. It might be reasonably argued that these four can be collapsed into just two – to wit, the interpersonal–deliberative style and the charismatic– creative style. This consolidation can be defended on two grounds: The two sets of styles have strong positive correlations. Although there are exceptions, interpersonal leaders tend to be deliberative (r ¼ 0.55) and charismatic leaders tend to be creative (r ¼ 0.62). The two sets tend to correlate the same way with personality traits and performance criteria (and with biographical experiences; Simonton, 1988). To be sure, contrasts do emerge for each set, but these differences may be judged smaller than the affinities. If so, then we might offer two correspondences: Most obviously, charismatic leadership would be associated with some integration of charismatic and creative styles. More tenuously, pragmatic leadership might be identified with some combination of interpersonal and deliberative styles. If we then consolidate these four styles into the two leadership types, we obtain the new set of scores shown in Table 2. The highest-scoring presidents on the charismatic–creative dimension are Franklin Roosevelt and Jackson; the lowest-scoring presidents are Grant, Coolidge, and Taft. In contrast, the highest scorers on the interpersonal–deliberative dimension are Fillmore and Washington, and the lowest scorers are Jackson and Wilson. For good or ill, chief executives since Franklin Roosevelt have not scored more than one standard deviation above or below the mean on either of the two collapsed dimensions. The attractions of this unification notwithstanding, we are still left with an irksome question: What happened to ideological leadership? One possibility is that the original IPAR presidential style inventory did not Presidential Leadership Styles 131

Table 2. Standardized Scores Combining Charismatic–Creative and Interpersonal–Deliberative Styles.

President Charismatic–Creative Interpersonal–Deliberative

Washington 0.04 1.72 J. Adams 0.15 0.23 Jefferson 0.84 1.08 Madison 0.61 1.47 Monroe 0.66 0.72 J. Q. Adams 0.44 0.09 Jackson 2.18 2.12 Van Buren 0.53 0.26 W. Harrison 1.26 0.30 Tyler 0.52 1.36 Polk 0.24 0.89 Taylor 0.15 0.46 Fillmore 1.35 2.12 Pierce 0.32 0.27 Buchanan 0.68 0.44 Lincoln 0.38 0.93 A. Johnson 0.37 1.19 Grant 1.86 1.64 Hayes 0.94 0.65 Garﬁeld 0.62 0.44 Arthur 0.69 0.13 Cleveland 0.92 0.07 B. Harrison 0.60 0.31 McKinley 0.30 1.12 T. Roosevelt 0.64 1.23 Taft 1.84 0.76 Wilson 0.21 2.07 Harding 1.08 0.24 Coolidge 1.86 0.08 Hoover 0.01 0.51 F. Roosevelt 2.16 0.89 Truman 0.45 1.22 Eisenhower 0.21 0.19 Kennedy 1.27 1.14 L. Johnson 1.51 0.53 Nixon 0.94 1.15 Ford 0.32 1.06 Carter 0.14 0.48 Reagan 1.24 0.01 G. H. W. Bush 0.06 0.47 Clinton 1.51 0.69

Note: All scores are published here for the first time. Scores for Bush and Clinton are based on imputed scores for the two presidents based on the Big Five factor personality scores in Rubenzer and Faschingbauer (2004). Accordingly, these must be considered more tentative than the others. In any case, the two combined factors have a correlation of 0.33 (po0.05), indicating a slight antagonism between the two leadership types. Kennedy is the only chief executive who was more than one standard deviation above the mean on both styles. 132 DEAN KEITH SIMONTON contain items tapping this aspect of presidential leadership. Unfortunately, a glance at the items indicates that this may be the case (Simonton, 1988, Table 1). Only two items seem relevant to capturing this dimension: ‘‘uncompromising in matters of personal principle’’ and ‘‘is idealistic.’’ Worse still, neither of these items entered the factor analysis. Their reliability coefficients were too low. Accordingly, it was impossible to tap variation in ideological leadership among the 39 U.S. presidents sampled. Even so, these same deficiencies can be interpreted in a more positive fashion. Perhaps the inventory omitted lots of items concerning ideological leadership because that form of leadership is not a prominent feature of the U.S. presidency. Consequently, there is too little variation on that dimension. This interpretation would also explain the low reliabilities of the items that do seem relevant to ideological leadership. If the sampled presidents varied very little on these traits, then the variance is not large enough to support a conspicuous consensus. Thus U.S. presidents may really come in just two types: charismatic (charismatic and creative styles) and pragmatic (interpersonal and deliberative styles). Admittedly, it is possible to extract from historians’ ratings (published in Maranell, 1970) a bipolar dogmatism factor that features idealistic- inflexibility at one pole and pragmatic-flexibility at the other pole (Simonton, 1986, 1987). Moreover, posthumously assessed presidential performance is a U-shaped function of this dogmatism assessment. The greatest presidents are either idealistically inflexible or pragmatically flexible! However, this finding does not solve the problem because the dogmatism factor places ideological and pragmatic presidents on opposite poles of a single dimension. Hence, a pure ideological type of leadership remains elusive among those who have assumed the highest political office in the United States. Naturally, some observers might like to brand this or that president as an ideologue. In truth, such ideological leaders are very rare, if nonexistent. Perhaps George W. Bush came closest, yet he was probably a far cry from a genuine ideological leader from either right or left (but see Simonton, 2006). Because the U.S. system of electing presidents requires that each candidate seek the largest electorate, and because voters as a whole tend to place themselves in the middle of the political spectrum, a forthright ideologue may be unelectable – as libertarian Republican Barry Goldwater learned in the 1964 presidential election. Presidents of the United States may really come in only two alternative flavors: charismatic (and creative) or pragmatic (interpersonal and deliberative). Whether this binary decision is an asset or a deficit of American democracy, I let others determine. Presidential Leadership Styles 133

REFERENCES

Emrich, C. G., Brower, H. H., Feldman, J. M., & Garland, H. (2001). Images in words: Presidential rhetoric, charisma, and greatness. Administrative Science Quarterly, 46, 527–557. Historical Figures Assessment Collaborative. (1977). Assessing historical figures: The use of observer-based personality descriptions. Historical Methods Newsletter, 10, 66–76. Ludwig, A. M. (2002). King of the mountain: The nature of political leadership. Lexington, KY: University Press of Kentucky. Maranell, G. M. (1970). The evaluation of presidents: An extension of the Schlesinger polls. Journal of American History, 57, 104–113. Mumford, M. D., Hunter, S. T., Friedrich, T. L., & Caughron, J. J. (2009). Charismatic, ideological, and pragmatic leadership: An examination of multi-level influences on emergence and performance. In: F. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Rubenzer, S. J., & Faschingbauer, T. R. (2004). Personality, character, and leadership in the White House: Psychologists assess the presidents. Washington, DC: Brassey’s. Simonton, D. K. (1980). Land battles, generals, and armies: Individual and situational determinants of victory and casualties. Journal of Personality and Social Psychology, 38, 110–119. Simonton, D. K. (1986). Presidential personality: Biographical use of the Gough Adjective Check List. Journal of Personality and Social Psychology, 51, 149–160. Simonton, D. K. (1987). Why presidents succeed: A political psychology of leadership. New Haven, CT: Yale University Press. Simonton, D. K. (1988). Presidential style: Personality, biography, and performance. Journal of Personality and Social Psychology, 55, 928–936. Simonton, D. K. (1993). Putting the best leaders in the White House: Personality, policy, and performance. Political Psychology, 14, 537–548. Simonton, D. K. (2001). Kings, queens, and sultans: Empirical studies of political leadership in European hereditary monarchies. In: O. Feldman & L. O. Valenty (Eds), Profiling political leaders: Cross-cultural studies of personality and behavior (pp. 97–110). Westport, CT: Praeger. Simonton, D. K. (2006). Presidential IQ, openness, intellectual brilliance, and leadership: Estimates and correlations for 42 US chief executives. Political Psychology, 27, 511–639. Simonton, D. K. (2008). Presidential greatness and its socio-psychological significance: Individual or situation? Performance or attribution? In: C. Hoyt, G. R. Goethals, & D. Forsyth (Eds), Leadership at the crossroads: Psychology and leadership (Vol. 1, pp. 132–148). Westport, CT: Praeger. This page intentionally left blank

CHARISMATIC, IDEOLOGICAL, AND PRAGMATIC LEADERSHIP: WHERE WE ARE, AND WHERE DO WE NEED TO GO?

Michael D. Mumford, Jay J. Caughron and Tamara L. Friedrich

ABSTRACT

Mumford, Friedrich, Hunter, and Caughron (2009) propose a multilevel theory of leadership intended to allow for the emergence and performance of charismatic, ideological, and pragmatic leaders. In their commentaries, Hunt and Davis (2009) and Simonton (2009) provide additional support for this model. Their observations also broach questions about how charismatic, ideological, and pragmatic leadership styles should be measured, and how hypotheses should be developed with regard to multilevel inﬂuence on leader emergence and performance. This commentary discusses the implications of these observations for future research on styles of outstanding leadership.

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 135–143 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008007 135 136 MICHAEL D. MUMFORD ET AL.

INTRODUCTION

Leadership comes in many forms (Bass, in press; Yukl, 2005). As a result, most scholars conceive of leadership as a domain-specific phenomenon. For example, one might focus on the leadership of creative ventures (Mumford, Scott, Gaddis, & Strange, 2002), leadership under conditions of complexity (Uhl-Bien, Marion, & McKelvey, 2007) or ethical leadership (Trevino & Brown, 2004). Of the many domains in which one might study leadership, scholars often stress the importance of outstanding, historically notable, leadership (Bass, in press) – that is, the leaders who change the world in which we live. Although there has long been an interest in outstanding leadership, little real substantive progress was made in advancing this field until some 20 years ago, when theories of charismatic leadership emerged (Conger & Kanungo, 1998; House, 1995; Sharmin, House, & Arthur, 1993). Broadly speaking, research on charismatic leadership served to demonstrate the importance of vision, vision articulation, and follower infusion of identity in the vision being articulated by the leader (Kirkpatrick & Locke, 1996). Although we would not dispute the importance of this work, by the same token it is clear that charismatic leadership does not always lead to success (Lowe, Kroeck, & Sivasubramaniam, 1996; Tosi, Misangyi, Fanelli, Waldman, & Yammarino, 2004). Even more importantly, it is not clear where vision comes from and why it has such a powerful effect on followers, at least under certain circumstances. These observations were the impetus for our development of the theory of charismatic, ideological, and pragmatic leadership (Mumford, 2006; Mumford, Hunter, Friedrich, & Caughron, 2009; Mumford & Van Doorn, 2001; Strange & Mumford, 2005); this theory is outlined in the chapter to which Hunt and Davis (2009) and Simonton (2009) address their commentaries. In response to the commentaries of Simonton as well as Hunt and Davis, the remainder of this commentary contains three sections. The first section focuses on describing the theory of charismatic, ideological, and pragmatic leadership as proposed by Mumford and colleagues (Mumford, 2006; Mumford et al., 2009). The second section emphasizes issues surrounding measurement. While measurement was not a focal point of the Mumford, Hunter, Friedrich, and Caughron’s chapter, the Simonton review raises some interesting points about measurement that should be recognized and addressed. Finally, the last section addresses the comments of Hunt and Davis as they pertain to the development of hypotheses about and examination of charismatic, ideological, and pragmatic leadership styles. It is Charismatic, Ideological, and Pragmatic Leadership 137 our hope that this response addresses the most substantive issues relating to the current theory under examination to the fullest extent possible given the available research on this recently developed conceptualization of leadership.

THE THEORY

Before turning to the commentaries of Hunt and Davis (2009) and Simonton (2009), it is useful to brieﬂy reconsider the theoretical model being proposed. Hunt and Davis provide a succinct, and accurate, summary of this theory, which holds that (1) outstanding leadership emerges in response to crises; (2) crises – as complex, ambiguous, high-risk events – require sense making for effective leadership; and (3) the basis for sense making lies in the formation of a viable mental model that allows others to understand and respond to the crisis. In fact, the available evidence indicates that this is a plausible basis for understanding the emergence (Hunt, Boal, & Dodge, 1999) and psychological basis (Strange & Mumford, 2005)of outstanding leadership. Although this theory strikes Hunt and Davis as well as Simonton as plausible, the key aspects of this theory raise a number of questions. For example, which characteristics of crises lead people to engage in sense making (Weick, 1995)? What are the origins and key characteristics of leaders’ mental models (Ligon, Hunter, & Mumford, 2008)? How do leaders translate these mental models to followers (Senge, 1990)? Of course, all of these questions need to be far more thoroughly addressed than has occurred up to this point. Indeed, a new wave of research on normative aspects of outstanding leadership is clearly needed. The charismatic, ideological, and pragmatic theory under consideration in the Hunt and Davis (2009) and Simonton (2009) commentaries is a micro- theory within the context of this broader theory. The charismatic, ideological, and pragmatic theory is based on a simple, well-established proposition (Goldvarg & Johnson-Laird, 2001). Mental models are complex cognitive constructions, and different mental models may be constructed by different people to account for the same phenomenon. Thus, we argue that people may formulate mental models with respect to crisis events, different temporal frames, experience, outcomes sought, nature of causes emphasized, agents of causation, and the sources of control in these causes (Mumford, 2006). Preferred patterns in constructing these causal explana- tory systems, in turn, give rise to the three styles of outstanding leadership referred to as charismatic (e.g., future time frame, people as causes, positive 138 MICHAEL D. MUMFORD ET AL. outcomes sought), ideological (e.g., past time frame, situations as causes, negative outcomes to be avoided), and pragmatic (e.g., present time frame, interactive causes, both positive outcomes sought and negative outcomes to be avoided). With regard to the charismatic, ideological, and pragmatic theory proposed by Mumford et al. (2009), four key points should be noted. First, this theory is an inherently cognitive theory. Second, the styles by which people formulate these mental models are not mutually exclusive. More speciﬁcally, different combinations of base elements will give rise to different mental models. Thus, one might view people as causes of negative events in mental model formation – an alternative type of outstanding leadership that warrants consideration. Third, this theory refers to three generic, or prototypical, styles by which leaders construct mental models for understanding crisis-related events. Fourth, these mental models are held to represent stable organizing systems that, although they may not be applied in addressing every issue confronting a leader, develop early in life (Ligon et al., 2008) and are used to describe, explain, and predict (Rouse & Morris, 1986) the effects of crises, and guide interactions with others in addressing these crises. These stable organizing principles, in turn, broach the question addressed by Mumford, Hunter, Friedrich, and Caughron: When does an individual expressing a particular style – charismatic, ideological, or pragmatic – emerge, and when does this style prove effective in crisis resolution?

MEASUREMENT

Simonton (2009), in his commentary, is ultimately examining the applicability of this theory to one area where outstanding leadership is commonly observed – U.S. Presidents. In his study of these leaders, Simonton (1988) obtained behavioral and personality ratings based on sketches provided of their actions. Five styles were identified: (1) interpersonal, (2) charismatic, (3) deliberative, (4) creative, and (5) neurotic. For example, charismatics were characterized by behaviors such as ‘‘uses rhetoric effectively’’ and ‘‘has a flair for the dramatic.’’ Subsequently, Simonton agrees, in part based on correlations observed among those styles, that U.S. Presidents can be characterized by a charismatic or pragmatic style when those five dimensions are collapsed. Thus, bearing in mind the point that ideologues are rarely elected to presidential office, a finding that confirms our arguments with regard to selective emergence of charismatic, ideological, and pragmatic leaders, Simonton’s findings provide some critical support for our proposal Charismatic, Ideological, and Pragmatic Leadership 139 that outstanding leaders may exhibit distinct styles – styles that Mumford, Hunter, Friedrich, and Caughron refer to as charismatic, ideological, and pragmatic. Like most scholars, we applaud any effort that provides support for our theory. When one considers Simonton’s research in the context of the theory sketched out earlier, however, certain questions come to the fore with regard to the measurement of those styles. These questions are of fundamental importance to advancing our understanding of charismatic, ideological, and pragmatic leadership in particular, and perhaps leadership in general. Those issues do not constitute a criticism of Simonton’s work per se, but rather point to a set of broader issues that must be considered. To begin, the theory at hand is ultimately a cognitive theory focusing on how leaders construct different mental models for understanding their world. Those mental models, in turn, affect behavior, interactions, and followers’ perceptions. What should be recognized here is that Simonton (2009) did not assess the mental models applied by leaders, but rather derivative behaviors. This use of second-order measures is not inappropriate, but it does result in certain ambiguities with regard to interpretation. For example, in prior work (Mumford, 2006), Ronald Reagan, based on the mental model he applied, was classified as an ideological leader rather than a charismatic one, as was the case in Simonton’s (2009) study. Thus, in studies seeking to measure charismatic, ideological, and pragmatic leadership, one must carefully consider the specific nature of the measures being applied vis- a-vis this theory. Although we can envision cases where behaviorally based measures might prove appropriate, development of these measures may call for scaling based on differential expression of key behaviors distinguishing these three types. More centrally, for the assessment of leader style, we now generally prefer measures based on direct measurement of the mental model applied by the leader (Marshall-Mies et al., 1996). Recently, Bedell-Avers, Hunter, and Mumford (2008) took an alternative approach to the assessment of the charismatic, ideological, and pragmatic leadership styles. In their work, study participants were presented with an illustration of charismatic, ideological, and pragmatic actions in addressing certain kinds of leadership problems. People were presented with an illustration of each style and asked to indicate the incident most like their own leadership style. The resulting measure evidenced good validity in accounting for leader’s problem-solving performance, with people expressing a certain style – charismatic, ideological, or pragmatic – displaying good performance on problems that matched their style. What is notable with regard to this procedure for measuring leadership style is that it is 140 MICHAEL D. MUMFORD ET AL. consistent with the idiographic nature of mental models. Simonton (2009), of course, applied a normative model. Although it is interesting that a normative approach produced convergent results, the question remains as to whether we should be applying a normative or an idiographic measurement model in style assessments. Also, it must be recognized that these two approaches may not always yield equivalent results. A final issue bearing on assessment of the charismatic, ideological, and pragmatic leadership styles relates to their distribution. The theory on which these styles are based is a cognitive theory that makes no assumptions about the normative distribution of these styles. Rather, Bedell-Avers, Hunter, and Mumford’s (2008) findings indicate that in an undergraduate sample, a pragmatic leadership style is frequently observed (55%), a charismatic style is somewhat frequently observed (30%), and an ideological leadership style is not frequently observed (15%). As Simonton (2009) notes, the relative frequency of style expression must be taken into account in drawing conclusions. Most centrally, measures must be developed that take into account the nonconstant frequency with which these styles are expressed.

HYPOTHESES

Mumford et al. (2009) are not primarily concerned with measurement issues. Rather, they focus on developing a theory about when charismatic, ideological, and pragmatic leaders emerge, and when they perform well, taking into account contextual variables at the individual, group, organizational, and environmental levels that will influence emergence and performance. The Hunt and Davis (2009) commentary directly speaks to this issue, focusing on the propositions proposed. Hunt and Davis focus on the unique propositions that would account for the emergence and performance of charismatic, ideological, and pragmatic leaders. Essentially, they ask a question about differential validity. Broadly speaking, these authors find the evidence acquired for this differential validity of these three styles compelling. That said, we would caution readers with regard to the application of differential validity arguments in part based on the theory at hand and in part based on the status of current research. The theory on which the charismatic, ideological, and pragmatic leadership styles were based is, ultimately, a general theory of outstanding leadership. As noted earlier, this theory holds that all outstanding leaders emerge under conditions of crisis, where the crisis conditions require sense making, and this sense making is held to depend on the viability of the Charismatic, Ideological, and Pragmatic Leadership 141 mental models formulated. Charismatic, ideological, and pragmatic leaders alike emerge under these kinds of crisis conditions. Moreover, their performance is ultimately based on the quality or viability of the mental models used in sense making. Of course, this observation implies that we need studies, far more studies than are currently available, examining key attributes of crises, sense-making activities engaged in at the individual and social levels, and the characteristics of viable mental models. In fact, we believe that studies examining these cross-style variables may prove as important, if not more important, than studies contrasting charismatic, ideological, and pragmatic leadership. We suspect that Hunt and Davis (2009) would agree with this point. However, they argue that by proposing charismatic, ideological, and pragmatic styles, Mumford et al. (2009) essentially have proposed a model in which differential propositions are critical to assessing the validity of the theory. When one reads the theoretical basis for these three styles, we believe this assumption is questionable. As Mumford (2006) has pointed out, these styles differ not in their basic elements, all mental models involve assumptions about causation, but rather in their specific instantiation of these elements (e.g., positive vs. negative causes, people as causes vs. situations as causes) and the organization of these elements in the creation of a mental model. What is noteworthy about this observation is that the mental models of charismatic, ideological, and pragmatic leaders are not viewed as independent entities, but instead are observed as overlapping entities. Thus, both charismatic and pragmatic leaders focus on external causes in formulating their mental models, whereas ideological leaders focus more on internal crises. Similarly, although pragmatic leaders focus on the present, exactly what constitutes ‘‘the present’’ is subject to vagaries in the time frame applied (Jaques, 1976). As a result, pragmatic leaders may, in forming mental models, like charismatics, consider the future. The implication of this observation, of course, is that distinctions between charismatic, ideological, and pragmatic leaders cannot be absolute but rather will be based on a pattern of effects (Mumford, Stokes, & Owens, 1990). This observation, in turn, leads one to question whether differential hypothesis formation and testing will be desirable as Hunt and Davis (2009) would assume. Perhaps even more fundamental to Hunt and Davis’s critique is their argument with regard to the pattern of hypothesis formation. More specifically, unique hypotheses were not formulated with respect to each style of leadership at each level. At one level, this is a fair criticism. On another level, we would caution readers against accepting this argument too 142 MICHAEL D. MUMFORD ET AL. quickly. More specifically, Mumford et al. (2009) formulated hypotheses based on the available evidence to clarify similarities and differences across leader types – charismatic, ideological, and pragmatic. What should be recognized here is that discussion of charismatic, ideological, and pragmatic leadership as a set of distinct styles is a recent event. Often adequate evidence is not available concerning ideological and pragmatic leaders. Although evidence is often available for charismatic leaders, it is rare for studies to have contrasted charismatic leaders with ideological and pragmatic leaders. Although we recognize this inherent limitation of the Mumford, Hunter, Friedrich, and Caughron’s effort, we hope this effort, and the commentaries of Hunt and Davis as well as Simonton will provide the impetus for future studies intended to address this issue.

ACKNOWLEDGMENTS

We would like to thank Sam Hunter and Katrina Bedell-Avers for their contributions to the present effort.

REFERENCES

Bass, B. M. (in press). Handbook of leadership. New York: Free Press. Bedell-Avers, K. E., Hunter, S. T., & Mumford, M. D. (2008). Conditions of problem-solving and the performance of charismatic, ideological, and pragmatic leaders: A comparative experimental study. Leadership Quarterly, 19, 89–106. Conger, J. A., & Kanungo, R. S. (1998). Charismatic leadership in organizations. Thousand Oaks, CA: Sage. Goldvarg, E., & Johnson-Laird, P. N. (2001). Naı¨ve causality: A mental model theory of causal meaning and reasoning. Cognitive Science, 25, 565–610. House, R. J. (1995). Leadership in the twenty-ﬁrst century: A speculative inquiry. In: A. Howard (Ed.), The changing nature of work (pp. 411–450). San Francisco, CA: Jossey- Bass. Hunt, J. G., Boal, K. B., & Dodge, G. E. (1999). The effects of visionary and crisis-responsive charisma on followers: An experimental examination of two kinds of charismatic leadership. Leadership Quarterly, 10, 423–448. Hunt, J. G., & Davis, J. N. (2009). Levels of performance: Multi-level perspectives on outstanding leadership. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Jaques, E. (1976). A general theory of bureaucracy. London: Heinemann. Charismatic, Ideological, and Pragmatic Leadership 143

Kirkpatrick, S. A., & Locke, E. A. (1996). Direct and indirect effects on three core charismatic leadership components on performance and attitudes. Journal of Applied Psychology, 81, 36–51. Ligon, G. M., Hunter, S. T., & Mumford, M. D. (2008). Development of outstanding leadership: A life narrative approach. Leadership Quarterly, 19, 312–334. Lowe, K. B., Kroeck, K. G., & Sivasubramaniam, N. (1996). Effectiveness correlates of transformational and transactional leadership: A meta-analytic review of the MLQ literature. Leadership Quarterly, 7, 385–425. Marshall-Mies, J. C., Martin, J. A., Fleishman, E. A., Zaccaro, S. J., Baughman, W. A., & McGee, M. L. (1996). Development and evaluation of cognitive and metacognitive measures for predicting leadership potential. Bethesda, MD: Management Research Institute. Mumford, M. D. (Ed.) (2006). Pathways to outstanding leadership: A comparative analysis of charismatic, ideological, and pragmatic leadership. Mahwah, NJ: Erlbaum. Mumford, M. D., Hunter, S. T., Friedrich, T. L., & Caughron, J. J. (2009). Charismatic, ideological and pragmatic leadership: An examination of multi-level influences on leader emergence and performance. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Mumford, M. D., Scott, G. M., Gaddis, B., & Strange, J. M. (2002). Leading creative people: Orchestrating expertise and relationships. Leadership Quarterly, 13, 705–750. Mumford, M. D., Stokes, G. S., & Owens, W. A. (1990). Patterns of life history: The ecology of human individuality. Hillsdale, NJ: Erlbaum. Mumford, M. D., & Van Doorn, J. R. (2001). The leadership of pragmatism: Reconsidering Franklin in the age of charisma. Leadership Quarterly, 12, 274–309. Rouse, W. B., & Morris, N. M. (1986). On looking into the black box: Prospects and limits in the search for mental models. Psychological Bulletin, 100, 349–363. Senge, P. M. (1990). The fifth discipline. London: Century. Sharmin, B., House, R. J., & Arthur, M. B. (1993). The motivational effects of charismatic leadership: A self-concept based theory. Organizational Science, 4, 577–594. Simonton, D. K. (1988). Presidential style: Personality, biography, and performance. Journal of Personality and Social Psychology, 55, 928–936. Simonton, D. K. (2009). Presidential leadership styles: How do they map onto charismatic, ideological, and pragmatic leadership? In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi- Level Issues. Bingley, UK: Emerald. Strange, J. M., & Mumford, M. D. (2005). The origins of vision: Effects of reflection, models and analysis. Leadership Quarterly, 16, 121–148. Tosi, H. L., Misangyi, V. F., Fanelli, A., Waldman, D. A., & Yammarino, F. J. (2004). CEO charisma, compensation and firm performance. Leadership Quarterly, 15, 405–420. Trevino, L. K., & Brown, M. E. (2004). Managing to be ethical: Debunking five business ethics myths. Academy of Management Executive, 18, 69–81. Uhl-Bien, M., Marion, R., & McKelvey, B. (2007). Complexity leadership theory: Shifting leadership from the industrial age to the knowledge era. Leadership Quarterly, 18, 298–318. Weick, K. E. (1995). Sensemaking in organizations. Thousand Oaks, CA: Sage. Yukl, G. (2005). Leadership in organizations. Englewood Cliffs, NJ: Prentice-Hall. This page intentionally left blank

PART III LEADERSHIP AND SOCIAL RELATIONS

This page intentionally left blank

A COMPONENTIAL ANALYSIS OF LEADERSHIP USING THE SOCIAL RELATIONS MODEL

David A. Kenny and Stefano Livi

ABSTRACT

The social relations model (SRM; Kenny, 1994) explicitly proposes that leadership simultaneously operates at three levels of analysis: group, dyad, and individual (perceiver and target). With this model, researchers can empirically determine the amount of variance at each level as well as those factors that explain variance at these different levels. This chapter shows how the SRM can be used to address many theoretically important questions in the study of leadership and can be used to advance both the theory of and research in leadership. First, based on analysis of leadership ratings from seven studies, we ﬁnd that there is substantial agreement (i.e., target variance) about who in the group is the leader and little or no reciprocity in the perceptions of leadership. We then consider correlations of leadership perceptions. In one analysis, we examine the correlations between task-oriented and socioemotional leadership. In another analysis, we examine the effect of gender and gender composition on the perception of leadership. We also explore how self-ratings of leadership differ from member perceptions of leadership. Finally, we discuss how the model can be estimated using conventional software.

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 147–191 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008008 147 148 DAVID A. KENNY AND STEFANO LIVI

INTRODUCTION

Leadership is a complex phenomenon that can be explained using a variety of models. In spite of the impressive amount of empirical findings in more than 7,500 articles reviewed by Bass (1990), however, leadership research has not accumulated many consistent results. Part of the explanation for this failure is that scholars have been influenced by periodic fashions in research theory (Chemers, 2000), fluctuating from leader characteristics and perceptual factors of leadership to a more complex view integrating cultural and contextual factors at a wider level of analysis. As a result, leadership research might appear to some as disconnected and directionless (Zaccaro & Klimoski, 2001) with little cross-fertilization of ideas (Van Vugt, 2006). These criticisms have forced leadership researchers to make efforts both to build a theoretical integration of current knowledge (Chemers, 2000)andto find conceptual and methodological instruments able to empirically clarify the relative weight of different approaches at the correct level of analysis (Livi, Kenny, Albright, & Pierro, 2008; Yammarino, Dionne, Chun, & Dansereau, 2005). At least three levels have been postulated to be important for leadership: the individual, the dyad, and the group. (Culture could potentially be viewed as a fourth level, but it is very often at the level of individual when groups contain members of different cultures.) Moreover, the individual level can be conceptualized in two ways: as the target (i.e., who is perceived to be a leader) and as the perceiver (i.e., who is judging leadership). The statistical model considered in this chapter examines leadership at all three of these levels. Before we discuss that model, we review theories of leadership at these levels to enhance understanding of the conceptual, as opposed to statistical, meaning of these different levels. Perhaps, the dominant view is that leadership is the level of the target (the individual). In these theories, leadership depends on a stable characteristic that may be linked to a consistent behavioral disposition, an internal characteristic of the person (i.e., a personality trait), or shared perception about leadership (i.e., a prototype). From this point of view, the oldest scientific approach to creating a psychological explanation of leadership is based on individual differences and is rooted on the nineteenth-century notion of Carlyle’s great man theory of leadership (Chemers, 2000). This theory proposed that the individual characteristics of a leader distinguish leaders from subordinates or an effective leader from inadequate ones, regardless of the situation (group, culture, or organization) or personal relationships with followers. A Componential Analysis of Leadership Using the SRM 149

The most traditional view of the individual differences approach of leadership is based largely on studies that correlate ratings of leadership with the targets’ personality characteristics. The focus on traits had declined after Stogdill (1948), and others reported the unreliable results of reported traits and found that no trait consistently predicted leadership, but interest in the trait approach has since been revived. For example, studies have shown that leadership demonstrates a substantial stability of leadership across situations (Bono & Judge, 2004; Foti & Hauenstein, 2007; Zaccaro, Foti, & Kenny, 1991). Moreover, the meta-analysis by Lord, De Vader, and Alliger (1986) found that some traits do consistently predict leadership. Research on this topic found consistent evidence that some traits are systematically associated with leadership emergence and effectiveness – namely, intelligence, dominance, masculinity, adjustment, conservatism, social sensitivity, self-monitoring, extraversion, and efficacy, among others (Chemers, 2000; Lord et al., 1986; Zaccaro et al., 1991). Scholars continue to have a strong interest on leader personality, as evidenced by the fact that 12% of articles about leadership still focus on this topic (Bono & Judge, 2004). Another fertile line of research examining individual characteristics of leaders highlights the importance of behavioral styles (Bass & Avolio, 1993). Beginning with the pioneering work of Bales (Bales & Strodtbeck, 1951)on task and socioemotional leadership, research has pointed to two main clusters related to leader behavior: consideration and initiation of structure (Hemphill, 1950). Behavioral observations from laboratory studies show that leader emergence is clearly associated with behaviors such as talkativeness (Dabbs & Ruback, 1987; Mullen, Salas, & Driskell, 1989) and interruptions (Hall, Coats, & LeBeau, 2005; Malloy & Janowski, 1992; Ng, Brooke, & Dunne, 1995). Moreover, a great deal of attention has been devoted to the examination of more complex styles of leadership, such as charismatic leadership (House & Shamir, 1993) and transformational leadership (Bass & Avolio, 1993). Successful charismatic leaders gain their position by exhibiting self-confidence, task ability, emotional expressiveness, and conviction in their beliefs, and they are able to induce in their subordinates a high level of loyalty, intellectual stimulation, and consideration (Bass, 1990). Although individual differences in personality or in behavioral styles can influence leadership emergence, other authors posit that certain variations emerge as a function of the context (e.g., group or organization) in which the leader is embedded. These theories vary in placing anywhere from a strong to a moderate emphasis on context, depending on the relative 150 DAVID A. KENNY AND STEFANO LIVI influence attributed to the leader. In essence, these theories emphasize other levels besides the target. A variety of theories attempt to describe leadership behavior in a more dynamic, interpersonal system of perceivers and the perceived. The cognitive perspective to leadership has emphasized that leadership is as much a property of the person who is doing the perceiving as it is a property of the person who is being perceived (Norris-Watts & Lord, 2004). According to Calder’s (1977) attribution theory, leadership is an inferred disposition or internal quality based on shared beliefs about a leader’s behaviors and traits that affect members’ perceptions, information encoding, and retrieval of relevant information (Kenney, Blascovich, & Shaver, 1994). Attributions play a prominent role in most implicit theories of leadership as well as in Meindl, Ehrlich, and Dukerich’s (1985) theory about the romance of leadership. In these models, the perceiver’s schemas guide the perception of a target more than the target’s actual behaviors. In this way, leadership traits are not a stable property of a person, but rather a perceptual abstraction that followers use to categorize leaders and to make sense of the leader’s behavior (Epitropaki & Martin, 2004). Following categorization theory (Lord, Foti, & De Vader, 1984), the prototype of leadership is an abstract conception or representation of the most representative leader (Phillips, 1984). The attributes of a particular leader, both traits and behaviors, are compared to the prototypic attributes of a preexisting leader category (Rush & Russell, 1988). Prototypes are also used by leaders as a vehicle for expressing their behavior and to manage their leadership perception. The result is a holistic, context-free, implicit theory of leadership (Epitropaki & Martin, 2004). Interestingly, recent developments of these theories, such as the connectionist models of leadership (Lord, Brown, & Harvey, 2001), hypothesize that the contextual constraints play a more important role, suggesting that strong variation across and within perceivers is a function of different contexts (e.g., when the context changes or is based on their experience; Brown & Lord, 2001). The social identity theory of leadership (Hogg, 2001; Hogg & van Knippenberg, 2003) emphasizes that a leader’s effectiveness heavily depends on that leader’s similarity to the group’s prototype. In this way, leadership perceptions vary among groups depending not only on the leader fit to the group prototype, but also on the group identification of the followers (Cicero, Pierro, & van Knippenberg, 2007). Thus, these theories of leadership perception suggest that if an organization or a workgroup shares a common knowledge of leadership owing to shared cultural values, members should agree about who is and who A Componential Analysis of Leadership Using the SRM 151 is not the leader in the group. Conversely, if group members are involved in different socialization processes or if leaders have different functions for different followers, less agreement will be apparent. Classic contingency theories, such as Fiedler’s (1978) contingency theory, or path-goal theory (House, 1971) and normative decision theory (Vroom & Yetton, 1973), assume that leadership effectiveness depends on fit between the leader’s personal characteristics and style, and the demands of the situation. For example, Fiedler’s theory assumes that leader effectiveness is a joint function of leader’s personal style (task or relationship oriented) and situational control (the ability of a leader’s relations to provide follower clarity in group tasks and authority). Relationship-oriented leaders are predicted to be more effective when they have a moderate control over the situation, whereas task-oriented leaders are more effective when they have low or high control. Other theories have pointed out that the contingency is at a dyadic level between leader and follower, and that the development of the relationship is maintained through the exchange of valued resources (Hollander, 1958). Thus, leadership is viewed as a reciprocal process in which leader and follower exist in a mutual relationship, such that leadership cannot be determined independently of followership (Chemers, 1997; Hogg, 2005; Hollander, 1992; Messick, 2005). Beginning with the early intuition of Hollander (1958) in his theory of idiosyncratic credit, many theories have emphasized that leadership is a process that involves followership and reciprocal action between leader and follower – as in leader–member exchange theory (Yukl, 1994) or vertical dyad linkage theory (Dansereau, Graen, & Haga, 1975). For example, the leader–member exchange theory assumes that leaders have different relationships with different subordinates; as a consequence, leadership effectiveness depends on a dyadic relationship between leader and follower (Levine & Moreland, 1998). Finally, other theories give more importance to context and, therefore, emphasize effects at the group level (Emrich, 1999). For instance, Lord, Binning, Rush, and Thomas (1978) showed that greater levels of leadership are perceived when a group is successful than when it fails. Schyns and Sanders (2003) demonstrated that mood can affect the perception of leadership. Pierro, Mannetti, De Grada, Livi, and Kruglanski (2003) showed the positive effect of temporal pressure in groups on leadership perception. According to the theory of substitutes of leadership (Kerr, 1977; Kerr & Jermier, 1978), environmental, social, and organizational factors all strongly influence task performance and are able to reduce the necessity of leadership: Subordinates (e.g., ability, experience), task (e.g., low 152 DAVID A. KENNY AND STEFANO LIVI ambiguity), and organization (e.g., group cohesion, support by supervisors, organizational formalization) may all potentially render the function of leadership unnecessary. Thus, leadership varies as a function of type of organization, the structure, and group characteristics in which the leader operates. Other researchers have emphasized cultural factors that affect leadership perception and the expressions of leader’s and followers’ personal and common needs (Ayman & Chemers, 1991). In this case, successful leaders are able to help followers to reach goals and personal needs, but those needs are oriented by socialized values that are settled at the cultural level (Chemers, 2000). With some notable exceptions (Brown & Lord, 2001), few theories of leadership have explicitly hypothesized effects due to the perceiver, although some measures of the implicit theories of the leadership recently developed assume differences in the way individuals perceive the leadership prototypes (e.g., Epitropaki & Martin, 2004). If, for some people, their prototype would be easier to match than it would be for others, these perceivers would see more leadership than others.

ORGANIZATION OF THIS CHAPTER

As we have just described, leadership has been shown to operate at the level of the relationship, the person (both perceiver and target), and the group. A fundamental challenge of leadership research is to simultaneously study leadership at multiple levels. The social relations model (SRM) represents one way of responding to that very difﬁcult challenge. The remainder of this chapter begins by presenting a detailed description of the components of the SRM. We discuss both variances of and correlations between these components. In the next section, we follow the lead of Lord et al. (2001) and examine the sources of variance in group members’ perceptions of leadership. Using data from seven studies that included multiple measures of leadership, we partition the variance in leadership ratings to determine the relative importance of group, relationship, perceiver, and target in the perception of leadership. In addition, we examine reciprocity of the perception of leadership: If Jack thinks that Jill is a leader, does Jill think that Jack is a leader? After reviewing the results of these SRM analyses, we consider them within the context of leadership theory. A Componential Analysis of Leadership Using the SRM 153

Next, we consider correlations of variables with leadership using the SRM. First, we explore the differences between task-oriented and socioemotional leadership, using the model to study convergent and discriminant validity of these two leadership styles. Second, we examine gender differences in the perception of leadership, considering gender differences at the three levels of analysis. These ﬁrst two sections illustrate how we can attempt to explain the components of leadership. Third, we consider the overlap between self-perception of leadership and the perception of leadership by group members. The fundamental question here is how self- perception of leadership differs from the perception of others. The ﬁnal and most technical section of the chapter focuses on the statistical analysis of the SRM. We show how standard statistical packages (i.e., SAS and SPSS) and structural equation modeling (SEM) packages (i.e., Amos) can be used to estimate the model’s parameters. We have tried as much as possible to keep the technical, statistical details to a minimum in this chapter. These details, while important, are likely to not be of interest to leadership researchers. Nevertheless, our discussion of the estimation of the model probably includes more detail than most readers would prefer to see.

SOCIAL RELATIONS MODEL

Most multi-level models consider just two levels of analysis, in which the two levels have a hierarchical structure – that is, units from one level are nested within the other level (e.g., persons in groups). Another multi-level model allows for an examination of all of the levels – group, individual (perceiver and target), and dyad. Known as the social relations model (SRM; Kenny, 1994; Kenny & La Voie, 1984), it can be used to empirically determine the amount of variance in a particular measure at each level. To do so, there must be a group of persons such that each group member rates every other member in the group on leadership. Such a design is commonly called a round-robin. Self-ratings may be constructed as well, but they are set aside because they may be qualitatively different from ratings of others. (In a later section of this chapter, we return to the question of what to do with self-ratings.) The SRM includes ﬁve random variables, each of which is a source of variance: Group: differences among groups in how much leadership is perceived. Perceiver (individual): differences among individuals in how much leadership they perceive in others, in general. 154 DAVID A. KENNY AND STEFANO LIVI

Target (individual): differences among individuals in how much each is perceived as a leader, in general, by the others. Relationship (dyad): idiosyncratic perceptions of the extent to which someone is perceived as a leader by a particular other. Error: measurement noise in the judgment of leadership.

We review each of these components in more detail. Is there more leadership in some groups than others? Group variance would be indicated by a large amount of variance in the group means of leadership. There are two opposing ways to think about whether leadership would show group effects. The first perspective on leadership is that it is a zero-sum phenomenon; that is, in a group one person becomes the leader and the others become followers, as in a rank ordering. In this case, there would be no expectation that leadership would vary by group, because one member tends to emerge as a leader in each group. The second view is that leadership represents a potential that may or may not be met. For example, if a group fails, the cause for the failure may be that it lacked actual or perceived leadership (Lord et al., 1978). Within this perspective, one would expect group variance, because leadership could be unequally allocated among groups due to different leadership styles, differential participation of the individual group members, unequal power expressed or assigned to each group member, different leadership prototypes (Hogg, 2001), or different group social constructions of leadership responsible for a specific or co-constructed perception of leadership (Meindl et al., 1985). The perceiver effect represents how much individuals see leadership in others in general: Do some people view others in general as leaders, whereas others do not? For leadership, the perceiver is probably the least theoretically interesting component. However, for measures of socioemotional leadership (i.e., leadership focusing on member relations, rather than productivity), the perceiver effect might reflect how much the person enjoys being in the group, how much the individual likes the other group members, and how much the person is involved and participates in group activities. The target effect represents the extent to which a person is seen as a leader in the group; target variance taps into whether people agree as to how much leadership each member exhibits. With social perception, a qualification becomes necessary: A person may be seen one way by members of the group, but the person may not actually be that way. In other words, perception may not necessarily match reality. At the same time, it can be argued that a major part of being a leader is being seen by other group A Componential Analysis of Leadership Using the SRM 155 members as a leader. In this sense, individuals who are perceived as showing more leadership are actually leaders, whereas those perceived as showing less leadership are actually followers. Consensus about the degree to which members are leaders in the group could be based on the amount of influence each person actually has or on shared stereotypes about how leaders are supposed to behave (e.g., leaders are tall, attractive, white males who talk a great deal). The relationship effect represents the idiosyncratic perception that a perceiver has of a target’s leadership. In the context of leadership ratings, the bulk of relationship variance would represent disagreement between perceivers, but it might also reflect personal preferences or alliances between particular group members. Moreover, leadership, especially socioemotional leadership, might potentially exist at the level of the relationship, as it better reproduces dyadic asymmetries among group members. That is, some members of the group may be more in need of leadership than other members of the group, and those members who are in need of leadership may find it from different members. Yet another view of the relationship effect is that it represents the biases and needs of the perceiver. We return to this notion when we discuss the difference between self-perception and other perception later in this article. Error represents sources of variance in the judgments that are neither systematic nor replicable. To separate error variance from the relationship variance in the SRM, at least two measures of the construct must be present – either repeated measures in time or different measures of the same construct. For instance, leadership might be measured by five items: contribution to the task, influence over the group’s solution, leadership exhibited, control of the group’s activities, and preference to have as a leader (Lord, Phillips, & Rush, 1980). As discussed later in this chapter, the SRM makes it possible to conceptualize four different types of error. The SRM is a random effects model. That is, its components are random variables. Consider the target effect: The interest in a study of leadership is not in knowing which individuals have the highest and lowest target effects, but rather in the amount of variance produced by perceptions of targets in the population. Thus, the major focus in a univariate SRM analysis is the sources of variance in the perception of leadership. A further complication of the SRM is that it includes two sources of nonindependence that can be interpreted as reciprocity correlations. The first source is a correlation between a person’s perceiver and target effect: If Jill is seen by others as leader, does she think that others are leaders? This correlation might well be negative. That is, if Jill is seen as leader, 156 DAVID A. KENNY AND STEFANO LIVI she may think the others in the group are not leaders. This correlation, which has been called generalized reciprocity, is defined only if there is both perceiver and target variance. The second source involves a correlation between two relationship effects from the same dyad. If Jack especially thinks that Jill is a leader (i.e., more than Jack thinks others are leaders and more than others think Jill is a leader), then does Jill especially think that Jack is a leader (i.e., more than Jill thinks others are leaders and more than others think Jack is a leader)? This correlation, which has been called the dyadic reciprocity correlation, could be negative – for instance, if the group is very hierarchical. Later in this chapter, we discuss the correlations between two variables measured in a round-robin design. The estimation of the SRM is complicated because it contains five random variables and two covariances. We consider the rather complicated statistical details in the estimation of the SRM variances and covariances in the last section of this chapter.

SOURCES OF VARIANCE AND COVARIANCE IN JUDGMENTS OF LEADERSHIP

Recently, Lord et al. (2001) reviewed several studies of leadership that examined the sources of variance in judgments of leadership. In our review, we did two things a bit differently from these authors. First, we included only studies that had multiple measures or indicators of leadership. With multiple measures, we can separate error from relationship variance. Second, we used the SRM to estimate sources of variance for all of the studies, whereas Lord et al. (2001) sometimes used alternative methods.

Studies

To qualify, the study had to employ round-robin measures of leadership (i.e., each member of the group rated every other member in the group on leadership). Also, either the studies needed to report the SRM variance partitioning or we had to have access to the raw data so that we could perform the analysis ourselves. The seven studies with multiple measures of leadership are as follows: Campbell, Simpson, Stewart, and Manning (2003) studied 57 four-person, all-male groups who were given a leaderless group discussion task. A Componential Analysis of Leadership Using the SRM 157

Dabbs and Ruback (1984) studied 20 five-person, same-gender groups who were asked to get acquainted. Dabbs, Ruback, and Evans (1987) studied 20 five-person, mixed-gender groups who were asked to discuss what they would bring from a list of items to a desert island. Lord et al. (1980) studied 24 four-person, mixed-gender groups who worked on four problem-solving tasks. Malloy and Janowski (1992) studied 10 six- to eight-person, mixed-gender groups. Montgomery (1986) studied 45 four- and five-person, mixed-gender groups. Groups met for three sessions, and the last two ratings of leadership are used as two measures of leadership. Ruback, Dabbs, and Hopper (1984) studied 20 five-person, mixed-gender groups who were asked how they could increase tourism for a community.

Across the seven studies, there were 820 participants who were placed in 180 groups.

Variance Components

The seven studies used multiple measures of leadership, so we were able to examine the extent to which the variances remained stable across these multiple measures. This allowed us not only to separate error from relationship variance, but also to differentiate four sources of ‘‘error.’’ This process yielded eight components that gave rise to the eight SRM sources of variance:

Stable group: differences among groups in how much leadership is perceived that is consistent across measures. Unstable group: differences among groups in how much leadership is perceived that is inconsistent across measures. Stable perceiver: differences among individuals in how much leadership they perceive in others in general that is consistent across measures. Unstable perceiver: differences among individuals in how much leadership they perceive in others in general that is inconsistent across measures. Stable target: differences among individuals in how much each individual is perceived as a leader that is consistent across measures. Unstable target: differences among individuals in how much each individual is perceived as a leader that is inconsistent across measures. 158 DAVID A. KENNY AND STEFANO LIVI

Stable relationship: disagreement about the extent to which someone is a leader that is consistent across measures. Unstable relationship: disagreement about the extent to which someone is a leader that is inconsistent across measures.

For instance, stable target variance implies that some persons are seen as leaders and others as followers across all the measures, whereas unstable target variance implies that someone who is seen as leader on one measure is not seen as a leader on the other measures. In past SRM studies, the group variance, whether stable or unstable, has been ignored, and all remaining unstable variances have been added together and treated collectively as error variance. However, unstable relationship variance comes closest to what most investigators would treat as pure error variance. Thus, the ability to separate stable and unstable variance across measures of leadership provides a more detailed picture of the variance partitioning. If all indicators measured the same construct and were equally reliable, then ideally no unstable perceiver and target variance would exist. Under normal circumstances, then, we would hope to have small amount of unstable perceiver and target variance.

Results

To help us better understand the meaningfulness of each component, we have included its reliability (Bonito & Kenny, 2007) assuming a six-person group with three measures of leadership – fairly typical values for the research that we reviewed. Reliability represents how much ‘‘true’’ variance is present in a measurement of a given component.

Variances To determine whether the group has an effect on judgments of leadership, we estimated the random variance due to the group. When we examined the data from several studies, we did not ﬁnd any evidence for group variance. In fact, we often obtained estimates of negative variance – a result consistent with the view that leadership is a zero-sum game. On the basis of this evidence, we conclude that groups do not vary in the amount of leadership perceived. For this reason, we do not include the group factor, both stable and unstable, in our table of results. Table 1 presents the results from the variance partitioning. On average, the amount of target variance is 0.427, which indicates that group members A Componential Analysis of Leadership Using the SRM 159

Table 1. Proportions of SRM Variance from Seven Studies.

Study Stablea Unstableb

Perceiver Target Relative Perceiver Target Relative

Campbell et al. (2003) 0.126 0.355 0.179 0.034 0.039 0.267 Dabbs and Ruback (1984) 0.135 0.457 0.140 0.030 0.070 0.168 Dabbs et al. (1987) 0.124 0.465 0.200 0.021 0.039 0.151 Lord et al. (1980) 0.122 0.363 0.132 0.093 0.022 0.267 Malloy and Janowski (1992) 0.061 0.449 0.241 0.033 0.069 0.148 Montgomery (1986) 0.000 0.304 0.315 0.082 0.006 0.294 Ruback et al. (1984) 0.040 0.598 0.136 0.014 0.059 0.153 Average 0.087 0.427 0.192 0.044 0.043 0.207 Reliabilityc 0.558 0.861 0.736 0.714 0.707 –d aStable variance is variance that correlates across different measures of leadership. bUnstable variance is variance that is specific to a given measure of leadership. cPresuming that there are six people in the group and three measures. dUndefined. agree substantially on the degree of leadership displayed by each participant. A small variance would have indicated that targets are perceived differentially by perceivers. The reliability of the target is 0.861, the highest reliability of any component. Notably, the variance partitioning from these seven studies is very similar to the variance partitioning from studies in which there is a single measure of leadership and when leadership is considered a rank-order measure (Livi et al., 2008). The literature demonstrates that substantial consensus has been reached regarding variance and leadership. First, agreement concerning leadership is greater than agreement found for personality traits, which averages to 0.275 (Kenny, 1994). Second, roughly equal amounts of relationship and target variance for physical attractiveness ratings have been identified (Ho¨nekopp, 2006). Thus, perceivers agree much more about leadership than they agree about physical attractiveness. Third, in an interesting study, Vargas (1986) found that college athletes nearly agreed about as much who was a leader on their team (proportion of target variance ¼ 0.404) as they agreed about athletic skill (0.413). Fourth, there is more target variance than error variance in leadership studies, another unusual result that indicates people agree about who is the leader of the group. Relatively small but not trivial amounts of perceiver variance have been found, with the average being about 10% of the variance. The reliability of the stable perceiver effect is somewhat disappointing, being only 0.558. 160 DAVID A. KENNY AND STEFANO LIVI

There is more evidence of stable relationship variance, which averages around 20% of the total variance and is much more reliable (0.736). Because perceiver variance and relationship variance both index disagreement about who is a leader in the group, it is interesting to note that target variance or agreement (0.427) is larger on average than disagreement (0.279 ¼ 0.087 þ 0.192). In summary, about 10% of the variance is due to perceiver variance, 20% to relationship variance, 40% to target variance, and 30% to error variance.

Correlations As we outlined in the previous section, the SRM encompasses two types of reciprocity correlations. The ﬁrst, called generalized reciprocity, correlates the perceiver effect with the target effect: If someone is seen as a leader by others, does that person see others as leaders? The second, called dyadic reciprocity, correlates relationship effects: If Alice thinks Bob is more of a leader than do others, does Bob see Alice as leader more than do others? Reciprocity in the perception of leadership, both negative and positive, could arise for several reasons. First, reciprocity would emerge when leaders have legitimacy. That is, if a leader is legitimated, he or she would also be perceived as an authority. Explicit asymmetries might lead to negative reciprocity correlations (i.e., if Bob is seen as a leader, he sees others as followers). Additionally, groups under temporal pressure are more likely to be centralized, forming a more clear structure of asymmetries among members (Pierro et al., 2003). This could lead to a stronger target effect, and possibly negative reciprocities. Positive or no correlations could be revealed in groups where leadership is more distributed, as in groups characterized by decentralized or laissez-faire leadership style (Bass, 1990). Second, certain factors could lead to positive reciprocity. Because leadership is ordinarily positively valued, then if two people like each other, they would both be biased to see each other as leaders, given that liking shows strong positive reciprocity (Kenny, 1994). This would imply a positive correlation at the dyadic level. Table 2 presents the reciprocities for a stable perceiver with a stable target (generalized reciprocity) and between the stable relationship effects (dyadic reciprocity). Although there is some variation in the correlation shown in Table 2, on average there is little correlation at both the generalized and dyadic levels. Reciprocity appears not to be present in the perception of leadership, at least in leaderless groups. Nonetheless, it is useful to know that these correlations are near zero, for two reasons. First, when group size is 3, as in the study conducted by A Componential Analysis of Leadership Using the SRM 161

Table 2. Generalized (Perceiver–Target) and Dyadic Reciprocity Correlations for Seven Studies.

Study Generalized Dyadic

Campbell et al. (2003) 0.203 0.034 Dabbs and Ruback (1984) 0.194 0.020 Dabbs et al. (1987) 0.274 0.106 Lord et al. (1980) 0.321 0.163 Malloy and Janowski (1992) 0.272 0.084 Montgomery (1986) –a 0.331 Ruback et al. (1984) –a 0.177 Average 0.144 0.034 aInsufﬁcient variance to compute a meaningful correlation.

Zaccaro et al. (1991), the full SRM cannot be estimated (i.e., the model is not identified). If we can assume that the dyadic reciprocity correlation equals zero, the model is identified with as few as three persons per group. Thus, the dyadic correlation is beneficial in that, if we assume it to be zero, we are making a valid assumption. Second, the SRM may be estimated using a SEM approach, as described later in this chapter. That model is very complicated, so adopting the assumption that some correlations are zero would greatly simplify an otherwise overly complicated model.

DISCUSSION OF THE SRM RESULTS

In this section, we relate the empirical ﬁndings in Tables 1 and 2 to our earlier review of the theoretical literature. Recall that our results came from leaderless discussion groups where there was no formal leadership role, so the perception of leadership was determined only by peers in the group discussion, and not by formal roles.

Target Variance

We have seen that the dominant component in leadership judgments is the target: Group members agree that some members in the group are leaders and other members are followers. In our survey of seven studies that used multiple measures of leadership, we found that 43% of the variance was target variance. Livi et al. (2008) report variance partitioning in the four 162 DAVID A. KENNY AND STEFANO LIVI studies that employed a single measure of leadership and found 49% of this variance to be target variance. Also, in the five studies of leadership ranking, Livi et al. found that 48% of the variance was target variance. Clearly, leadership is not totally ‘‘in the eye of the beholder,’’ and the obstinacy that many authors demonstrate in continuing to analyze target variance has a strong empirical basis. Livi et al. did find that there is less target variance and more relationship and perceiver variance when leadership is socioemotional than when it is task oriented. Thus, task-oriented leadership is more clearly consensual, in the sense that members agree on task-oriented leadership rather than on socioemotional leadership. Because leadership consensus on transformational and laissez-faire leadership have been shown to moderate the emotional exhaustion–work commitment relationship (Cole & Bedeian, 2007), it would be interesting to determine whether more transformational leadership styles (e.g., charismatic leadership) are less consensual and laissez-faire style is more consensual. Given that the target component represents the leadership perception of a member, controlling for other sources of variance, it can be correlated with external criteria; in this way, we may examine the behavior and trait correlates of consensually perceiver leadership. We might first examine which behaviors leaders, rather than followers, exhibit. We might then take that behavioral profile and try to match it with the group prototype of leadership (Epitropaki & Martin, 2004; Hogg, 2001) or leadership stereotypes (Lord, Brown, Harvey, Hall, 2001). Alternatively, we could examine which personality characteristics of the leader predict the perception of leadership (Lord et al., 1986). We could then examine whether behaviors or alternative prototype or stereotype matches mediate the effect of traits on the perception of leadership. Another interesting question is the extent to which agreement about leadership reflects appearance-based (not behavior based) stereotypes. In zero acquaintance (Albright, Kenny, & Malloy, 1988) and thin-slices (Ambady & Rosenthal, 1992) studies, perceivers with minimal information have been shown to arrive at consensual and accurate judgments of targets. To date, no studies have specifically examined the perception of leadership using minimal information, although a thin-slices study did focus on the effectiveness of negotiation (Curhan & Pentland, 2007). It is also known that baby-faced targets are seen as low in dominance, making it unlikely that they would be perceived as leaders (Zebrowitz-McArthur & Apatow, 1983–1984). Finally, stereotypes about gender and leadership A Componential Analysis of Leadership Using the SRM 163

(Ridgeway, 2001) have been extensively studied. (The effect of gender on the target effect is discussed later in this chapter.) Livi et al. (2008) found that target variance increases with group size. Perhaps, leadership becomes both more necessary and more evident in larger groups, much in the same way that extroversion becomes more apparent in groups than in dyads (Kenny, 2004). These results provide further empirical support for the claim that the emergence of a more hierarchical structure within groups is determined in part by a larger group size. Finally, just because members agree about a target’s leadership in their group, it does not necessarily imply that the same person would be a leader in another group. Although information on this topic is limited, data from rotation designs (Foti & Hauenstein, 2007; Zaccaro et al., 1991) and other sources (Albright & Forziati, 1995) indicate that leaders in one group are very likely to be leaders in other groups.

Perceiver Variance

We also ﬁnd evidence for perceiver variance that indicates the extent to which some perceivers see others as leaders, and other perceivers see others as followers, though this component explains only 10% of the variance. In addition, a small, positive generalized reciprocity (correlation between perceiver and target effect) is present. This result may reﬂect the fact that leaders believe leadership is important and are then biased to see leadership in others. This correlation is not very large, however, averaging only 0.144. Although the level of perceiver variance is small, it is important to underscore, as other authors have done (Lord et al., 2001), that the perceiver effect could be an important factor in leadership perception. As Livi et al. (2008) show, much more perceiver variance is present in the perception of socioemotional leadership.

Relationship Variance

Relationship variance reﬂects idiosyncratic perception of leadership. For example, perhaps Peter sees Paul as a leader, but no one else does, and Peter does not just see everyone as a leader. Oddly enough, although relationship variance is the second largest source of variance (after target variance), it has traditionally received very little attention from researchers in this ﬁeld. 164 DAVID A. KENNY AND STEFANO LIVI

Future research might, therefore, focus on within-group dyadic dynamics, as proposed by leader–member exchange theory (Yukl, 1994) and by vertical dyad linkage theory (Dansereau et al., 1975). An alternative explanation for relationship variance might also involve the presence of triadic effects (Hoff, 2005). For instance, a group might have two factions, such that each faction has a different leader. In this case, subgroups might disagree about the identity of the larger group’s leader. Ideally, future leadership research studies will investigate such phenomena. Categorization and attribution analysis of leadership perception (Lord et al., 2001) might beneﬁt by focusing more intensely on relationship variance. In particular, to avoid confounding factors in relationship and error variances, a study would need to employ multiple measures of leadership. Intragroup dyadic dynamics – which are especially important in long-term groups, such as organizations and teams (Tsui, Porter, & Egan, 2002) – could be intensively investigated in this manner. The results in Table 2 show a weak dyadic reciprocity effect, and we expect this result might be likewise found in other studies of leaderless groups, such as those involving college students. Members in leaderless groups, where status is not predetermined and where interaction is limited to a short discussion session, have less of a chance to make transactions or exchanges among members (e.g., leader–member exchange theory; Yukl, 1994). Perhaps in more long-standing groups, dyadic reciprocity of leadership might be found. To examine this hypothesis, we reviewed the Montgomery (1986) study, which did examine groups overtime. Her data do, indeed, show evidence of increasing positive dyadic reciprocity overtime.

Group Variance We failed to find any evidence of group variance in the perceptions of leadership. Although most prior SRM studies of person perception in college students have not attempted to find group variance, studies of families that have used the SRM have looked for variance in family effects (i.e., group effects). For instance, Cook (2001) studied influence and control in families and found that only 4% of the total variance was at the level of the group. It may be especially difficult to find group variance in leadership because it may operate in zero-sum fashion. In other words, leadership, much like status and power, may be a finite resource that is divided among group members: Not everyone in the group can be a leader. A Componential Analysis of Leadership Using the SRM 165

Our failure to ﬁnd group variance may be somewhat surprising, given that various theories propose that groups could have signiﬁcant differences relative to one another (e.g., Brown & Lord, 2001; Hogg & van Knippenberg, 2003). All of the studies that we reanalyzed involved leaderless, largely college student groups, however, and it is likely that there are only small differences between groups owing to their shared stereotypes and socialization histories. Also, our results were obtained in laboratory experimental settings, where there is low acquaintanceship and likely shared prototypes that do not differ from group to group. Moreover, many theories have explicitly pointed out that groups should have differences in leadership perception as a function of culture (Chemers, 2000) or group context characteristics (Kerr, 1977). Ideally, to properly assess these factors, the research design should manipulate group composition to measure the effect of these variables.

UNDERSTANDING THE DIFFERENCE BETWEEN TASK-ORIENTED AND SOCIOEMOTIONAL LEADERSHIP

We now consider the correlation between two or more variables, at least one of which is an SRM component. In this section, we take up the question of correlation between task-oriented and socioemotional leadership; that is, we take an SRM component in task-oriented leadership and correlate it with its parallel component in socioemotional leadership. The interest here focuses on convergent and discriminant validity. In the next section, we consider gender as a predictor of leadership. In the third section, we consider the relationship between self-perceptions of leadership and the perception by fellow group members. This examination reveals that self-ratings can be thought to have perceiver, target, and relationship components. A classic distinction in leadership research is task-oriented versus socioemotional leadership (Bales & Strodtbeck, 1951). Simply put, task- oriented leaders are more concerned with productivity, whereas socioemotional leaders are more concerned with member relations. Livi et al. (2008) examined the differences between task-oriented and socioemotional leadership, and found a very different pattern of variance partitioning for the two types of leadership. First, more than twice as much target variance was observed when leadership was task oriented than when it was socioemotional. Second, nearly twice as much relationship variance and 166 DAVID A. KENNY AND STEFANO LIVI perceiver variance were observed in conjunction with socioemotional leadership than with task-oriented leadership. We consider here the correlations between SRM components across leadership types. In the Dabbs and Ruback (1984) study, 20 groups of size 5 – speciﬁcally, same-gendered, leaderless discussion groups with college students – were involved in a get-acquainted session. The indicators of task- oriented leadership were attention, leadership, and contribution, and the indicators of socioemotional leadership were encouragement of participation and showing interest. The correlations of the stable components were as follows: Perceiver: 0.869; Target: 1.000; Relationship: 0.431. We were somewhat surprised by the poor discriminant validity at the level of target in this study. The results indicate that task-oriented and socioemotional leadership are most distinct at the level of relationship. One way of understanding this result is to note that task-oriented leadership resides more in the target, whereas socioemotional leadership resides more in the relationship. Recall Livi et al.’s (2008) ﬁnding that task-oriented leadership showed nearly twice as much target variance as socioemotional leaders, whereas the reverse pattern was observed for the relationship. When we performed the same analysis on the data gathered by Ruback et al. (1984), we found no discriminant validity between the two types of leadership for any of the three components. That is, the correlation between the two types of leadership was perfect for stable effects due to perceiver, target, and relationship. Recall that the Ruback et al. study involved a production-focused task (i.e., trying to increase tourism in a community), whereas the task in the Dabbs and Ruback study was simply ‘‘to get acquainted.’’ In these short-term groups, the separation of these two types of leadership may not occur as it does in more task-oriented groups.

GENDER AND PERCEPTIONS OF LEADERSHIP

Over the years, many studies have examined the effect of gender on leadership. Since the 1980s, two analyses have made particularly important contributions to this line of research. In the ﬁrst study, Deaux (1984) argued that there was no biological basis in gender differences in social behavior. In the second contribution, the meta-analyses of Eagly and colleagues A Componential Analysis of Leadership Using the SRM 167

(Eagly & Johnson, 1990; Eagly & Karau, 1991; Eagly, Makhijani, & Klonsky, 1992) demonstrated that some weak effects are present in terms of gender differences in leadership effectiveness and perception of leadership. These differences are mainly due to social categorization. In particular, men are more likely to emerge as leaders when measures of general or task leadership are used, whereas women are more likely to emerge as leaders if socioemotional measures are used (Eagly & Karau, 1991). It seems that when the situational requirements match those associated with the feminine stereotype, women emerge as leaders. According to the social-role theory of gender differences (Eagly, 1987), these stereotypes tend to work against women by creating negative expectations, which mitigate the likelihood that women will be perceived as leaders in task-oriented situations. Another relevant issue is the effect of solo status or tokenism. Of particular concern is the extent to which women’s contributions might be devalued when the group includes only one woman (Fuegen & Biernat, 2002). Using the SRM can clarify the empirical and theoretical meanings of gender differences in the study of leadership, answering long-standing questions and proposing other interesting and intriguing novel queries. We begin by noting that there is not just one gender-related issue to be addressed in the study of leadership. Among the questions that have been considered are the following: Are men or women more likely to be seen as leaders? Do men or women see more leadership? Are men more likely to be seen as leaders when men (rather than women) are judges? How are men and women affected by the gender composition of the group? For instance, are women less likely to be seen as leaders when the group includes fewer women? We examined each of these questions simultaneously using the SRM. In so doing, we used the data gathered by Lord et al. (1978). In that study, there were 24 groups, each with 4 persons in laboratory settings. Speciﬁcally, groups of Carnegie–Mellon students worked together on several tasks. To simplify the analysis, we averaged the ﬁve measures of leadership and created three variables: gender of the perceiver, called Pgender (1 ¼ female; 1 ¼ male); target gender, called Tgender (1 ¼ female; 1 ¼ male); and number of females, called Nfem (0–4). Note that the interaction of Pgender with Tgender captures the difference between same-gendered pairs (þ1) and mixed-gendered pairs (1). We also allowed Nfem to interact with 168 DAVID A. KENNY AND STEFANO LIVI

Pgender, Tgender, and their interaction. In terms of levels of analysis, Pgender and Tgender are at the individual level of analysis, their interaction is at the dyadic or relationship level, and Nfem is at the group level. So far in our discussion of the SRM, we have focused on the components of perceiver, target, and relationship. Each of these is a random variable. Considering the target effect, our interest in that component lies not in determining which person in particular is seen as a leader in the group, but rather in estimating the population variance due to target. The gender variables are all fixed variables. We are interested in particular values (i.e., the means) of the levels of that variable. For instance, we want to know if men are more likely to be seen as leaders than are women. The addition of the fixed variables to the SRM greatly complicates the statistical model and the analysis. For technical reasons, the complications are especially problematic when the fixed variables are at the level of relationship (e.g., Pgender Tgender). Standard SRM software cannot properly handle such a model. Fortunately, advances in multi-level modeling have been made and computer programs such as SAS can accomplish such analyses. We used SAS to estimate the model and controlled for nonindependence due to the SRM variances and covariances. In a later section of this chapter, we describe the details of the estimation method. The model includes seven fixed effects; and the four SRM variances of group, perceiver, target, and relationship; and the two covariances, perceiver– target and relationship. Table 3 shows the results for the fixed effects.

Table 3. Solution for the Fixed Effects of Gender from Lord et al. (1978) Study.

Effect Estimate dft p

a Pgender 0.1503 118 2.01 0.0469 b Tgender 0.1286 199 1.55 0.1231 d Pgender Tgender 0.0338 141 0.54 0.5879 c Nfem 0.0384 77.3 0.45 0.6538 Pgender Nfem 0.0528 153 0.75 0.4564 Tgender Nfem 0.0434 129 0.53 0.5971 Pgender Tgender Nfem 0.0669 106 0.88 0.3805 aPerceiver gender coded 1 ¼ female and 1 ¼ male. bTarget gender coded 1 ¼ female and 1 ¼ male. cNumber of females in the group, 0–4. dSame-gender (1) versus mixed-gender (1) dyads. A Componential Analysis of Leadership Using the SRM 169

Note that some of the degrees of freedom in the t tests are fractional; fractional degrees of freedom can occur within multi-level modeling because error variances are pooled across levels. Interestingly, we found only one statistically significant fixed effect: Female perceivers see more leadership than do male perceivers. The gender of target effect was not significant, and we even found that women tend to be seen as leaders more than men. If men devalue women’s contributions more than women do (Eaglyetal.,1992), there should be an interaction between gender of the perceiver and gender of target – something we did not find. Moreover, unlike Fuegen and Biernat (2002), we did not find any statistically significant effects of the gender composition of the group. The analysis in this case resembles a multiple regression, so all effects are controlling for all other effects in the analysis. Unlike a traditional multiple regression, the analysis allows for all sorts of nonindependence. Of course, our study had some serious limitations. First, it took place at Carnegie–Mellon University, a school with a reputation for strong quantitative ability. Second, of the 96 participants in the study, only 42 were female. Thus, there were very few groups with large numbers of females. Nonetheless, our work illustrates how many interesting and important questions about leadership and gender can be analyzed. The results emerging from this study confirm the results from previous studies showing a lack of target gender differences in leadership when measured in a university context (Eagly & Karau, 1991). Moreover, the absence of the interaction effects of the actor’s gender with the gender of target, even when controlling for the composition of the group, confirms a substantial fairness in leadership perceptions. A more intriguing and perhaps original finding is the significant effect due to the perceiver: It seems, at least in this study, that the bias of a gendered evaluation of leadership is mainly in the eye of the beholders. We know of no prior studies or theories that explicitly analyze gender differences in the perceiver effect of leadership perceptions. Nevertheless, we can speculate that, following the social-role theory of gender differences, women more than men see more influence in the group. Another explanation of the gender differences in the perceiver effect could be an extension of the ‘‘contrastive effects’’ of stereotyped judgment (Biernat & Kobrynowicz, 1997): Members seem to judge leadership expressed in groups by comparing others to the expectations of their own gender. Thus, women perceive a greater extent of leadership in groups because they judge the influence of other members to be stronger in 170 DAVID A. KENNY AND STEFANO LIVI comparison to their negative leadership expectations and not in comparison to their actual expressed influence. In most of the previous studies, this gender difference in the perceiver effect was not measured and controlled. If a study included more men than women, and if women also tended to see more leadership than men, then it would falsely appear that men would be seen more as leaders than women because of the hidden gender difference in the perceiver effect. One last issue should be addressed. When we study a fixed effect, such as Pgender, we might ask how much of the variance in the perceiver effect gender explains. To answer this question, we conducted the run twice – once with gender variables in the model and once without them – and each time we determined the perceiver variance. Conducting such an analysis, we found that the variance due to the perceiver declined by 12.8% when we included gender variables in the model. Thus, we conclude that gender variables, mainly Pgender, explain 12.8% of the variance in the perceiver effect. Because the perceiver effect itself explains approximately 12.2% of the total variance (see Table 1), gender explains only 1.6% of the total variance, at least when only the perceiver effect is considered. Including the fixed effects of gender in the model reduced the perceiver variance while having little or no effect on the target or relationship variance. This result bolsters the conclusion that the effect of gender in this study is at the level of the perceiver. The strategy illustrated here could also be applied to the study of variables that predict the target effect. If such research were carried out, we would have a much better understanding of the percentage of variance that traits explain in the random effect of target than we currently have. That is, because error variance is measured in SRM analyses, a much larger proportion of variance would almost certainly be explained than is possible when using conventional methods.

SELF-PERCEPTIONS OF LEADERSHIP

The relationship between self-perception and the perception of others has been a focus in the leadership literature. Notably, Atwater and Yammarino (1997) have examined the role of self–other agreement. More recently, Anderson, Srivastava, Beer, Spataro, and Chatman (2006) examined the self-enhancement of status in groups: Do people over- or underestimate their status in the group, and do people know how others view their status? Anderson et al. (2006) found little or no evidence of self-enhancement of A Componential Analysis of Leadership Using the SRM 171 status, but did find appreciable levels of understanding of how others view their status. Kwan, John, Kenny, Bond, and Robins (2004) proposed applying the SRM to the study of self-perception. Kenny and West (2008) have extended the Kwan et al. (2004) approach and have developed a detailed, and relatively complicated, SRM approach to the comparison between self- perceptions and the perceptions of other members in the group. In this section, we apply their approach to the study of leadership. Kenny and West (2008) propose four ways in which self- and other perception of leadership can differ: Self-enhancement: Do people, on average, see themselves more as leaders than the other members of the group see them and more than they see others as leaders? Assumed similarity (the weighting of perceiver effect): If a person tends to see others in the group as a leader, does the person also see himself or herself as a leader? Self–other agreement (the weighting of the target effect): If a person tends to be seen by others in the group as a leader, does the person see himself or herself as a leader? Relationship variance: Is there more relationship variance in self-ratings (i.e., self-enhancement) than in the ratings of others (i.e., other enhancement)? We detail these four differences next: First, the equality of the means of self-ratings and the ratings of others may produce differences. If the self on average sees himself or herself as more of a leader than do the members of the group and more of a leader than he or she sees others, then self-enhancement would be present. Also possible is self-effacement: People see themselves as less of a leader than they see others as a leader and less than others see them as a leader. This difference refers to the average across all perceivers and targets; later, we discuss individual differences. Second, the weight of the perceiver effect in self-ratings, a parameter denoted as k in Kenny and West (2008), may lead to differences. If k is 1, then the idiosyncratic tendency of people to see others as leaders (the perceiver effect) is reflected equally for perceptions of the self. That is, self- ratings operate in the same way as ratings of others. If k is less than 1 but greater than zero, there is some assumed similarity, but self is seen somewhat differently than others. While possible, we do not usually find much evidence for k to be greater than 1. It is also possible for k to be 172 DAVID A. KENNY AND STEFANO LIVI zero (i.e., no assumed similarity) and even for k to be negative (i.e., a contrast effect). Third, the weight of the target effect in self-ratings, a parameter denoted as q in Kenny and West (2008), may produce differences. If q is 1, then the tendency of others to see a person as a leader is the same for the perception of self. That is, the target effect is equally well reflected in self-ratings as in ratings of others. If q is less than 1 but greater than zero, some self–other agreement is present, but self is seen somewhat differently than others. If q is greater than 1, then self has some sort of special insight: Self agrees more with a member in the group than two members agree with each other. It is possible for k to be zero (i.e., no self–other agreement) or even negative. Although negative or zero values for k seem implausible for traditional measures of leadership, they may be more plausible for certain types of transformational leaders. Certain sorts of leaders (e.g., Gandhi) may feel that they are not leaders at all, but see themselves as followers instead. 2 Fourth, there are differences in relationship variances, denoted as sg, for self and others. This is perhaps the most difficult comparison to understand. Kwan et al. (2004) conceptualized relationship variance as individual differences in enhancement. Thus, for self, it would represent individual differences in self-enhancement. For example, persons who experience high self-efficacy (Foti & Hauenstein, 2007) may have larger scores in relational self-perceptions. If someone has a positive relationship effect for self, then that person would think he or she is better than others – even better than members of the group see the person. Following the reasoning of Kwan et al. (2004), perhaps this measure of the individual difference in self- enhancement might moderate leader effectiveness (Atwater & Yammarino, 1992). For other perception, we can interpret the relationship effect as other enhancement. That is, the relationship effect represents the biased perception of a perceiver of a given target. Thus, the variances in both of the relationship effects quantify the biases of the perceivers. The question in terms of the comparison of self and other is whether there is more or less variance in the two types of judgments. Kenny and West (2008) argue that, because self-perception is generally considered to be more biased than the perception of others, we might expect more relationship variance in self- perception than in other perception. To be clear, relationship variance reflects differences in bias, rather than the overall level of bias. Another useful aspect of the approach taken by Kenny and West (2008) is that we can evaluate for the first time whether self-ratings operate in the same way as do the ratings of others. In such a situation, we would find that the means and relationship variances were the same in the two types of A Componential Analysis of Leadership Using the SRM 173 ratings and that k ¼ q ¼ 1. For instance, Kenny and West (2008) found that for ratings of emotions in groups of strangers, the model for self- and other ratings was the same. The estimation of the model is complex and uses SEM, a method detailed later in this chapter. The parameters k and q can be viewed as factor loadings; that is, self-ratings load on perceiver and target factors. For instance, Jack’s self-rating of his leadership loads on Jack’s perceiver and target factors.

Example

The Dabbs and Ruback (1984) study included 20 same-gendered groups of size 5. To reduce the complexity of the analysis, we did not separate error from relationship; rather, we simply summed the indicators to form the measure. We consider both task-oriented leadership (the sum of ‘‘attention,’’ ‘‘leadership,’’ and ‘‘contribution’’) and socioemotional leadership (sum of ‘‘encourages participation’’ and ‘‘shows interest’’). Table 4 presents the parameters in ratings of self and others for the two types of leadership. Of key importance is the test that both k (assumed similarity) and q (self– other agreement) equal 1. Both assumed similarity parameters were less than 1 for self-ratings. Consistent with the prior analyses undertaken by Kenny (1994), we found that the k paths were greater than 0 but less than 1. However, for q (self–other agreement), the pattern was different for the two types of leadership. For task-oriented leadership, the pattern was the typical one of q being less than 1: Members agree more with one another than they

Table 4. Comparison of Ratings of Self and Others.

Effect Task Oriented Socioemotional

Others Self Others Self ma 15.307 15.140 10.260 10.680 kb 1.000c 0.656 1.000c 0.934 qd 1.000c 0.706 1.000c 1.357 2e sg 5.785 6.152 4.094 1.455 aMean rating. bAssumed similarity parameter. cFixed to 1. dSelf–other agreement parameter. eVariance in enhancement. 174 DAVID A. KENNY AND STEFANO LIVI do with the self. Conversely, for socioemotional leadership, q is greater than 1: The self is a better informant concerning socioemotional leadership than are the other group members. In addition, we examined whether the self-mean and dyadic means were different. As can be seen in Table 4, they were not. Although mean differences were found, they were rather small. These results are consistent with the results of Anderson et al. (2006), who found little or no self- enhancement for perceptions of status. We also examined whether relationship variance differed for the perception of self and others. We did ﬁnd a large difference for socioemotional leadership. Contrary to our expectations, less relationship variance was found in self-ratings than in the ratings of others. Another explanation for the task and socioemotional differences in self-enhancement is that socioemotional leadership is more closely linked to group belongingness than to more observable task activities. In essence, members refrain from making judgments that decrease social acceptance so as to maintain their belongingness in a group (Anderson et al., 2006). Generally, for both types of leadership, the parameter estimates are not all dramatically different in the two types of ratings, self and other. There is little or no self-enhancement or effacement, and both k and q are close to 1. Although some differences in relationship variance are apparent, they arise only with socioemotional leadership. The bottom line is that it would not be especially problematic to include self-ratings in the measurement of the target and perceiver effects.

ESTIMATION OF SRM USING SPECIALIZED SOFTWARE If the SRM were so useful in the study of leadership, then why is it applied so infrequently? There are several reasons for its lack of popularity, but one of the most important is the difﬁculty in estimating the SRM variances and covariances using standard software, such as SPSS. In this section, we show how it is possible to estimate the model with such software. We warn the reader in advance that this section is much more technical than the rest of this chapter. In constructing our example, we consider the simple univariate model. We also use traditional SRM terminology, meaning that we replace ‘‘perceiver’’ with ‘‘actor’’ and ‘‘target’’ with ‘‘partner.’’ The SRM equation for actor i A Componential Analysis of Leadership Using the SRM 175 with partner j in group k is

Xijk ¼ mk þ aik þ bjk þ gijk where Xijk is the score for person i rating (or behaving with) person j, mk the group mean, aik person i’s actor effect, bjk person j’s partner effect, and gijk the relationship effect. The terms m, a, b, and g are random variables and 2 2 2 2 the following variances are parameters of the model: sm; sa; sb; and sg. The SRM also speciﬁes two different correlations between the SRM components of a variable, both of which can be viewed as reciprocity correlations:

At the individual level, a person’s actor effect can be correlated with that person’s partner effect and can be denoted as sab, generalized reciprocity. At the dyadic level, two members’ relationship effects can be correlated and can be denoted as sgg, dyadic reciprocity.

The ﬁnal parameter is the mean of mk or m, which is an overall intercept or ‘‘grand mean.’’ Thus, there are seven univariate SRM parameters: one mean, four variances, and two covariances.

ANOVA Estimation

Prior to the current work, almost all published papers used the method of moments – sometimes called the ANOVA (analysis of variance) method –to estimate these variances and covariances. This method is described in some detail in Kenny, Kashy, and Cook (2006). Basically, a term is computed, such as an estimate of the target effect; its variance is computed; and then the expected value of that variance is determined. This process is then repeated for the other SRM components. What results is a series of simultaneous equations, such that linear algebra can be used to solve for the unknown SRM variances and covariances from the variances and covariances of the estimated components. A computer program for the estimation of these components called SOREMO (SOcial RElations MOdel; Kenny, 1998) has been developed to estimate the SRM parameters from round-robin designs. In our example, we use data gathered by Lord et al. (1980). Their study, which involved a total of 24 four-person groups, measured how much each person in the group stated the other member contributed to the group, using a rating scale from 1 to 6. We denote this variable as LEAD. The data 176 DAVID A. KENNY AND STEFANO LIVI

Table 5. Summary of Results Using Different Programs.

Term Symbol SOREMO SAS Ia SPSSa SAS IIb SEM

Mean m 3.868 3.868 3.868 3.868 3.868 2 Actor variance sa 0.233 0.198 0.198 0.198 0.233 2 Partner variance sb 0.240 0.192 0.192 0.204 0.240 2 Group variance sm 0.091 0.000 0.000 0.000 0.094 c c Actor–partner covariance sab 0.059 0.000 0.000 0.024 0.059 2 Error variance sg 0.222 0.237 0.237 0.230 0.222 Error covariance sgg 0.014 0.032 0.032 0.022 0.014 aActor–partner covariance is assumed to be zero. bDummy variables with equality constraints. cFixed to zero. structure is a round-robin design, which has an n n structure in which the diagonal (self-ratings) is missing. Table 5 presents a summary of the results from the computer program SOREMO. We now consider how conventional software can estimate the SRM variances and covariances. Table 1 provides the SOREMO estimates of the seven parameters, which include a negative estimate of the group variance.

Conventional Multi-Level Modeling: SAS and SPSS

Increasingly, multi-level models can estimate models with cross-classiﬁed or crossed variables. In these models, the actor–partner covariance is assumed to be zero, which is a major limitation of this method. We describe this approach in three steps, and describe how both SAS and SPSS can be used to carry out the model’s calculations. Presumably, the program HLM could be used for this purpose as well.

Step 1: Data Organization and Preparation Create a data set in which each record is the response of one person about another person in a dyad on all variables (e.g., person A’s rating of person B on leadership, talkativeness, and intelligence). For a one round-robin of 5 persons, there would be 20 records, assuming that self-ratings are not included. Make sure the following variables are included in each record: A unique actor number: For example, for group 1, the actor numbers would range from 1 to 5, and for group 2, the actor numbers would range from 6 to 10. A Componential Analysis of Leadership Using the SRM 177

A unique partner number. A unique dyad number: For a ﬁve-person group, there are 10 dyads. A unique group number.

Step 2: Syntax We ﬁrst present the syntax for SAS, followed by the syntax for SPSS. Note that the actor–partner covariance is not modeled and that LEAD is the outcome of interest. The syntax for SAS is as follows: PROC MIXED COVTEST; CLASS ACTOR PARTNER DYAD GROUP; MODEL LEAD ¼ /S DDFM ¼ SATTERTH NOTEST; RANDOM INTERCEPT/TYPE ¼ VC SUB ¼ ACTOR; RANDOM INTERCEPT/TYPE ¼ VC SUB ¼ PARTNER; RANDOM INTERCEPT/TYPE ¼ VC SUB ¼ GROUP; REPEATED/TYPE ¼ CS SUB ¼ DYAD; The syntax for SPSS is as follows: MIXED LEAD BY GROUP /FIXED ¼ /PRINT ¼ SOLUTION TESTCOV /RANDOM INTERCEPT|SUBJECT(GROUP) COVTYPE(VC) /RANDOM INTERCEPT|SUBJECT(ACTOR) COVTYPE(VC) /RANDOM INTERCEPT|SUBJECT(PARTNER) COVTYPE(VC) /RANDOM INTERCEPT|SUBJECT(DYAD) COVTYPE(VC).

With SPSS, the REPEATED statement cannot be used for the dyad, so one must presume that the dyadic covariance is positive. Also, for SPSS, error variance equals the dyad variance plus the error variance, and the dyadic correlation equals the dyad variance divided by the sum of the dyad variance plus the error variance. As can be seen in Table 5, the SAS and the SPSS programs yield the same results. These outcomes would have been different if the reciprocity covariance were negative. In that case, it would be estimated as zero by SPSS and properly estimated by SAS. Note also that the estimates are different from SOREMO. The major reason for this difference is the assumption of the zero actor–partner covariance. Because that covariance is small, the differences are small. As discussed earlier in this chapter, in the 178 DAVID A. KENNY AND STEFANO LIVI study of leadership, the actor–partner covariance is typically small, so this method can be used fruitfully.

SAS with Dummy Variables

An approach to estimating the full SRM was initially proposed by Snijders and Kenny (1999) and illustrated using the program MLwiN. Here, we apply their approach but use the package SAS. Currently, this approach cannot be applied using SPSS and HLM. We warn the reader that this approach is very complicated. With this approach, 2n dummy random variables are created and constraints are placed on the variance–covariance matrix of those dummy variables. We describe their approach in three steps. We used this approach to analyze gender effects in the Lord et al. (1978) data earlier in this chapter. This method is most appropriate when explicit predictors (i.e., ﬁxed effects) of the SRM components are available.

Step 1: Data Organization and Preparation Create a data set where each record refers to a data point or n(n1) data points for each group, where n is the group size (assuming no self-data). For each observation, we have a variable that designates which group the person is in, which dummy is used, and which observation is relevant. We create two sets of dummy variables: A(1) through A(n), where n is the largest group size. For a dummy variable A(i), if the actor is person i, the dummy equals 1; otherwise, the dummy equals 0. P(1) through P(n), where n is the largest group size. For a dummy variable P(i), if the partner is person i, the dummy equals 1; otherwise, the dummy equals 0. The following SAS code might be used to create the dummy variables for a four-person round-robin in which there is a variable for actor and partner that goes from 1 to 4: A1 ¼ 0; A2 ¼ 0; A3 ¼ 0; A4 ¼ 0; IF ACT ¼ 1THENA1¼ 1; IF ACT ¼ 2THENA2¼ 1; IF ACT ¼ 3THENA3¼ 1; IF ACT ¼ 4THENA4¼ 1; A Componential Analysis of Leadership Using the SRM 179

P1 ¼ 0; P2 ¼ 0; P3 ¼ 0; P4 ¼ 0; IF PART ¼ 1 THEN P1 ¼ 1; IF PART ¼ 2 THEN P2 ¼ 1; IF PART ¼ 3 THEN P3 ¼ 1; IF PART ¼ 4 THEN P4 ¼ 1;

Step 2: Force Constraints This step is the least intuitive, but the most important part of the analysis. We need to place constraints on the variance–covariance matrix of the dummy variables. For instance, the covariances of the actor’s effects are set to zero. So far as we know, SPSS and HLM (but not MLwiN) do not allow for such constraints. To do so, we create a data file, in this case called G, to set the n actor variances (parameter 1) equal, the n partner variances (parameter 2) equal, and the n actor–partner covariances (parameter 3) equal. The file for a four- person group has the following structure: DATA G; INPUT PARM ROW COL VALUE; DATALINES; 1111 1221 1331 1441 2551 2661 2771 2881 3151 3261 3371 3481 4991; The four values in each record in the data file are input in the following order: Parameter number (where 1 refers to actor variance, 2 refers to partner variance, 3 refers to actor–partner covariance, and 4 refers to group variance). Row of the variance–covariance matrix. 180 DAVID A. KENNY AND STEFANO LIVI

Column of the matrix. Value in the matrix (almost always 1). For instance, the first line refers to actor variance in row 1 and column 1 and that value is set to 1. In summary, the first four lines set the actor variances equal, the next four lines set the partner variances equal, the next four lines set the actor–partner covariances equal, and the last line refers to the group variance. If an element is not referenced, then it is fixed to zero.

Step 3: Run the Multi-Level Model Following is the SAS code for PROC MIXED: PROC MIXED COVTEST; CLASS DYAD GROUP; MODEL LEAD ¼ /S DDFM ¼ SATTERTH NOTEST; RANDOM A1 A2 A3 A4 P1 P2 P3 P4 INTERCEPT /G SUB ¼ GROUP TYPE ¼ LIN(4) LDATA ¼ G; REPEATED/TYPE ¼ CS SUB ¼ DYAD(GROUP); Note that the ‘‘ldata ¼ g’’ statement in the RANDOM statement sets the equality constraints. There are nine (2nþ1) terms in the RANDOM statement, A1 through INTERCEPT, which are ordered as in G (actor ﬁrst, then partner, and ﬁnally group). If the model included predictor variables (e.g., gender), they would be included in the MODEL record after the equality sign, but before the slash. As seen in Table 5, the results obtained with SAS are similar to those yielded by SOREMO. We did not include the analysis using MLwiN, but the results are essentially the same estimates as those produced by SAS, even though the two programs use somewhat different estimation methods. This method is currently complex, but we would expect that it will eventually become simpler to use and be adapted to other packages (e.g., SPSS and HLM).

Structural Equation Modeling

The method of SEM is a generalization of the method developed by Olsen and Kenny (2006) for dyadic analysis. When we analyzed the relationship between self- and other perception earlier in this chapter, we used an extension of the SEM approach. This method is not especially difﬁcult, but it is messy and awkward. A Componential Analysis of Leadership Using the SRM 181

Step 1: Data Preparation Unlike in the other methods, the group is the unit of analysis with the SEM approach. If n members are in the largest group size, there would be n(n1) scores read per group. For n ¼ 4, the variables would be X12, X13, X14, X21, X23, X24, X31, X32, X34, X41, X42, and X43. Self-measures and other predictors could be included. The order of the variables does not matter, and scores can be missing. Normally, there must be at least two more groups (i.e., ‘‘individuals’’) than the number of variables to use SEM. However, Amos can accommodate in the case in which there are fewer groups than variables if one tells the program to allow for nonpositive deﬁnite input matrices. So far as we know, this option is not currently available in other SEM computer programs.

Step 2: Latent Variables There would be n actor factors, n partner factors, and one additional latent variable for the group. A given measure loads on its actor factor, a partner factor, and the general mean factor. For instance, the measure X32 loads on the actor factor for person 3, the partner factor for person 2, and the mean factor. All factor loadings are ﬁxed to 1. Parallel actor and partner effects would be correlated. Thus, the actor factor for person 1 would be correlated with the partner factor for person 1. All other factor correlations are ﬁxed to zero. Additionally, there would be correlations between pairs of errors for the two members of the dyad (e.g., the errors of X12 and X21).

Step 3: Equality Constraints To achieve an identiﬁed model, many equality constraints must be applied. The n(n1) means would be set equal, as well as the n actor variances, the n partner variances, the n(n1) relationship variances, the n actor–partner covariances, and the n(n1)/2 error covariances. Because a large number of equality constraints are made and many parameters are ﬁxed to 1 (i.e., all factor loadings), the total number of parameters in the model is only seven. Thus, if n is 4, there are 78 elements in the moment matrix, and 71 degrees of freedom in the model.

Step 4: Model Testing The fit of the model does not matter, because the fit varies depending who is assigned to persons 1, 2, 3, and so on. The model is treated as an I-SAT (i.e., interchangeable saturated) model, as described by Olsen and Kenny (2006). Nevertheless, specialized models (e.g., setting actor and partner variances 182 DAVID A. KENNY AND STEFANO LIVI equal) can be estimated, and the changes in fit in these restricted models from the saturated model can be ascertained. Also note that the estimates are maximum likelihood estimates, rather than restricted maximum likelihood estimates like those obtained in the multi-level modeling program. We used the program AmosBasic to estimate the model. Using AmosGraphics worked, but the model included so many variables and constraints that it was difficult to implement and verify its correctness. (This is why we described this approach as ‘‘messy.’’) We suggest always outputting the ‘‘implied moments,’’ to determine whether the constraints were successfully implemented.

Comparison of Different Methods

Each of the various methods has its own advantages and disadvantages. We believe that the SAS dummy variable estimates would be the same as those produced with SOREMO, which uses the ANOVA method when there are equal group sizes, no missing data, and all variances greater than or equal to zero. SEM results are slightly biased because the program uses a maximum likelihood estimation. Note that SOREMO and SEM do allow for negative variances (although some SEM programs can prevent negative estimates of variance). The multi-level program MLwiN offers the option of allowing for negative variances. In the case of our example, we obtain the value of 0.091 for the group variance with this approach. Using conventional software offers several advantages over using SOREMO. First, conventional software can handle the situation in which there are missing data. By contrast, SOREMO presumes that there are no missing data; if there are, either the entire group must be removed or some method must be used to impute missing values. Second, SOREMO requires that the group contain a minimum of four people. With these alternative methods, smaller-sized groups can be retained. Third, when group sizes are unequal, the results from different groups are optimally weighted. Fourth, one can estimate specialized models, such as a model that sets group variance to zero, a model that sets the actor–partner and relationship covariances to zero, or a model that sets the actor and partner variances equal. SOREMO estimates only the ‘‘saturated model’’ – that is, the model with all parameters. For instance, using SAS with dummy variables and setting the group variance to zero with our example yields actor variance of 0.1989, partner variance of 0.2056, actor–partner covariance of 0.04404, dyadic covariance of 0.03828, relationship variance of 0.2098, and intercept of 3.8640. A Componential Analysis of Leadership Using the SRM 183

The one major advantage of SOREMO is that it can estimate, in a single run, all the variances and correlations for a large number of variables. Perhaps, SOREMO might be used for initial analyses of the variables, and these other programs used once more speciﬁc models are available. The discussion here has considered only univariate models. Elsewhere, we have used the dummy variable approach with SAS to estimate a bivariate model (Kenny, West, & Cillessen, 2007). In this case, we constructed two additional dummy variables for the means of each variable and ﬁxed the error variance to a very small value, essentially zero. We have also used the SEM approach to estimate a path model in which the actor and partner effects caused self-ratings (Kenny & West, 2008), a method applied earlier in this chapter.

CONCLUSION

Although this chapter has highlighted many interesting and theoretically important uses of the SRM in the study of leadership, it has not discussed all of the possibilities. Here, we brieﬂy review other types of analyses that could be performed using this model.

Further Extensions

Two major topics were omitted from this chapter: group-level outcomes (e.g., group productivity) and data gathered from persons who are members of multiple groups or persons who are not members of the group (e.g., observers). We brieﬂy consider each of these topics here. Some of the major outcomes of analyses in the study of leadership in natural settings are at the level of the group: group productivity, cohesiveness, and type of team (Greguras, Robie, & Born, 2001). Group- level predictor variables are not problematic, they can just be added as predictors in the model. For instance, we considered a group-level predictor of gender composition when we discussed gender differences. Group-level outcomes, however, do present challenges. To study group outcomes, we can follow the pioneering work of Dabbs and Ruback (1987). They performed a separate SRM analysis for each group and correlated SRM variances and covariances with group-level outcomes. For instance, we could ask whether agreement about the target 184 DAVID A. KENNY AND STEFANO LIVI effect (i.e., the proportion of total variance that is due to the target) predicted group productivity. Sometimes, the same person may be a member of more than one group. One key question is whether a component varies across different group compositions and tasks. A rotation design (Kenny & Hallmark, 1992) places the same person in multiple groups with different persons; the analysis then considers whether the person is seen the same way in all of those groups. Using a rotation design, which involves a series of carefully constructed round-robin designs, Zaccaro et al. (1991) examined the extent to which the target effect correlated across four different tasks. In their study, the correlation of the target effect of leadership across four highly heterogeneous tasks was 0.59. More recently, Foti and Hauenstein (2007) used a rotation design to study leadership in a military context. They found that the correlation of the target effect of leadership across heterogeneous tasks was 0.44. It is also possible to examine the stability of perceiver and even relationship effects (If Alex is in a group with Bob, is it more likely that Alex will be a leader?). For more information about the rotation design, consult Kenny and Hallmark (1992). In our earlier example, we included data from only the members of the group. Often, data are available from other persons, besides those who are members of the group. For instance, others might observe or supervise the group. If multiple people observe each group member (as in studies of 360-degree feedback; Yammarino & Atwater, 1997), the SRM design called block design is applicable (Kenny, 1994). With this design, one set of persons constitutes the perceivers and the other set forms the targets. In addition, sometimes the group interacts with another group. If there are ratings of persons in the other group or out-group judgments, the study is said to have a block round-robin design. A very interesting question is whether someone who is seen as a leader of his or her own group is also seen as a leader by members of the other group. Boldry and Kashy (1999) have tested this interesting hypothesis and have found agreement in the perceptions of leadership across groups (r ¼ 0.52).

Correlations of Components

This chapter began with a traditional partitioning of variance into SRM components. If we are to understand the theoretical meaning of those components, we need to determine the correlations between variables. Earlier in this chapter, we illustrated the correlations between two random A Componential Analysis of Leadership Using the SRM 185 variables (task-oriented and socioemotional leadership), and the correlation between a fixed variable (gender) with a random variable (dyadic perceptions of leadership). Here, we provide examples of other potential correlations. One obvious possibility is to correlate behavior in the group with leadership perceptions. For instance, we could measure how much a person participated in the group and correlate that behavior with perceiver, target, and relationship effects in leader perceptions. Most likely, the strongest correlations would be the target effect, but we should be open to other possibilities. Almost certainly, the major focus in future research will be in the target effect, representing the tendency to be perceived by others as a leader. This component could be correlated with other external variables, such as personality variables. In this way, SRM allows us examine the correlation between this ‘‘true’’ leadership score with individual differences, such as extraversion, intelligence, and self-efficacy (Malloy & Janowski, 1992). In groups with formal leaders, a dummy variable could be used to code that variable, which could then be correlated with the target effect. In that way, we would learn how much of the random variance in consensual leadership is explained by the fixed variable of formal leadership. Perceiver effect could also be correlated with individual characteristics, much as we have done in this chapter with gender. We might also investigate correlations of followers’ personality traits (extraversion and agreeableness) with the perception of transformational leadership (Schyns & Felfe, 2007).

Final Comments

Although the SRM is a very powerful tool for the study of leadership, it is a tool that is often best left in the toolbox. That is, many questions of leadership do not require the SRM. In particular, when the data collected employ designs that are not round-robin (each member judging each other member), the SRM need not be employed. The SRM can be used to answer many important and neglected questions in the study of leadership, but it most surely does not answer every question. One problem with the SRM is the sheer volume of information in an analysis. If one measures only five variables in a round-robin study with self-ratings, for example, more than 100 correlations are possible – many of which are of little theoretical and empirical use. Thus, the researcher needs to focus on the variances and correlations that are particularly relevant to 186 DAVID A. KENNY AND STEFANO LIVI the research. For instance, we have seen that group variance and reciprocities seem to contain little information that is valuable for the study of leadership, so they might be safely ignored. The SRM paints a very complex picture of leadership perception because leadership perceptions are, indeed, very complex. For example, traditional leadership emergence studies use as the influence measure the mean rating given to the person by the other members of the group; that measure contains not just target effect but also relationship, perceiver, and group effects. The SRM applied to round-robin data greatly enlarges the leadership perception picture, disambiguating different sources of variance that could be explored either to gain a clearer picture of perception itself or to correlate each component level with external criteria. The analysis in this chapter has concentrated on the perception of leadership as perceived influence within group. Nonetheless, many other leadership-related measures (e.g., leadership styles) could be employed effectively, and with such measures, the SRM components could be analyzed. The SRM is not a simple method to employ. It requires complicated research designs, such as round-robin designs, complicated data structures, and difficult statistical analyses that require specialized software or conventional software to be used in unconventional ways. It also yields many complicated results. Nevertheless, these complications are, in some sense, required by the study of leadership. Leadership itself is very complicated, in that this construct has created great interest in the social and behavioral sciences. No doubt part of that interest derives from the fact that the construct operates simultaneously at multiple levels. As a means to help scholars both manage the integration theory challenge and comply with a multi-level view of leadership, SRM represents a valuable instrument for both theory development and testing. This methodological application and statistical tool is able to answer old questions and stimulate new ones. We encourage readers to tap its potential despite the fact that it requires a lot of work and a rethinking of many ideas. This hard work has the potential to deliver a large payoff because the model yields a mother lode of rich information.

ACKNOWLEDGMENTS

We thank Marialuisa Farnese, Tessa V. West, and Randi Garcia, who provided feedback on an earlier version of this chapter. We also thank A Componential Analysis of Leadership Using the SRM 187

Andrew Knight and Joseph Olsen, who provided helpful suggestions concerning software implementation. Signiﬁcant portions of this article were published in Leadership Quarterly, Vol. 19, Stefano Livi, David A. Kenny, Linda Albright, and Antonio Pierro, A social relations analysis of leadership, pp. 235–248, Copyright Elsevier 2008.

REFERENCES

Albright, L., & Forziati, C. (1995). Cross-situational consistency and perceptual accuracy in leadership. Personality and Social Psychology Bulletin, 21, 1269–1276. Albright, L., Kenny, D. A., & Malloy, T. E. (1988). Consensus in personality judgments at zero acquaintance. Journal of Personality and Social Psychology, 55, 387–395. Ambady, N., & Rosenthal, R. (1992). Thin slices of expressive behavior as predictors of interpersonal consequences: A meta-analysis. Psychological Bulletin, 111, 256–274. Anderson, C., Srivastava, S., Beer, J. S., Spataro, S. E., & Chatman, J. A. (2006). Knowing your place: Self-perceptions of status in social groups. Journal of Personality and Social Psychology, 91, 1094–1110. Atwater, L. E., & Yammarino, F. J. (1992). Does self–other agreement on leadership perceptions moderate the validity of leadership and performance predictions? Personnel Psychology, 45, 141–164. Atwater, L. E., & Yammarino, F. J. (1997). Self–other rating agreement: A review and model. Research in Personnel and Human Resources Management, 15, 121–174. Ayman, R., & Chemers, M. M. (1991). The effect of leadership match on subordinate satisfaction in Mexican organisations: Some moderating inﬂuences of self-monitoring. Applied Psychology: An International Review, 40, 299–314. Bales, R. F., & Strodtbeck, F. L. (1951). Phases in group problem-solving. Journal of Abnormal and Social Psychology, 46, 485–495. Bass, B. M. (1990). Bass and Stogdill’s handbook of leadership. New York: Free Press. Bass, B. M., & Avolio, B. J. (1993). Transformational leadership: A response to critiques. In: M. M. Chemers & R. Ayman (Eds), Leadership theory and research: Perspectives and directions (pp. 49–80). San Diego, CA: Academic Press. Biernat, M., & Kobrynowicz, D. (1997). Gender- and race-based standards of competence: Lower minimum standards but higher ability standards for devalued groups. Journal of Personality and Social Psychology, 72, 544–557. Boldry, J. G., & Kashy, D. A. (1999). Intergroup perception in naturally occurring groups of differential status: A social relations perspective. Journal of Personality and Social Psychology, 77, 1200–1212. Bonito, J., & Kenny, D. A. (2007). The measurement of reliability of social relations components from round-robin designs. Unpublished paper, the University of Arizona. Bono, J. E., & Judge, T. A. (2004). Personality and transformational and transactional leadership: A meta-analysis. Journal of Applied Psychology, 89, 901–910. Brown, D. J., & Lord, R. G. (2001). Leadership and perceiver cognition: Moving beyond ﬁrst order constructs. Mahwah, NJ: Lawrence Erlbaum Associates. 188 DAVID A. KENNY AND STEFANO LIVI

Calder, B. J. (1977). An attribution theory of leadership. In: B. M. Staw & G. R. Salancik (Eds), New directions in organizational behaviors (pp. 179–204). Chicago, IL: St Clair Press. Campbell, L., Simpson, J. A., Stewart, M., & Manning, J. G. (2003). Putting personality in social context: Extraversion, emergent leadership, and the availability of rewards. Personality and Social Psychology Bulletin, 29, 1547–1559. Chemers, M. M. (1997). An integrative theory of leadership. Mahwah, NJ: Lawrence Erlbaum Associates. Chemers, M. M. (2000). Leadership research and theory: A functional integration. Group Dynamics: Theory, Research, and Practice, Special Issue: One Hundred Years of Group Research, 4, 27–43. Cicero, L., Pierro, A., & van Knippenberg, D. (2007). Leader group prototypicality and job satisfaction: The moderating role of job stress and team identification. Group Dynamics: Theory, Research, and Practice, 11, 165–175. Cole, M. S., & Bedeian, A. G. (2007). Leadership consensus as a cross-level contextual moderator of the emotional exhaustion–work commitment relationship. Leadership Quarterly, 18, 447–462. Cook, W. L. (2001). Interpersonal influence in family systems: A social relations model analysis. Child Development, 72, 1179–1197. Curhan, J. R., & Pentland, A. (2007). Thin slices of negotiation: Predicting outcomes from conversational dynamics within the first five minutes. Journal of Applied Psychology, 92, 802–811. Dabbs, J. M., Jr., & Ruback, R. B. (1984). Vocal patterns in male and female groups. Personality and Social Psychology Bulletin, 10, 518–525. Dabbs, J. M., Jr., & Ruback, R. B. (1987). Dimensions of group process: Amount and structure of vocal interaction. In: L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 20, pp. 123–169). San Diego, CA: Academic Press. Dabbs, J. M., Jr., Ruback, R. B., & Evans, M. S. (1987). Grouptalk: Sound and silence in group conversation. In: A. W. Siegman & S. Feldstein (Eds), Nonverbal behavior and communication (2nd ed., pp. 501–520). Hillsdale, NJ: Lawrence Erlbaum Associates. Dansereau, F. J., Graen, G., & Haga, W. J. (1975). A vertical dyad linkage approach to leadership within formal organizations: A longitudinal investigation of the role making process. Organizational Behavior and Human Performance, 13, 46–78. Deaux, K. (1984). From individual differences to social categories: Analysis of a decade’s research on gender. American Psychologist, 39, 105–116. Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Mahwah, NJ: Lawrence Erlbaum Associates. Eagly, A. H., & Johnson, B. T. (1990). Gender and leadership style: A meta-analysis. Psychological Bulletin, 108, 233–256. Eagly, A. H., & Karau, S. J. (1991). Gender and the emergence of leaders: A meta-analysis. Journal of Personality and Social Psychology, 60, 685–710. Eagly, A., Makhijani, M., & Klonsky, B. (1992). Gender and the evaluation of leaders: A meta- analysis. Psychological Bulletin, 111, 3–22. Emrich, C. G. (1999). Context effects in leadership perception. Personality and Social Psychology Bulletin, 25, 991–1006. Epitropaki, O., & Martin, R. (2004). Implicit leadership theories in applied settings: Factor structure, generalizability, and stability over time. Journal of Applied Psychology, 89, 293–310. A Componential Analysis of Leadership Using the SRM 189

Fiedler, F. E. (1978). The contingency model and the dynamics of the leadership process. In: L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 11, pp. 209–225). New York: Academic Press. Foti, R. J., & Hauenstein, N. M. A. (2007). Pattern and variable approaches in leader emergence and effectiveness. Journal of Applied Psychology, 92, 347–355. Fuegen, K., & Biernat, M. (2002). Reexamining the effects of solo status for women and men. Personality and Social Psychology Bulletin, 28, 913–925. Greguras, G. J., Robie, C., & Born, M. P. (2001). Applying the social relations model to self and peer evaluations. Journal of Management Development, 20, 508–525. Hall, J. A., Coats, E. J., & LeBeau, L. S. (2005). Nonverbal behavior and the vertical dimension of social relations: A meta-analysis. Psychological Bulletin, 131, 898–924. Hemphill, J. K. (1950). Relations between the size of the group and the behavior of ‘‘superior’’ leaders. Journal of Social Psychology, 32, 11–22. Hoff, P. D. (2005). Bilinear mixed-effects models for dyadic data. Journal of the American Statistical Association, 100, 286–295. Hogg, M. A. (2001). A social identity theory of leadership. Personality and Social Psychology Review, 5, 184–200. Hogg, M. A. (2005). Social identity and leadership. In: D. M. Messick & R. M. Kramer (Eds), The psychology of leadership: New perspectives and research (pp. 53–89). Mahwah, NJ: Lawrence Erlbaum Associates. Hogg, M. A., & van Knippenberg, D. (2003). Social identity and leadership processes in groups. San Diego, CA: Elsevier Academic Press. Hollander, E. P. (1958). Conformity, status, and idiosyncrasy credit. Psychological Review, 65, 117–127. Hollander, E. P. (1992). The essential interdependence of leadership and followership. Current Directions in Psychological Science, 1, 71–75. Ho¨nekopp, J. (2006). Once more: Is beauty in the eye of the beholder? Relative contributions of private and shared taste to judgments of facial attractiveness. Journal of Experimental Psychology: Human Perception and Performance, 32, 199–209. House, R. J. (1971). A path–goal theory of leader effectiveness. Administrative Science Quarterly, 16, 321–338. House, R. J., & Shamir, B. (1993). Toward the integration of transformational, charismatic, and visionary theories. In: M. M. Chemers & R. Ayman (Eds), Leadership theory and research: Perspectives and directions (pp. 81–107). San Diego, CA: Academic Press. Kenney, R. A., Blascovich, J., & Shaver, P. R. (1994). Implicit leadership theories: Prototypes for new leaders. Basic and Applied Social Psychology, 15, 409–437. Kenny, D. A. (1994). Interpersonal perception: A social relations analysis. New York: Guilford. Kenny, D. A. (1998). SOREMO Version V.2. Unpublished manuscript, University of Connecticut, Storrs, CT. Kenny, D. A. (2004). PERSON: A general model of interpersonal perception. Personality and Social Psychology Review, 8, 265–280. Kenny, D. A., & Hallmark, B. W. (1992). Rotation designs in leadership research. Leadership Quarterly, 3, 25–41. Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis. New York: Guilford. Kenny, D. A., & La Voie, L. (1984). The social relations model. In: L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 18, pp. 142–182). Orlando, FL: Academic Press. 190 DAVID A. KENNY AND STEFANO LIVI

Kenny, D. A., & West, T. V. (2008). How is self-perception fundamentally different from the perception of others? Unpublished paper, University of Connecticut, Storrs, CT. Kenny, D. A., West, T. V., Cillessen, A. H. N., et al. (2007). Accuracy in the judgment of aggressiveness. Personality and Social Psychology Bulletin, 33, 1225–1236. Kerr, S. (1977). Substitutes for leadership: Some implications for organizational design. Organization and Administrative Sciences, 8, 135–146. Kerr, S., & Jermier, J. M. (1978). Substitutes for leadership: Their meaning and measurement. Organizational Behavior and Human Performance, 22, 375–403. Kwan, V. S. Y., John, O. P., Kenny, D. A., Bond, M. H., & Robins, R. W. (2004). Reconceptualizing individual differences in self-enhancement bias: An interpersonal approach. Psychological Review, 111, 94–110. Levine, J. M., & Moreland, R. L. (1998). Small groups. New York: McGraw-Hill. Livi, S., Kenny, D. A., Albright, L., & Pierro, A. (2008). A social relations analysis of leadership. Leadership Quarterly, 19, 235–248. Lord, R. G., Binning, J. F., Rush, M. C., & Thomas, J. C. (1978). The effects of performance cues and leader behavior on questionnaire ratings of leadership behavior. Organizational Behavior and Human Performance, 21, 27–39. Lord, R. G., Brown, D. J., & Harvey, J. L. (2001). System constraints on leadership perceptions, behavior, and influence: An example of connectionist level processes. In: M. A. Hogg & R. S. Tindale (Eds), Blackwell handbook of social psychology: Group processes (pp. 283–310). Oxford, UK: Blackwell. Lord, R. G., Brown, D. J., Harvey, J. L., & Hall, R. J. (2001). Contextual constraints on prototype generation and their multilevel consequences for leadership perceptions. Leadership Quarterly, 12, 311–338. Lord, R. G., De Vader, C. L., & Alliger, G. (1986). A meta-analysis of the relation between personality traits and leader perceptions. Journal of Applied Psychology, 71, 402–410. Lord, R. G., Foti, R. J., & De Vader, C. L. (1984). A test of leadership categorization theory: Internal structure, information processing, and leadership perceptions. Organizational Behavior and Human Performance, 34, 343–378. Lord, R. G., Phillips, J. S., & Rush, M. C. (1980). Effects of sex and personality on perceptions of emergent leadership, influence, and social power. Journal of Applied Psychology, 65, 176–182. Malloy, T. E., & Janowski, C. (1992). Perceptions and meta-perceptions of leadership: Components, accuracy and dispositional correlates. Personality and Social Psychology Bulletin, 18, 700–708. Meindl, J. R., Ehrlich, S. B., & Dukerich, J. M. (1985). The romance of leadership. Administrative Science Quarterly, 30, 78–102. Messick, D. M. (2005). On the psychological exchange between leader and follower. In: D. M. Messick & R. M. Kramer (Eds), The psychology of leadership: New perspectives and research (pp. 53–89). Mahwah, NJ: Lawrence Erlbaum Associates. Montgomery, B. (1986). An interactionist analysis of small group peer assessment. Small Group Behavior, 17, 19–37. Mullen, B., Salas, E., & Driskell, J. E. (1989). Salience, motivation, and artifact as contributions to the relation between participation rate and leadership. Journal of Experimental Social Psychology, 25, 545–559. Ng, H., Brooke, S., & Dunne, M. (1995). Interruption and influence in discussion groups. Journal of Language and Social Psychology, 14, 369–381. A Componential Analysis of Leadership Using the SRM 191

Norris-Watts, C., & Lord, R. G. (2004). Attribution processes. In: G. R. Goethals, G. J. Sorenson & J. M. Burns (Eds), Encyclopedia of leadership (pp. 1192–1195). Thousand Oaks, CA: Sage. Olsen, J. A., & Kenny, D. A. (2006). Structural equation modeling with interchangeable dyads. Psychological Methods, 11, 127–141. Phillips, J. S. (1984). The accuracy of leadership ratings: A cognitive categorization perspective. Organizational Behavior and Human Performance, 33, 125–138. Pierro, A., Mannetti, L., De Grada, E., Livi, S., & Kruglanski, A. W. (2003). Autocracy bias in informal groups under need for closure. Personality and Social Psychology Bulletin, 29, 405–417. Ridgeway, C. L. (2001). Bias in the evaluation of women leaders: Gender, status, and leadership. Journal of Social Issues, 57, 637–656. Ruback, R. B., Dabbs, J. M., Jr., & Hopper, C. H. (1984). The process of brainstorming: An analysis with individual and group vocal parameters. Journal of Personality and Social Psychology, 47, 558–567. Rush, M. C., & Russell, J. E. (1988). Leader prototypes and prototype-contingent consensus in leader behavior descriptions. Journal of Experimental Social Psychology, 24, 88–104. Schyns, B., & Felfe, J. (2007). The personality of followers and its effect on the perception of leadership: An overview, a study, and a research agenda. Small Group Research, 37, 522–539. Schyns, B., & Sanders, K. (2003, December 1). Mood and the evaluation of leaders. Current Research in Social Psychology, 9. Retrieved December 9, 2003, from http://www. uiowa.edu/%7Egrpproc/crisp/crisp.9.4.html Snijders, T. A. B., & Kenny, D. A. (1999). The social relations model for family data: A multilevel approach. Personal Relationships, 6, 471–486. Stogdill, R. M. (1948). Personal factors associated with leadership: A survey of the literature. Journal of Psychology, 25, 35–71. Tsui, A. S., Porter, L. W., & Egan, T. D. (2002). When both similarities and dissimilarities matter: Extending the concept of relational demography. Human Relations, 55, 899–929. Van Vugt, M. (2006). Evolutionary origins of leadership and followership. Personality and Social Psychology Review, 10, 354–371. Vargas, P. R. (1986). A social relations analysis of leadership perceptions among athletes. Unpublished master’s thesis, Sam Houston State University, Huntsville, AL. Vroom, V. H., & Yetton, P. W. (1973). Leadership and decision-making. Pittsburgh, PA: University of Pittsburgh Press. Yammarino, F. J., & Atwater, L. E. (1997). Do managers see themselves as others see them? Organizational Dynamics, 25, 35–44. Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. Leadership Quarterly, 16, 879–919. Yukl, G. (1994). Leadership in organizations (3rd ed.). Englewood Cliffs, NJ: Prentice Hall. Zaccaro, S. J., Foti, R. J., & Kenny, D. A. (1991). Self-monitoring and trait-based variance in leadership: An investigation of leader ﬂexibility across multiple group situations. Journal of Applied Psychology, 76, 308–315. Zaccaro, S. J., & Klimoski, R. J. (2001). The nature of organizational leadership: Understanding the performance imperatives confronting today’s leaders. San Francisco, CA: Jossey-Bass. Zebrowitz-McArthur, L., & Apatow, K. (1983–1984). Impressions of baby-faced adults. Social Cognition, 2, 315–342. This page intentionally left blank

CONSIDERATIONS IN APPLYING THE SOCIAL RELATIONS MODEL TO THE STUDY OF LEADERSHIP EMERGENCE IN GROUPS: A LEADERSHIP CATEGORIZATION PERSPECTIVE

Rosalie J. Hall, Robert G. Lord and Katey E. Foster

ABSTRACT

This commentary on Kenny and Livi (2009) expands on aspects of leadership categorization theory that are consistent with the application of the social relations model (SRM). It critically considers limits to the generalization of the variance components analysis results described in Kenny and Livi’s first example, and briefly summarizes results of an additional study that supports their findings. Finally, this commentary suggests interpretational issues of interest to researchers who wish to continue to apply the SRM to multi-level issues in the study of leadership.

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 193–213 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008009 193 194 ROSALIE J. HALL ET AL.

INTRODUCTION

We have long been fans of David Kenny’s social relations model (SRM), and believe that leadership researchers can gain from a thorough understanding of both the theoretical implications of the SRM and the unique design and analytical techniques developed or refined in association with it. For those who want a more in-depth treatment of issues raised by Kenny and Livi (2009), we heartily recommend both the Kenny (1994) book titled Interpersonal Perception: A Social Relations Analysis, and the more recent Kenny, Kashy, and Cook (2006) book titled Dyadic Data Analysis. In addition, readers may wish to consult the recent article appearing in The Leadership Quarterly by Livi, Kenny, Albright, and Pierro (2008). In this volume, Kenny and Livi cover substantial ground by introducing the topic and main ideas of the SRM, providing two illustrations of the potential benefits of applying the model to the analysis of existing data, and showing how to implement the variance components analysis techniques using commonly available statistical software packages. The SRM seems well suited to address issues of leadership perceptions. Questions surrounding leadership perceptions play an important role in the extensive literature on leadership in work or political organizations. The study of leadership perceptions focuses on factors determining the extent to which a given individual is seen by others to be a leader. Leadership perception is clearly important to the understanding of leader emergence (Lord, Brown, & Harvey, 2001) and to the effective maintenance of a leadership position (Brown & Lord, 2001; Epitropaki & Martin, 2005; Maurer & Lord, 1991). Especially if meta-perceptions are considered, self- perceptions as leaders may also potentially play a role in the ability of leaders to themselves develop and change (e.g., Lord & Brown, 2004; Lord & Hall, 2005). [Interestingly, the perception of leadership may – but does not necessarily – directly relate to one’s actual performance in the leader role as noted by Kaiser, Hogan, and Craig (2008).] To fully benefit from applications of the SRM, leadership researchers now need to carefully consider how the concepts and techniques of the general SRM approach mesh with the particulars of the theoretical and empirical work on the topic of leadership in organizations. This will likely mean thought and empirical research on the issue of how the particular group type and context might (or might not) influence the values of the variance components and correlations estimated as part of the SRM. For example, we will argue later in this chapter that although the empirical results from the reanalyses presented by Kenny and Livi provide a compelling ‘‘proof of A Leadership Categorization Perspective 195 concept’’ for applying SRM to the study of leader emergence, we do not at this time know the extent to which the generalization of their results is limited by the characteristics of the studies that were available to them. Among other things, the available studies involve artificial groups performing very short- term tasks and tend to involve student samples, which are relatively homogeneous in age and experience. First, however, we briefly describe our theoretical perspective on leadership. This discussion is followed by comments on the first empirical example presented by Kenny and Livi (2009). After that, we briefly describe an additional SRM study that two of the current chapter authors (Hall and Foster) are in the process of analyzing. Although this study suffers from some of the same boundary conditions as those reanalyzed by Kenny and Livi, the longitudinal design allows a focus on longer-term group interactions in a context where task performance has meaningful consequences. Finally, we summarize the implications for researchers.

LEADERSHIP CATEGORIZATION APPROACHES TO LEADERSHIP PERCEPTION

The theoretical leadership framework from which we evaluate the work of Kenny and Livi (2009) is a leader categorization approach (Lord, Foti, & Phillips, 1982; Medvedeff & Lord, 2007), which draws heavily from basic concepts in social and cognitive psychology. Past theory and empirical research suggests that perceiving leadership may occur through at least two general types of social-cognitive processes: recognition processes, which involve a match of perceived leader characteristics to the internal cognitive structures used by the perceivers in an effort to identify and categorize leaders, and inferential processes, in which the leadership (or lack of leadership) of a target individual is attributed based on the outcomes of important organizational events (Lord & Maher, 1990, 1991). Both of these perceptual processes imply the existence of both a perceiver and a target. Thus leadership perception may be understood as involving, at a minimum, two persons (later comments in this chapter will address the role of larger groups). We will refer in our discussion to a subordinate follower as the ‘‘perceiver,’’ and to either a peer or a hierarchical superior other individual as the ‘‘target.’’ Leadership categorization theory suggests that both of these individuals are influenced by internal cognitive structures relating to leadership. The leader categorization approach has evolved over time. Essentially, this approach emphasizes the importance of cognitive structures, variously 196 ROSALIE J. HALL ET AL. considered as implicit leadership theories (ILTs), schemas, prototypes, exemplars, and so forth, in the recognition and attribution of leadership. Within individuals, these cognitive structures develop over time and with experience (Lord & Maher, 1991). Leadership schemas influence both the behavioral observations of perceivers and the categorization of behavior as being leader-like (or not); they also influence the thoughts and actions of the target person who generates those behaviors. When leadership schemas are highly available, prototypical behaviors are easily encoded in terms of leadership. In contrast, when leadership schemas are inhibited – as may happen when inconsistent cues, such as being female, are present – prototypical behaviors are less easily categorized in terms of leadership (Scott & Brown, 2006). Similarly, schemas guide the behaviors of leaders and may be partially responsible for individual differences in transactional and transformational leadership (Wofford, Goodwin, & Whittington, 1998). Recent approaches to leadership schemas suggest (1) that they are not immutably fixed in content and structure, but rather are constituted anew in an accessible cognitive workspace at each activation, and (2) that they are context sensitive (Hanges, Lord, & Dickson, 2000; Lord & Brown, 2004; Lord et al., 2001). Thus somewhat different instantiations of a leadership schema are currently activated depending on numerous factors, including characteristics of the perceiver and the target such as activated identity level (e.g., individual vs. relational vs. collective), the social context, the task, and current goals. Research shows that certain personal characteristics or attributes are commonly associated with leadership and, therefore, are often present in a given individual’s leadership schema (Cronshaw & Lord, 1987; Epitropaki & Martin, 2004; Lord, Foti, & De Vader, 1984; Lord, De Vader, & Alliger, 1986; Offerman, Kennedy, & Wirtz, 1994). These characteristics include physical attributes such as physical attractiveness, height or being male, personality traits such as dominance or extraversion, and cognitive characteristics such as intelligence. In other words, whether or not these characteristics are actually associated with higher leadership performance, their mere presence increases the likelihood that a given individual will be perceived as a leader. Much empirical research has been done to abstract common elements of leadership schema, many of which show considerable generality across cultures (House, Hanges, Javidan, Dorfman, & Gupta, 2004). In addition, some individuals tend to see themselves as being leaders. For such individuals, leadership is a central – rather than peripheral – aspect of the self, and is used to guide both that person’s own behavior and the perceptions of others (Gradwohl, Lord, Brown, & Engle, 1996). Interestingly, A Leadership Categorization Perspective 197 individuals who rate high in leadership self-schema may be more stringent in evaluating the leadership status of others (Brown, Marchioro, Tan, & Lord, 1998), possibly because they hold more rigorous standards or have more complex prototypes that are more difficult to match. Leadership self- schema may also provide important cognitive structures that facilitate the organization and development of leadership skills, in part because seeing oneself as a leader helps one to proactively seek the leadership experiences required to develop leadership skills (Chan & Drasgow, 2001; Lord & Hall, 2005). With specific reference to the SRM, the leadership categorization approach provides theoretical explanations for a variety of phenomena of interest, including consensus across perceivers, dyadic uniqueness in perceptions, reciprocity correlations, and self-perceptions. For example, both consensus (or consistency) across different followers in leadership perceptions and some degree of stability in within-follower perceptions of leadership across time are explained by the presence of cognitive structures that share common elements across perceivers. Also, although schema may show some within-person change over time, they often do so in an incremental fashion or at least are instantiated somewhat similarly when there is consistency in environmental cues over time (Epitropaki & Martin, 2004). These same cognitive structures also help explain why agreement across observers is not perfect (e.g., observers will vary at least somewhat in their prior experiences of leadership and may have incomplete overlap in opportunities to observe targets) and why changes in perceptions may occur over time (e.g., changes in contextual cues, changes in the leadership self-schemas of perceivers that redefine their rating criteria). Different experiences can lead to person-to-person variability in leadership schema content. Also, aspects of schema such as their learned nature and sensitivity to context suggest the activated portion of the schema will differ over time within the same person. For example, Brown et al. (1998) found that observers changed from a more masculine leadership schema to a more feminine leadership schema when they repeatedly observed an experimental group in which the dominant leadership role shifted from a male target to a female target. Moreover, when this change was modeled with catastrophe analysis, an additional 20% of the variance in leadership was explained. These considerations imply that researchers will make the following observations:

Variability in leadership perceptions across perceivers due to consistent differences in leadership schema content 198 ROSALIE J. HALL ET AL.

Variability in the performance of behaviors (which may be generated in accord with idiosyncratic leadership self-schemas) across the targets of leadership perceptions Instability or change in leadership perceptions within an individual over time

In addition, negative generalized reciprocity correlations (in which an individual’s perceiver effect negatively relates to his or her target effect) might be expected given the tendency for persons who are self-schematic for leadership to themselves emit more leader-like behaviors while evaluating the leadership behaviors of others more harshly.

ADDITIONAL RELEVANT LEADERSHIP THEORY

With apologies to the large number of additional theoretical and empirical leadership works that could be tied into this topic, we briefly describe here two other theoretical frameworks that may be helpful to understanding the application of the SRM to leadership. First, the extensive literature on leader–member exchange (LMX) theory (and its earlier version known as vertical dyadic linkage theory) suggests the preeminence of dyadic relationships in leadership perceptions (Dansereau, Graen, & Haga, 1975; Graen & Scandura, 1987). From this perspective, the same leader is viewed as having multiple dyadic relationships that vary in quality across different leader–follower dyads. Thus LMX theory implies that perceptions of the same leader may vary systematically from one follower to another because of differences in the quality of their exchange relationship, even though some level of agreement may be apparent among the followers on the relative ranking of a given leader with respect to other target leaders. Empirical studies suggest that the quality of dyadic relationships becomes established fairly quickly as the dyadic relationship is formed, and that this quality tends to remain relatively stable over time (Liden, Wayne, & Stilwell, 1993). Recent work has also made links between LMX and ILTs (i.e., leadership schemas). For example, Epitropaki and Martin’s (2005) longitudinal study demonstrated the extent of match between an ideal leadership schema and perceptions of actual leader characteristics predicted LMX quality. Thus LMX theory provides a solid theoretical foundation for expecting what the SRM terms ‘‘relationship effects.’’ Relationship effects refer to variability in perceptions, affective reactions, or behaviors that is not A Leadership Categorization Perspective 199 explainable as simple perceiver or target effects. In other words, they reflect something about the unique dyadic relationship that has been established between two particular persons. These effects may be based on any number of factors, such as similarity in background or attitudes or an affective bond that develops over time. A leadership-related variable that might particularly be expected to show strong relational effects is liking (Liden et al., 1993), which may be subsumed in some LMX measures such as Liden and Maslyn’s (1998) affect dimension of the MDM. LMX theory also provides a theoretical basis for expecting positive dyadic reciprocity correlations, in which a correlation exists between the relationship effects of two paired members (e.g., if Bob likes Dave or perceives him as particularly leader-like, Dave has a similarly uniquely high perception of Bob). Second, the leadership literature suggests group-level influences may affect leadership perceptions, so that factors such as common experiences, shared contexts, social norms, group identities and goals, and communications among group members may lead to greater agreement in leadership perceptions within groups. For example, recent research that takes a social identity theory perspective indicates that with experience and increased group identification, groups develop their own unique prototype of an effective leader (Hogg, 2001). Thus, when members’ identification with the group is high, perceivers may rely more heavily on group-level prototypes than on general leadership prototypes (van Knippenberg, van Knippenberg, De Cremer, & Hogg, 2004). Such a process would likely produce larger target effects in leadership perceptions than when group identification is low. In sum, there is a strong theory-based logic for expecting leadership perceptions to show perceiver, target, and relational effects, as well as negative generalized reciprocity correlations and positive dyadic reciprocity correlations. At the same time, it is not clear to what extent the estimates of variance components from the Kenny and Levi application of the SRM may be generalized to naturally occurring groups, because of the nature of the studies available to them for reanalysis. We turn now to a general description of the characteristics of their studies as compared to groups of interest to organizational leadership researchers.

CONSIDERATIONS RELATED TO KENNY AND LIVI’S FIRST EMPIRICAL EXAMPLE

In their first illustrative analysis, Kenny and Livi (2009) reanalyzed data from seven studies that involved the collection of ratings either of overall 200 ROSALIE J. HALL ET AL. leadership or of a more circumscribed aspect of leadership perceptions or leadership-related behavior following a period of group interaction. All of these studies used a round-robin design, in which each group member rated each other group member. These studies were notable not only because of their round-robin designs (which are demanding to implement), but also because they provided leadership ratings on multiple measures. This design allowed for an assessment not only of the relative magnitudes of different sources of variance in leadership ratings (i.e., perceiver, target, and relationship), but also a determination of the stability of those effects across multiple measures. In brief, key empirical results found by Kenny and Livi were as follows: Groups do not vary in the total extent of leadership perceived. Target variance is substantial and has high reliability, indicating that perceivers tend to agree with one another in their perceptions of a specific other person’s leadership ability. Different measures of leadership tend to give similar results. Perceiver variance is much lower and not nearly as reliable. Relationship variance, which is the unique effect associated with a specific perceiver and a specific target, is moderate and approximately as reliable as the target variance. Kenny and Livi comprehensively demonstrate how the variance component and reciprocity correlation estimates are interpreted, and tie the results back to relevant leadership literature. In our comments here, we consider a slightly different issue. We are already convinced that this technique is valuable, and for the most part we agree with Kenny and Livi in their general interpretations of the meanings of the different SRM components. We would like to address the issue of what their specific estimates do and do not tell us about leadership in typical applied settings, and identify the implications of the limitations of that knowledge for where leadership studies using this technique may want to go in the future. In other words, we consider the kind of studies that the leadership field needs to be more confident that SRM estimates of the variance components describe real-life leadership situations and that reveal the potential effects of contextual factors.

Group History and Composition

We begin by summarizing some salient aspects of the studies that went into the analysis. All seven studies were conducted in the laboratory, with college A Leadership Categorization Perspective 201 students as research participants. Six of the seven studies involved groups that had a very short ‘‘life’’ of less than one hour, and the seventh study (Campbell, Simpson, Stewart, & Manning, 2003) was only nominally better in this respect. (In the Campbell et al. study, participants were first assigned to a group in an initial session, and then a week later interacted for roughly one-half hour.) It is difficult to gauge from the written accounts of the studies whether the persons interacting in the groups knew one another previously. There may have been some degree of previous relationship, given that the participants may have been drawn from a subject pool that included preexisting acquaintances and friendships. (In fact, it is a good practice for researchers utilizing these types of study designs to collect a measure of prior acquaintance in case it is needed for explanation or as a covariate in later analyses.) In none of the studies did there appear to be an explicit expectation that the group members would continue to interact as an intact group in the future. The short-term nature of the group members’ acquaintances, especially coupled with the knowledge that the group itself would not continue to exist after the study ended, contrasts with most groups of interest to organizational researchers. Even those sorts of workgroups intentionally designed to have a short life (e.g., committees, project groups) typically exist substantially longer than an hour. In addition, even though members of these groups may not have met face-to-face until the group is constituted, their reputations may be known to one another before they even meet, and they may reasonably expect that there will be further interactions after the group is dissolved. Also, some of the same group members may interact in multiple groups that pursue differing goals, both within the same organization and in the larger community, thus bringing experiences with one another from other contexts. One implication of the inclusion of short-term groups is that the resulting ratings of leadership perceptions may primarily be driven by recognition processes, because there may have been fewer opportunities for performance information to accumulate and engender inferential processes. This could yield underestimates of relationship effects relative to target effects. It might also imply the potential for changes in the relative proportions of target and relationship variance components over time as information leading to inferences becomes more available. In addition, in contrast to the groups that are characteristically found in many applied settings, the groups in the analyzed studies appeared to be fairly homogeneous in age composition (and, as a consequence, probably in task-relevant experience). That is, the participants in most studies were 202 ROSALIE J. HALL ET AL. students, so in terms of age they were likely primarily in their late teens to mid-twenties. Replications in groups with greater heterogeneity in age and experience might result in changes in the estimates of both target and perceiver effects. For example, older and more highly experienced perceivers might have systematic tendencies to underrate the targets, thus resulting in larger perceiver variance components. There was less of an issue with homogeneity in terms of the gender composition of the groups. Depending on the study, the groups were either single gender (e.g., Campbell et al., 2003; Dabbs & Ruback, 1984) or mixed gender (Dabbs, Ruback, & Evans, 1987; Lord, Phillips, & Rush, 1980; Ruback, Dabbs, & Hopper, 1984); one study appears to have deliberately created groups that had balanced gender (Montgomery, 1986). Of the studies that assigned participants to groups without considering gender, we do not know what the gender balance was. This factor could influence the variance component results, as masculinity has been shown repeatedly to be a predictor of leadership perceptions (Eagly & Karau, 2002). For example, target variance may be increased in situations where a group contains a single male relative to groups with more balanced compositions. This result might also be consistent with Kenny and Livi’s reanalysis of the Lord et al. (1980) data set, in which they found a systematic effect of female perceivers perceiving greater amounts of leadership. Group sizes in the seven reanalyzed studies ranged from four to eight persons. Having undertaken research using round-robin designs ourselves, we can completely understand the attractiveness of using such relatively small group sizes simply for practical reasons. Collecting and analyzing data from larger groups requires extensive coordination and a large pool of research assistants. If group-level analyses are planned, it also means that a large number of individuals must participate to have a large enough N at the group level. Yet, although many naturally occurring organizational groups have sizes similar to those in the Kenny and Livi reanalysis, organizational researchers may also desire to have data from samples that provide information on leadership perceptions in larger groups, which may have different dynamics for leader emergence. Furthermore, no formally designated hierarchical status differences occurred among the members of the groups in the studies reanalyzed by Kenny and Livi, and the study designs did not appear to impose any formal role differences among group members or acknowledge expertise-based differentiations. In contrast, in many workgroups the authority of at least one member is legitimated externally, as signaled by a formal title such as supervisor or chair. In addition, both the designated group leader and other A Leadership Categorization Perspective 203 group members may be further differentiated in terms of specialized roles or knowledge. Adding even more to the complexity, groups composed of persons from different functional backgrounds may show shifts in attributed expert power as the task domain shifts over time. Both in the current chapter and in the Kenny and Livi chapter, the SRM analyses were conducted using methods that do not explicitly take such role or expertise differences into account. In some cases, a more appropriate approach might be to consider Kenny et al.’s (2006) SRM designs with roles.

Task and Reward Structure Considerations

The assigned group tasks in the seven reanalyzed studies may or may not have been perceived by research participants as meaningful. For example, these tasks included a ‘‘get acquainted’’ session (Dabbs & Ruback, 1984), the desert island survival task (Dabbs et al., 1987), a brainstorming task in which participants were asked to generate ideas for increasing tourism to the United States (Ruback et al., 1984), and other tasks of a similar nature. In addition, other than potentially as a source of research credit, there does not seem to have been much in the way of rewards or sanctions for the behaviors exhibited by group members. In contrast, in a typical work setting, a variety of factors might make task meaningfulness much higher. In real work settings, performance on a task may be important for several reasons. For example, performance may be seen as reflecting on one’s career identity or work reputation, people may work at jobs or tasks that are high in intrinsic motivation, or cultural and normative forces may help define tasks as meaningful to employees. Given these possibilities, we wonder if the research contexts may have cued more social loafing than might be expected in tasks occurring in an organizational setting. If so, this behavior could have the effect of increasing the variance in target effects. Also, in real-world workgroups, one’s superior typically has the formal authority to dispense rewards and punishments, such as pay increases, promotions, sanctions, or firing decisions. Other group members may also have some control over rewards and punishments, especially those of a social nature (e.g., acceptance by the group or ostracism, assignment or division of favorable vs. unfavorable tasks). We also note that an important aspect to consider is whether group members’ performance will be evaluated and rewarded on an individual basis or a group-level basis. Individual-level rewards tend to create individual-level identities and qualitatively different 204 ROSALIE J. HALL ET AL. social exchanges than the collective identities that result from a practice of bestowing group-level rewards (Flynn, 2005). Collective identities may create strong norms for cooperation among individual group members and strong norms for leaders to treat all members alike, whereas individual-level identities promote dyadic differentiation in leadership behavior and more competition among group members. As a consequence, reward structures may indirectly affect the behavior of group members and leaders. Further, because collective identities highlight the importance of group norms, as noted by social identity theory, they result in a different basis for leadership perceptions than individual identities (Hogg, 2001; van Knippenberg et al., 2004).

Miscellaneous Considerations

All of the reanalyzed studies were conducted in a laboratory setting, which itself may have influenced the leadership rating process. For example, perhaps none of the group members in the study was perceived as particularly leader-like compared to the researcher who conducting the study (this might have washed out group-level effects). Also, several of the studies included the presence of an evaluator (Campbell et al., 2003) or use of recording devices such as microphones or video cameras (Dabbs & Ruback, 1984; Dabbs et al., 1987; Ruback et al., 1984; Lord et al., 1980). These cues may have elevated participant self-awareness and influenced their behaviors, as they bear a similarity to experimental inductions of increased self-focus. This could produce greater perceiver effects if raters’ leadership self-schema become activated due to increased self-focus. Increased self-focus may result in more severe ratings of others by persons who are more leadership schematic (Brown et al., 1998). Such awareness-heightening cues may or may not be present in natural settings. Another design consideration is the choice of leadership measures used in the seven reanalyzed studies. Those measures varied in nature from direct ratings of leadership perceptions to ratings of likeability, judgments of smiling and friendliness, desire to interact with the target, judgments of task performance, influence, and openness. There is room for debate about the extent to which some of these measures reflect leadership perceptions per se or are more appropriately considered antecedents and/or consequences of leadership perceptions (e.g., likeability). Regardless, we might expect variance component estimates to differ systematically across some of these different measures. For example, we would expect that more affectively loaded A Leadership Categorization Perspective 205 measures such as liking would show higher proportions of relationship variance. Finally, all of the studies are embedded in a middle-class, mainstream American culture. We do not know how the results would differ (if at all) across different socioeconomic and cultural groups. For example, an increasing amount of work is focusing on distinctions in leadership in more individualistic versus more collectivistic cultures (House et al., 2004). It is not clear how these differences along the individualism–collectivism continuum might affect study outcomes. For example, more collectivistic settings might be associated with decreased target variance because information- processing schemas that direct attention to context rather than focal individuals are more common in such environments (Kuhnen, Hannover, & Schubert, 2001). Conversely, more collectivistic settings might be characterized by more collectively held norms about authority associated with target characteristics such as age, gender, family background, and position, which might have the opposite effect of increasing target variance. In sum, the groups in the studies summarized by Kenny and Livi differed in many characteristics from the naturally occurring applied groups to which leadership researchers would like to generalize their results. Their results likely very closely describe population values for the variance components and reciprocity correlations that might be expected from relatively small, homogeneous groups in which people have very short-term acquaintance and work at tasks with few serious personal consequences. We do not know for certain how such differences between these groups and the groups more typically found in real-world organizations might affect the values of the estimated variance components and reciprocity correlations for studies conducted either in organizational settings or in other artificial settings that have greater fidelity. Some theoretical considerations would suggest that Kenny and Livi’s results might be fairly accurate estimates of leadership perceptions in many different types of groups. These considerations include the rapidity with which LMX relationships become established, perceivers’ reliance on recognition processes that depend on fairly easily observable and stable surface characteristics of targets, and cumula- tive, self-perpetuating effects of the relationship between attributed leadership status and social power. At the same time, changes in task and environmental demands, the development of social processes including identities over time, and other potential factors that were discussed suggest that their results may be limited in generalizability. Studies of more realistic groups can help us determine the boundary conditions of Kenny and Livi’s results. 206 ROSALIE J. HALL ET AL.

Because we believe that the extended acquaintance of group members potentially may inﬂuence SRM variance component and reciprocity correlation estimates, we next present results from the analysis of an additional data set that repeatedly assessed leadership perceptions. Although it is only one sample with a relatively small N, the results may help us begin to gauge the extent to which Kenny and Levi’s results can be generalized from ad hoc to ongoing leadership processes.

BRIEF DESCRIPTION OF ADDITIONAL SRM STUDY OF LEADERSHIP

This section briefly reports partial results from a study of leadership perceptions that used an extended longitudinal design and a realistic task in which the group level of performance has meaningful implications for group members. These data were collected from 95 students in three sections of a Social Psychology course at the University of Akron. An initial measurement period involved the collection of individual difference measures early in the semester before groups were formed. It was followed by six additional data collections of perceptions of group members (and other variables) that occurred roughly every 2–3 weeks across the 15-week semester. Members of the class were randomly assigned to four- or five-person groups, and the groups worked on an assigned class project that accounted for 25% of students’ final course grade. The project involved the design, conduct, analysis, and written report of the results of a simple psychological study on a topic related to the course material. Here we report the results of two sets of analyses of the composite global leadership impression (GLI) score collected in this study. The GLI (Cronshaw & Lord, 1987) is a short, cognitively oriented scale measuring leadership categorization. One set of analyses was based on the 77 persons in 20 groups with complete or virtually complete data. The group sizes for these analyses ranged from 3 to 5 persons (some initial 4- and 5-person groups experienced dropout). The analyses were then repeated using only the 45 persons who were in 10 groups of size 4 or 5. The second set of analyses was performed to estimate the relationship (dyadic) reciprocity correlation, which cannot be calculated for groups of 3 or fewer persons. Recall that the studies reanalyzed by Kenny and Livi were chosen at least in part because they used multiple leadership measures, allowing for an estimate of composite leadership. As a consequence, these authors’ A Leadership Categorization Perspective 207 reanalysis was able to partition variance in leadership perceptions into stable and unstable components, by looking for consistency across different measures. We used a similar analytic strategy, except the repetition of measurement was over time rather than across different measures of leadership. In this study, ‘‘stability’’ refers to consistency in perceptions across time and across different indicators of leadership in the Kenny and Livi results. Despite some design differences between this study and the studies included in the Kenny and Livi reanalyses, the overall partitioning of variance into actor, partner, and relationship components showed very similar results (see Table 1). The largest variance component was stable target variance, which accounted for 48% of the variance in our full set of 20 groups, 47% for the subset of groups with more than 3 persons, and 43% averaged across the seven studies reanalyzed by Kenny and Livi. The ratios of stable to unstable target variance were quite similar across studies, even though the Kenny and Livi samples represent stability across measures while our sample assesses stability across measurement periods. Also, in both the Kenny and Livi reanalyses and our sample, relationship variance was the second largest component, with estimates tightly clustered between

Table 1. Estimation of Social Relations Model Variance Components and Reciprocity Correlations: Comparison of Studies.

Relative Variance Partitioning Reciprocity Correlation

Perceiver Target Relationship Generalized Dyadic

Current (our) study All groups – Stable component .15 .48 .15 .24 – Unstable component .01 .06 .15 .56 – Groups W3 persons Stable component .12 .47 .17 .09 .29 Unstable component .01 .05 .17 .30 .20

Kenny and Livi reanalysis Average of seven studies .14 .03 Stable component .09 .43 .19 Unstable component .04 .04 .21

Notes: Stability is across time using same measurement method (GLI) for our current study but is across multiple measures administered at the same time for the Kenny and Livi reanalyses. The number of groups is 20 in the ﬁrst set of our analyses and 10 in the second set of our analyses. Dyadic reciprocity correlations were not estimated in our full current sample because of the inclusion of groups with three persons. 208 ROSALIE J. HALL ET AL.

.15 and .21 for both the stable and unstable components. Although the proportion of stable to unstable variance differed a bit between our sample and the Kenny and Livi reanalyses, in both cases the perceiver variance estimates were smallest and of very similar magnitudes across the studies, although our study did show larger perceiver effects than the Kenny and Levi reanalyses. Finally, the results of our study supported the expected positive dyadic reciprocity correlation. This correlation indicated that there was reciprocity in ratings between dyad members after controlling for the ‘‘main effects’’ of perceiver and target. In total, these results are consistent across our and the Kenny and Livi analyses, especially given the variability in measurement and contexts across the studies. In contrast to the results obtained by Kenny and Livi, our sample may have shown meaningful group-level variance. Kenny and Livi do not explicitly describe how this component was calculated, so we do not directly reproduce their analysis here. Nevertheless, in our sample, the composite GLI values for each of the 20 groups, going across group members and across time, ranged from 2.28 to 3.89. (The composite value of the GLI across all groups was 3.01.) The GLI response scale was a 5-point Likert- type format, so this range seems meaningful. One expectation that was not supported was the idea that greater amounts of variance would be associated with the relationship component in our study. This may reflect the use of a more cognitively oriented measure of leadership (the GLI) or, alternatively, as Liden et al. (1993) found, dyadic effects may occur relatively early in leader–member exchanges and then remain quite stable. In support of the idea that more affectively laden measures may demonstrate greater relationship variance, in our data set a separate round-robin measure of perceived group contributions (e.g., helping others, contributing) showed a relationship variance component in the mid-.40’s, which was larger than that observed for the GLI. Another expectation that was not cleanly supported was generalized reciprocity correlation. We expected that the relationship between the ratings received by an individual and the ratings given by that individual would show a negative relationship, based on an argument that raters high in leadership self-schema would act more leader-like themselves, but would give more stringent ratings. When the generalized reciprocity correlation was estimated in the entire sample, which included groups consisting of three, four, and five persons, the relationship was unexpectedly positive. By contrast, when it was estimated in a subsample containing only groups of four or five persons, the correlation was negative, although the bulk of it was an unstable component. We do not know at this point if these A Leadership Categorization Perspective 209 contradictory results are artifactual or if they reflect different dynamics in the larger groups. The longitudinal nature of our study suggests a variety of additional interesting analyses, which we do not pursue here. For example, because the study involved project-focused groups, the relative contributions of different group members likely changed over time as different aspects of the class project became important. Likewise, the relative variance components may have shifted systematically in value. This issue could be addressed by moving from an inspection of the more aggregate-level results presented here (i.e., cumulated over time) to a consideration of the changes in variance components across the six time periods. Yet, despite these differences in tasks, motivation, and time, the most noteworthy result in Table 1 is the overall similarity to the results reported by Kenny and Levi. These similarities provide added confidence in their results.

CONCLUSION

In this final section, we raise (but do not answer) a few more theoretical issues that we feel deserve a more extended discussion among researchers of leadership and the SRM. First is the issue of how observed target effects should be interpreted. Target effects initially may appear to be a function solely of the individual characteristics of the target, which may include physical appearance, personality, and behaviors. However, further consideration suggests that the interpretation of the level of these target effects as purely due to individual characteristics is not so straightforward. For example, some aspects of the rating of the target will likely be affected by group-level processes and issues that vary with the nature of the group. For example, the development of a strong group identification may supplant the influence of a general (individual-level) leadership prototype (e.g., Hogg, 2001; van Knippenberg et al., 2004). Alternatively, group consensus might arise as a result of discussion among group members about their current (and potentially future) supervisor and about events in which other group members may have demonstrated the presence or absence of leadership. These factors could increase target effects, but would be interpreted by many leadership researchers as reflecting the existence of a group-level phenomenon. If the group has existed for an extended period of time, these sorts of group-level effects on target variance potentially could be quite large. 210 ROSALIE J. HALL ET AL.

In addition, more dynamic models of prototype activation (Hanges et al., 2000; Lord et al., 2001) maintain that the prototype used by perceivers is inﬂuenced, in part, by target characteristics. For example, male targets may elicit more masculine prototypes than female targets, and vice versa. Because gender is easily recognized and is strongly associated with leadership prototypes, such target-based cuing of perceiver prototypes would likely produce consistent effects across perceivers and, therefore, would appear as a target effect. To the extent that variability in the gendered aspects of leadership prototypes exists across perceivers, this target-based cuing of prototypes would become manifest as a relational effect. Another issue concerns the potential reciprocal relationship between some leadership behaviors and leadership perceptions. It is well known that individuals who talk more tend to emerge as group leaders. For example, a review by Stein and Heller (1979) reports a very consistent pattern across studies of correlation between these two variables. At the same time, it is also known that individuals whose verbal contributions are not favorably acknowledged by the group reduce their subsequent verbalizations. Thus some behavioral measures of verbal participation may reﬂect leadership differentiation that has already occurred. This problem can be avoided by using trait-based measures known to be related to verbal ability, such as intelligence and extraversion (Judge, Bono, Ilies, & Gerhardt, 2002; Lord et al., 1986). In short, verbal contributions and leadership perceptions co-develop over time so that manifest verbal activity may be problematic because it is a confounded process explanation of leadership emergence. Finally, while it may be tempting to consider the unstable portions of variance components as measurement error, they may, in fact, be a mixture of error and other systematic effects that indicate the need for further dissection. For example, unstable components in a longitudinal design should be further probed to determine whether they signal systematic changes or development in the group process or whether they might be associated with changes in task or environmental demands on the group. These demands might include new tasks or changes in the nature of the existing task, crisis situations that increase the relevance of leadership to group processes (e.g., Emerich, 1999), and changes in the relative importance of task contributions versus group maintenance (e.g., socioemotional) activities. Unstable components in cross-sectional designs with multiple measures may indicate a need to carefully consider whether the existing set of measures fully and reliably sample the relevant leadership content domain. In sum, Kenny and Livi (2009) highlight exciting possibilities for increasing our understanding of leadership through the application of the A Leadership Categorization Perspective 211

SRM. Both the SRM theory and the accompanying analytic techniques can be important tools that allow us to meet the increasing demands of scholars in the leadership domain to explicitly consider multi-level theories and processes (e.g., Yammarino & Dansereau, 2008). As leadership researchers, our mission is to tailor aspects of the SRM to include considerations already present in leadership theory and empirical ﬁndings, and to allow us to move beyond our current state of knowledge.

ACKNOWLEDGMENTS

We would like to acknowledge the assistance of Jeanne Makinney and Christopher Marchioro, who were graduate students in the Industrial/ Organizational Psychology program of the University of Akron at the time the empirical study described in this chapter was conducted. They played key roles in the design and implementation of the study.

REFERENCES

Brown, D. J., & Lord, R. G. (2001). Leadership and perceiver cognitions: Moving beyond ﬁrst order constructs. In: M. London (Ed.), How people evaluate others in organizations: Person perception and interpersonal judgment in I/O psychology (pp. 181–202). Mahwah, NJ: Lawrence Erlbaum. Brown, D., Marchioro, C., Tan, J. A., & Lord, R. G. (1998, April). Individual difference variables and leadership perceptions: Linear or catastrophic relationship? Poster paper presented at the annual meeting of the Society for Industrial and Organizational Psychology, Dallas, TX. Campbell, L., Simpson, J. A., Stewart, M., & Manning, J. (2003). Putting personality in social context: Extraversion, emergent leadership, and the availability of rewards. Personality and Social Psychology Bulletin, 29, 1547–1559. Chan, K. Y., & Drasgow, F. (2001). Toward a theory of individual differences and leadership: Motivation to lead. Journal of Applied Psychology, 86, 481–498. Cronshaw, S. F., & Lord, R. G. (1987). Effects of leader prototypicality, consensus information, and encoding processes on leadership perceptions. Journal of Applied Psychology, 72, 97–106. Dabbs, J. M., Jr., & Ruback, R. B. (1984). Vocal patterns in male and female groups. Personality and Social Psychology Bulletin, 10, 518–525. Dabbs, J. M., Jr., Ruback, R. B., & Evans, M. S. (1987). ‘‘Grouptalk’’: Sound and silence in group conversation. In: A. W. Siegman & S. Feldstein (Eds), Nonverbal behavior and communication (2nd ed., pp. 501–520). Hillsdale, NJ: Lawrence Erlbaum. Dansereau, F. G., Graen, G., & Haga, J. (1975). A vertical dyad linkage approach to leadership within formal organizations: A longitudinal investigation of the role making process. Organizational Behavior and Human Performance, 13, 46–78. 212 ROSALIE J. HALL ET AL.

Eagly, A. H., & Karau, S. J. (2002). Role congruity theory of prejudice toward female leaders. Psychological Review, 109, 573–598. Emerich, C. G. (1999). Context effects in leadership perception. Personality and Social Psychology Bulletin, 25, 991–1006. Epitropaki, O., & Martin, R. (2004). Implicit leadership theories in applied settings: Factor structure, generalizability, and stability over time. Journal of Applied Psychology, 89, 293–310. Epitropaki, O., & Martin, R. (2005). From ideal to real: A longitudinal study of the role of implicit leadership theories on leader–member exchanges and employee outcomes. Journal of Applied Psychology, 90, 659–676. Flynn, F. J. (2005). Identity orientation and forms of social exchange in organizations. Academy of Management Review, 30, 737–750. Gradwohl, W., Lord, R. G., Brown, G., & Engle, E. (1996, October). Leadership self-schemas and their effects on leader impressions. Paper presented at the Society for Organizational Behavior, Pittsburgh, PA. Graen, G., & Scandura, T. A. (1987). Toward a psychology of dyadic organizing. Research in organizational behavior, 9, 175–208. Hanges, P. J., Lord, R. G., & Dickson, M. W. (2000). An information processing perspective on leadership and culture: A case for connectionist architecture. Applied Psychology: An International Review, 49, 133–161. Hogg, M. A. (2001). A social identity theory of leadership. Personality and Social Psychology Review, 5, 184–200. House, R. J., Hanges, P. J., Javidan, M., Dorfman, P. W., & Gupta, V. (2004). Leadership, culture, and organizations: The GLOBE study of 62 societies. Thousand Oaks, CA: Sage. Judge, T. A., Bono, J. E., Ilies, R., & Gerhardt, M. W. (2002). Personality and leadership: A qualitative and quantitative review. Journal of Applied Psychology, 87, 765–780. Kaiser, R. B., Hogan, R., & Craig, S. B. (2008). Leadership and the fate of organizations. American Psychologist, 63, 96–110. Kenny, D. A. (1994). Interpersonal perception: A social relations analysis. New York: Guilford Press. Kenny, D. A., Kashy, D. A., & Cook, W. L. (2006). Dyadic data analysis. New York: Guilford Press. Kenny, D. A., & Livi, S. (2009). A componential analysis of leadership using the social relations model. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 Research in Multi-Level Issues. Bingley, UK: Emerald. Kuhnen, U., Hannover, B., & Schubert, B. (2001). The semantic–procedural interface (SPI) model of the self: The role of self-knowledge for context-dependent versus context- independent modes of thinking. Journal of Personality and Social Psychology, 80, 397–409. Liden, R. C., & Maslyn, J. M. (1998). Multidimensionality of leader–member exchange: An empirical assessment through scale development. Journal of Management, 24, 43–72. Liden, R. C., Wayne, S. A., & Stilwell, D. (1993). A longitudinal study on the early development of leader–member exchanges. Journal of Applied Psychology, 78, 662–674. Livi, S., Kenny, D. A., Albright, L., & Pierro, A. (2008). A social relations analysis of leadership. Leadership Quarterly, 19, 235–248. Lord, R. G., & Brown, D. J. (2004). Leadership processes and follower self-identity. Mahwah, NJ: Lawrence Erlbaum. Lord, R. G., Brown, D. J., & Harvey, J. L. (2001). System constraints on leadership perceptions, behavior and inﬂuence: An example of connectionist level processes. A Leadership Categorization Perspective 213

In: M. A. Hogg & R. S. Tindale (Eds), Blackwell handbook of social psychology: Group processes (pp. 283–310). Oxford, UK: Blackwell. Lord, R. G., De Vader, C. L., & Alliger, G. M. (1986). A meta-analysis of the relation between personality traits and leadership perceptions: An application of validity generalization procedures. Journal of Applied Psychology, 71, 402–410. Lord, R. G., Foti, R. J., & De Vader, C. L. (1984). A test of leadership categorization theory: Internal structure, information processing, and leadership perceptions. Organizational Behavior and Human Performance, 34, 343–378. Lord, R. G., Foti, R. J., & Phillips, J. S. (1982). A theory of leadership categorization. In: J. G. Hunt, U. Sekaran & C. Schrieshiem (Eds), Leadership beyond establishment views (pp. 104–120). Carbondale, IL: Southern Illinois University Press. Lord, R. G., & Hall, R. J. (2005). Identity, deep structure and the development of leadership skill. Leadership Quarterly, 16, 591–615. Lord, R. G., & Maher, K. J. (1990). Leadership perceptions and leadership performance: Two distinct but interdependent processes. In: J. Carroll (Ed.), Applied social psychology and organizational settings (pp. 129–154). Hillsdale, NJ: Lawrence Erlbaum. Lord, R. G., & Maher, K. J. (1991). Leadership and information processing: Linking perceptions and performance. London, UK: Unwin-Hyman. Lord, R. G., Phillips, J. S., & Rush, M. C. (1980). Effects of sex and personality on perceptions of emergent leadership, influence and social power. Journal of Applied Psychology, 65, 167–173. Maurer, T. J., & Lord, R. G. (1991). An exploration of cognitive demands in group interaction as a moderator of information processing variables in perception of leadership. Journal of Applied Social Psychology, 21, 821–840. Medvedeff, M. E., & Lord, R. G. (2007). Implicit leadership theories as dynamic processing structures. In: B. Shamir, R. Pallai, M. C. Bligh & M. Uhl-Bien (Eds), Follower-centered perspectives on leadership: A tribute to the memory of James R. Meindl (pp. 19–50). Greenwich, CT: Information Age Publishing. Montgomery, B. M. (1986). An interactionist analysis of small group peer assessment. Small Group Behavior, 17, 19–37. Offerman, L. R., Kennedy, J. K., Jr., & Wirtz, P. W. (1994). Implicit leadership theories: Content, structure, and generalizability. Leadership Quarterly, 5, 43–58. Ruback, R. B., Dabbs, J. M., & Hopper, C. H. (1984). The process of brainstorming: An analysis with individual and group vocal parameters. Journal of Personality and Social Psychology, 47, 558–567. Scott, K., & Brown, D. J. (2006). Female first, leader second: Gender bias in the encoding of leadership behavior. Organizational Behavior and Human Decision Processes, 101, 230–242. Stein, R. T., & Heller, T. (1979). An empirical analysis of the correlations between leadership status and participation rates reported in the literature. Journal of Personality and Social Psychology, 11, 1993–2002. van Knippenberg, D., van Knippenberg, B., De Cremer, D., & Hogg, M. A. (2004). Leadership, self, and identity: A literature review and research agenda. Leadership Quarterly, 15, 825–856. Wofford, J. C., Goodwin, V. L., & Whittington, J. L. (1998). A field study of a cognitive approach to understanding transformational and transactional leadership. Leadership Quarterly, 9, 55–84. Yammarino, F. J., & Dansereau, F. (2008). Multi-level nature of and multi-level approaches to leadership. Leadership Quarterly, 19, 135–141. This page intentionally left blank

THOUGHTS ON STUDYING LEADERSHIP IN NATURAL CONTEXTS

Stefano Livi and David A. Kenny

ABSTRACT

Hall, Lord, and Foster (2009) have commented on whether variance partitioning in the social relations model would be the same in long-term groups and when groups have formal leaders. This reply follows their lead and speculates on how the variance partitioning would change. It considers the design and analysis issues in natural workgroups as well as the estimation of group effects.

INTRODUCTION

We are most grateful that the editors asked Hall, Lord, and Foster (2009) to critique our earlier chapter. Our respect for these authors is clearly evident: We cited them nine times in our own chapter. Moreover, Dave Kenny worked closely with Bob Lord in the late 1970s in his ﬁrst social relations model (SRM) analysis of leadership data. He also worked with Rosalie Hall (as well as Fred Dansereau) in the 1980s on issues dealing with levels of analysis in organizational data. We ﬁnd Hall, Lord, and Foster’s commentary in this

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 215–224 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008010 215 216 STEFANO LIVI AND DAVID A. KENNY volume to be both thoughtful and instructive, and we are very gratiﬁed that they found potential in the use of the SRM to study leadership. We organize our comments in this reply around the following issues. First, we consider what SRM variance partitioning would be when studied in more naturalistic settings. We devote most of our attention to this issue. Second, we examine some design and analysis issues that arise especially in the study of non-laboratory groups. Third, we consider the estimation of group effects. In the remainder of this reply, we refer to the Hall et al. (2009) response as HLF and to our own paper as KL (Kenny & Livi, 2009).

VARIANCE PARTITIONING IN NATURAL SETTINGS

HLF raised a concern about whether the results found in KL would be replicated if we studied long-term groups in naturalistic settings. The prototypical group is a workgroup in an organization. We begin by reviewing what we found in our reanalysis of the original data. We reanalyzed seven laboratory studies using college students. We found that 48% of the variance is target variance, with this variance ranging from 30% to 60%. Thus nearly half of the variance in leadership perception is explained by consensus among members about who is the leader. The second largest component is the relationship (19%, ranging from 13% to 31%), showing that a consistent source of variance is explained by the unique evaluations among dyads within group. The perceiver variance, which refers to the group members’ personal evaluation of leadership within group, explains a small amount of variance (9%, ranging from 0% to 13%). Finally, the group appears to have no effect, indicating that leadership perceptions do not vary from group to group. Interestingly, the new longitudinal study reported by HLF basically replicates the average values of the variance partitioning that we report in Table 1 of KL. HLF speculate about changes in the variance partitioning that would occur in natural settings. In short-term, leaderless groups, leadership is attributed to members who exert significant influence over other members of the group, despite not having any formal authority over the members. Research applying SRM on laboratory groups may not generalize well to non-laboratory settings. As previously stated, the target is the largest variance component in laboratory studies. Clearly, most of the time, people agree about who is the most influential person and who is the least influential person within group. This result implies that the target effect deserves a great deal of attention in Studying Leadership in Natural Contexts 217 leadership perception research. We strongly agree with HLF that just because we find consensus in the perception of leadership, it does not necessarily imply that leadership is a trait. Leadership stereotypes and interpersonal processes can result in consensus. Nevertheless, the results from laboratory studies of leadership suggest that the focus should be the leader himself or herself and the process related to the leader perception, rather than the processes that focus on between-groups or within-follower perceptions. What would we expect in natural workgroup settings? We suspect that target variance might be even larger in naturally occurring groups. In other studies, we have found evidence for more consensus in task-oriented groups than in socioemotional groups (Livi, Kenny, Albright, & Pierro, 2008). In addition, other reasons that we might find greater consensus exist. First, as HLF pointed out, workgroups typically have formal designated hierarchical status differences among members, with these differences often being reinforced by specific titles, resources, and access to information. These asymmetries make leadership schemas highly available, such that prototypical leadership behaviors (such as talkativeness or interruptions) are easily encoded in terms of leadership. Moreover, the workgroup’s shared history, social norms, values, identities, and goals result in shared schemas that are highly available to all members. The formal authority structure of workgroups would, therefore, make it much easier to recognize the expected leadership behaviors. Moreover, as HLF point out, the group sizes in workgroups settings are considerably larger than the group sizes in laboratory studies. We have previously found that in leaderless groups, target variance increases with group size as asymmetries in individual contributions become more salient in larger groups (Livi et al., 2008). Moreover, in larger groups, leadership is needed for the group to achieve its goals. Interestingly, HLF suggested that collectivistic culture (compared to individualistic culture) might have a decreased target variance because the information processing schemas of those settings direct attention to context rather than focus on the individual. Evidence supporting this view can be found in the study by Malloy, Albright, Diaz-Loring, Dong, and Lee (2004). In terms of the perception of personality, there was much less agreement or consensus in China than was found in either the United States or Mexico. Relationship variance, the second largest component in our review of studies, might reflect different workgroups’ social dynamics – a factor that might be more salient in natural settings. As VDL (subsequently relabeled LMX) theory (Dansereau, Graen, & Haga, 1975) suggests, in workgroups the leadership role development process can vary in duration from brief to 218 STEFANO LIVI AND DAVID A. KENNY extended, but usually time is required for dyadic relationships to develop. Thus within groups, the exchange of rewards for effort, for good performance, or for reciprocal social support needs time to mature. Furthermore, long-term groups often develop personal preferences among particular members or alliances among two or more members, leading to the formation of subgroups or minorities. Each faction has its own internal hierarchical structure, leading to an increased relationship variance within the group as a whole. Other intra-group phenomena, such as conflict or competition among members, could favor dyadic relationship within groups, leading to ratings that are either too lenient to members’ friends or too severe toward their competitors (Greguras, Robie, & Born, 2001). We suspect that there would be less perceiver variance in organizational workgroups. Short-term, leaderless discussions raise a more personal representation of the leadership schema, which is increased by the vagueness of the experimental setting. Moreover, as HLF noted, self-awareness bias, which is typical of laboratory settings, would likely be reduced in naturally occurring groups. Alternatively, workgroups’ shared schemas or information about group goals and performance might enhance positive interdependence, reducing individual variability within group. Moreover, peer raters in workgroups may tend to rate other members similarly so as not to cause friction (Murphy & Cleveland, 1995). Nevertheless, we agree with HLF that variability in composition could lead in some cases to increased perceiver variance. For example, age, role, and gender heterogeneity, which is much greater in natural as opposed to laboratory settings with college students, would likely produce an increase in perceiver variance. The weakest component that emerged in terms of leadership partitioning is the group variance. We expect that this component would be more apparent in organizational settings. Workgroups structurally show between-group differences that couldlead to an increased group variance component – specifically, dissimilarity in formal leadership, such as assigned or expressed power to each group members, unequal participation in group members’ different leadership styles by the actual leader, or differences between groups in expertise level. At a group level, overall cohesion in groups could enlarge the overall influence of each member despite of the actual level of participation by individuals. Moreover, the overall level of members’ identification with their group may unequally allocate leadership perceptions among groups. The group component might be increased by leader similarity or dissimilarity to the group prototype or by shared beliefs about the need for specific abilities of the leader in special circumstances, such as crisis. Alternatively, we might Studying Leadership in Natural Contexts 219 expect a reduction of the group variance component in workgroups in which hierarchy is formally assigned, thus exacerbating the perception of leadership as a zero-sum phenomenon. Certain situational factors could decrease leadership perception (such as uncertainty or group failures) or, at the opposite extreme, increase leadership perception (such as successful groups or temporal pressure). Given that workgroups in organizations might differ on these variables, we would expect increased group variance in applied settings. Overall, we would expect to find group variance in non-laboratory settings. We return to those factors that affect group variance in a later section of this commentary. Finally, in organizational settings, we suspect that group size, formal hierarchy, and group history are factors that are likely to reduce error variance. Basically, our expectation is that much more systematic variation of critical factors would occur in non-laboratory settings, so error should play a much smaller role in studies carried out in this environment. Typical workgroups probably facilitate the activation of leadership schemas, varying especially the target and relationship component of perceptions. Short-term groups, such as those for which data were reviewed for this commentary, do not provide enough time to activate the stable information needed for leadership emergence, which tends to inflate error variance. HLF speculate that naturalistic groups may show more changes than do laboratory groups. Thus what might be considered ‘‘error’’ (i.e., unstable target variance) might actually reflect meaningful changes in leadership over time. As reported by Kenny (2004), perceptions of personality are remarkably constant over time, with the average correlation being .98. Thus we would be surprised if we found much change in leadership. One factor that might potentially inflate these stability correlations is the use of the same perceivers to rate personality at each time. If different perceivers were used at each time, researchers would have more opportunities to observe change. In organizational settings, we would expect to observe variation in both dyadic and generalized reciprocity correlations. We would expect negative reciprocity or compensation to increase when formal leaders have strong external legitimacy. Alternatively, for situations characterized by decentralized and democratic leadership, we would expect to see little or no reciprocity in the perceptions of members, as leadership is more widely distributed among members. In summary, we suspect that studies of naturalistic, long-term groups would reveal increased target, group, and relationship variances and reduced perceiver and error variances. Of course, that conclusion is just 220 STEFANO LIVI AND DAVID A. KENNY our guess. Research evidence might well tell a different story, as data all too often make liars out of all of us.

DESIGN AND ANALYSIS ISSUES

As discussed in KL, natural settings complicate the research design. In particular, one member of the group may be denoted as the formal leader. Thus there is role differentiation in the group. Moreover, in the natural setting, members may not rate leadership, but observers might be used. HLF also point out that when group sizes are large, as they often are in naturalistic contexts, gathering full round-robin ratings is infeasible. Fortunately, with the advent of modern methods of estimation of the SRM components, as reviewed in KL, a full round-robin design is not required. Perceivers need simply be asked about a random subset of targets – say, four items. With such a design, full information about SRM variances and covariances can still be estimated. One of the major goals of KL was to show researchers that the SRM, while being a model of random effects, could also be used to study fixed effects. Many of us were trained as psychologists and were taught to use analysis of variance to answer research questions. Most psychologists focus on fixed effects and pay little or no attention to the random parts of the model. In a complete analysis of data from groups, both random and fixed effects would be present. In fact, fixed effects can be reconceptualized to parallel the SRM random effects, as discussed by Kenny and Garcia (2008). Consider a simple fixed effect: similarity of gender. In measuring how much Jack likes Jill in the group, we can consider four different ways to conceptualize gender similarity in the group: Relationship: Are Jack and Jill the same gender? Perceiver: Is Jack’s gender the same as others, besides Jill, in the group? Target: Is Jill’s gender the same as others, besides Jack, in the group? Group: Are other members of the group similar to each other in terms of gender? These four types of similarity parallel the SRM components of relationship, perceiver, target, and group, respectively. To further complicate the discussion, these fixed effects may differ as a function of role – that is, leader versus follower. Kenny and Garcia (2008) Studying Leadership in Natural Contexts 221 present some suggestive evidence that similarity of the group members leads to liking of followers, but diversity leads to liking of leaders.

GROUP VARIANCE

We realized that after reading HLF that we were unclear about how to compute group variance. The specialized software that was created to estimate SRM variances and covariances, SOREMO, included group variances. However, because these variances were almost always very small and because at the time there were unresolved statistical issues related to the computation of such variances when group sizes varied, it was decided to drop the computation of such variances from that program. Moreover, more likely than not, the estimated variances were negative. For these reasons, SOREMO did not provide estimates of group-level variances and covariances. The appendix to this commentary provides the speciﬁc formulas for the computation of group variances, both stable and unstable. With the advent of multi-level modeling programs, these variances can be estimated. As stated in KL, they are almost always zero when we examine leadership data. More to the point, they are almost always estimated as negative. How are we to interpret these negative variances? We must begin by saying we are not exactly sure what the answer to this question is. We do have two ideas, however. First, leadership in a group might be thought of as a zero-sum process: The more that George is a leader, the less that Sally can be a leader. Consider the following points: We know that dominance predicts leadership (Lord, DeVader, & Alliger, 1986). That said, if the group contained members who were all very dominant, we would not expect a lot of leadership in that group. In a parallel fashion, if the group consisted of people who were all low in dominance, we would expect little or no leadership in the group. The statistical implication is that the target effect is not assumed to vary randomly from group to group. Rather, it might be assumed that the mean of the speciﬁc members’ target effects is zero in the group, not in the population. If this was the case, then we should remove the target variance and perceiver–target covariance terms in the formula for the estimate of group variance (see Appendix). The second alternative is even more complicated. It presumes that when a person judges others in the group, contrast effects arise. Thus the sum of relationship effects of a given perceiver is equal to zero. In this case, the 222 STEFANO LIVI AND DAVID A. KENNY

‘‘zero-sum’’ aspect of leadership is perceptual and not real. The SRM formulas could be modiﬁed to account for such a phenomenon.

CONCLUSION

As is evidenced by the consistency of the empirical results from our review of seven studies and the new data presented by HLF, we have a solid understanding of the components in the perception of leadership in leaderless, short-term, laboratory groups. We now need more analyses of ‘‘natural’’ groups to determine what happens to these components in complex settings such as organizational workgroups. We and HLF have speculated about such differences and have made suggestions for future research that applies to these settings. Moving from the laboratory to the ﬁeld creates design and analysis complications, and the interpretation of the variance components, as HLF suggested, should be carefully evaluated based on the theoretical level of analysis (Yammarino, Dionne, Chun, & Dansereau, 2005). Group, relationship, and individual levels can be viewed in combination or simultaneously, and cross-level effects need to be revealed using appropriate designs that control for those levels. Many variables operate at multiple levels of analysis. For instance, leadership prototypes theoretically may have some degree of variation at the individual level, but also have between-group variation as a reﬂection of the nature of the group. As KL noted, group- level variables and their interaction with individual-level variables are not problematic, as the group-level variables can just be added as predictors in the model. With this approach, cross-level effects could be estimated as long as higher-level constructs are assessed. We have the tools available that can handle many of the complications. We need to get busy and start using them!

REFERENCES

Dansereau, F., Jr., Graen, G., & Haga, W. J. (1975). A vertical dyad linkage approach to leadership within formal organizations: A longitudinal investigation of the role making process. Organizational Behavior and Human Performance, 13, 46–78. Greguras, G. J., Robie, C., & Born, M. P. (2001). Applying the social relations model to self and peer evaluations. Journal of Management Development, 20, 508–525. Hall, R. J., Lord, R. G., & Foster, K. E. (2009). Considerations in applying the social relations model to the study of leadership emergence in groups: A leadership categorization Studying Leadership in Natural Contexts 223

perspective. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Kenny, D. A. (2004). PERSON: A general model of interpersonal perception. Personality and Social Psychology Review, 8, 265–280. Kenny, D. A., & Garcia, R. (2008). Using the actor–partner interdependence model to study the effects of group composition. Unpublished paper, University of Connecticut, CT. Kenny, D. A., & Livi, S. (2009). A componential analysis of leadership using the social relations model. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Livi, S., Kenny, D. A., Albright, L., & Pierro, A. (2008). A social relations analysis of leadership. Leadership Quarterly, 19, 235–248. Lord, R. G., DeVader, C. L., & Alliger, G. (1986). A meta-analysis of the relation between personality traits and leader perceptions. Journal of Applied Psychology, 71, 402–410. Malloy, T. E., Albright, L., Diaz-Loring, R., Dong, Q., & Lee, Y. T. (2004). Agreement in personality judgments in non-overlapping social groups in collectivist cultures. Personality and Social Psychology Bulletin, 30, 106–117. Murphy, K. R., & Cleveland, J. N. (1995). Understanding performance appraisal: Social, organizational, and goal-based perspectives. Thousand Oaks, CA: Sage. Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. Leadership Quarterly, 16, 879–919.

APPENDIX. COMPUTATION OF GROUP VARIANCES

Assume that there are groups of size n, and for each group we compute the variance of the group means or Mk. The expected value of the variance of 2 these group means (sM) can be shown to equal

2 2 2 s þ s þ 2sab s þ sgg s2 þ a b þ g m n nðn 1Þ We can solve for the group variance by computing

2 2 2 s þ s þ 2sab s þ sgg s2 a b g M n nðn 1Þ

To partition the variance into stable and unstable and group variance, we need to estimate the covariance of the group effects across two measures. We take the group means of one variable and correlate them with the means of the other variables. The covariance in the means (sMM)u is the correlation times the product of the two standard deviations. The estimate of group 224 STEFANO LIVI AND DAVID A. KENNY

covariance or smmu is

saa0 þ sbb0 þ sab0 þ sba0 sgg0 þ sgg00 s 0 MM n nðn 1Þ where the prime notation refers the second variable. Thus saau refers to the covariance of the actor effect of one variable (a) with a second actor effect (au). Also sggv refers to the covariance of gij for one variable with gji for a second variable. If there are r measures, there are r variances and r(r 1)/2 covariances. We denote the average of the variances as V and the average of the covariances as C. The estimate of the stable group variance equals C, and the estimate of the unstable group variance equals V C. In the text, we consider changing assumptions of the model to treat some effects as ﬁxed. For instance, if we treat the target effect as ﬁxed, then the formula for the group variance becomes

2 2 s s þ sgg s2 a g M n nðn 1Þ

Because less is subtracted in the preceding formula, it is more likely that there would not be negative variance. To allow for treating relationship effects as ﬁxed within the perceiver is more complicated, and further work in this area is needed. Not only would it affect the estimate of the group variance, but it would also affect the estimates of all other variances in the model. We encourage the development of such a model.

PART IV LEADERSHIP SIMULATION

This page intentionally left blank

A LEVELS-BASED LEADERSHIP SIMULATION: INSIGHTS REGARDING GROUP DECISION OPTIMIZATION

Shelley D. Dionne and Peter J. Dionne

ABSTRACT

Previous literature has compared the effectiveness of different styles of leadership, yet most of this research has not compared different levels of analyses regarding leader styles or behaviors. This shortcoming often limits our understanding of how leadership acts on a phenomenon of interest to a single level of analysis. This article develops a computational model and describes a levels-based comparison of four types of leadership that represent three different levels: individual, dyad, and group. When examined across a dynamic group decision-making optimization scenario, group-based leadership is found to produce decisions that are closer to optimal than dyadic-based and individual-based leadership. An alternative computational model varying individual cognitive and experience-based components among group members also indicates that group-based leadership produces more optimal decisions. First published in Leadership Quarterly (Dionne, S. D., & Dionne, P. J. (2008). Levels- based leadership and hierarchical group decision optimization: A simulation. Leadership Quarterly, 19, 212–234), this version offers an updated

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 227–270 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008011 227 228 SHELLEY D. DIONNE AND PETER J. DIONNE

introduction discussing simulation as a theoretical development tool and supplies additional evidence regarding the growth of simulation methods in leadership research.

INTRODUCTION

More than 20 years ago, levels of analysis was introduced into the organizational and leadership literature, highlighting both the conceptual implications (Dansereau, Alutto, & Yammarino, 1984; Rousseau, 1985)and the methodological implications (Dansereau et al., 1984) of levels-based organizational and leadership research. Since that time, several calls have been issued for greater consideration and incorporation of levels of analysis into theory development and testing within organizational research (Dansereau, Yammarino, & Kohles, 1999; House, Rousseau, & Thomas- Hunt, 1995; Klein, Dansereau, & Hall, 1994), and especially within leadership research (Dansereau & Yammarino, 1998; Yammarino & Dansereau, 2002; Schriesheim, Castro, & Yammarino, 2000; Yammarino, Dionne, Chun, & Dansereau, 2005). Despite these calls, a recent ‘‘state-of- the-science’’ review within 17 areas of leadership research revealed that less than 10% of published leadership articles adequately address levels of analysis in theory formulation, hypothesis testing, measurement, and inference (Yammarino et al., 2005). The practical message of these findings can be translated as follows: Levels-based leadership theory development and testing remains a critical, yet ripe, area for investigation. Because only a handful of leadership studies (i.e., less than 10%) have adequately addressed levels-based theory development and testing, cautious and conservative interpretation of the extant leadership literature must prevail (Yammarino et al., 2005). This point begs an important question: If leadership scholars cannot definitively pinpoint appropriate and relevant entities and levels-based outcomes for the large majority of leadership theory and literature, how can practicing leaders understand the appropriate use and application of particular leader styles and behaviors? What does levels-based theory look like in action, and should levels of analysis, in any way, guide a leader’s selection of a leadership style? Previous literature has compared the effectiveness of different styles of leadership, especially transactional as compared to transformational and/or charismatic leadership. Nevertheless, most of this research has not considered a comparison of different levels of analysis regarding leader styles or behaviors (Yammarino et al., 2005; Yammarino, Dionne, & Chun, 2002), Leadership and Group Decision 229 although a few notable exceptions exist (cf., Schriesheim, Castro, Zhou, & Yammarino, 2001; Yammarino & Naughton, 1992; Somech, 2003). Thus, our understanding of how leadership style and behavior act on a phenomenon of interest is often limited to a single level of analysis, which usually has not been adequately specified (Yammarino et al., 2002, 2005). Leadership research can begin to remedy this lack of understanding regarding levels of analysis by examining particular organizational issues across a variety of levels. Increased reliance on groups and teams in organizations (Salas, Stagl, & Burke, 2004; Dionne, Yammarino, Atwater, & Spangler, 2004) offers a rich opportunity to view leadership across multiple levels – specifically regarding the pertinent issue of decision making in hierarchical groups. Hierarchical (i.e., leader led) groups dominate in organizations (Hollenbeck et al., 1995) and offer an opportunity to examine leader behaviors and styles at the individual, dyadic, and group levels. Can group decision making be optimized at only a single level of analysis (e.g., group), or are other leadership styles at different levels (e.g., dyad) appropriate to facilitate decision making in the group context? For example, is participative leadership (Koopman & Wierdsma, 1998) the best leadership style to optimize a group decision, or could an individualized dyadic leadership style (Dansereau et al., 1995) optimize a group decision as well? Hurdles encountered in this type of research include not just the multi- level conceptualization of the leadership process, but also the notion that group interactions are more complex than the most of our linear, antecedent–consequence type theory and methods would suggest (Kerr & Tindale, 2004). Given this level of complexity and difficulty, Kerr and Tindale (2004) point to recent advances in simulation technology and empirical work using dynamical models (e.g., Stasser, 1988, 1990, 2000)asa suggestion for further examination of group decision making. Decision- making simulation research joins other simulation-based organizational research in areas such as employee withdrawal (Hanisch, Hulin, & Seitz, 1996; Hanisch, 2000), strategic management (Lant & Mezias, 1992; Adner, 2002; Zott, 2003), pay-for-performance system effects (Schwab & Olsen, 1990; Schwab & Olsen, 2000), and faking implications on personality tests (Zickar & Robie, 1999; Zickar, 2000). Although some research is beginning to address the challenges of modeling dynamic processes (Black, Oliver, Howell, & King, 2006), the multi-level nature of a leadership/group decision process within a dynamic setting has yet to be examined. From a theoretical perspective, controversy exists regarding a simulation- based approach providing a means for theoretical development. Prior arguments both advocating simulation methods (cf., Zott, 2003; 230 SHELLEY D. DIONNE AND PETER J. DIONNE

Repenning, 2002; Carroll & Harrison, 1998) and cautioning against simulation methods (cf., Fine & Elsbach, 2000; Chattoe, 1998) exist in the organizational literature. Davis, Eisenhardt, and Bingham (2007), however, recently suggested a ‘‘road map’’ for how and when simulation models can be used to promote theoretical advancement. Central to their assertion was a use of simulation methods for simple theories, or theories where ‘‘propositions are in all likelihood correct, but are currently limited by weak conceptualizations of constructs’’ (p. 482) and/or have ‘‘interactions that are only vaguely, if at all, understood’’ (p. 482). In the absence of widespread levels-based theoretical development of leadership constructs and processes (Yammarino et al., 2005), many leadership theories may exist in a somewhat ‘‘simple’’ state, where interactions between leaders and followers, especially across various levels of analysis, may be only vaguely understood, and as such, could benefit from stronger conceptualizations of constructs. Thus, based on the road map outlined for use of simulation methods in theoretical development (Davis et al., 2007), simulation methods would be a useful means for exploring and developing leadership theory, as conceptualizations of appropriate levels of analysis issues surrounding leaders, followers, and their existing environments and situations remain in a somewhat simple state. Moreover, Hazy (2007) offers a review of 14 recent computer simulations in leadership research, where several different levels of analysis were examined, including individuals, dyads, groups, and organizations. Although these simulations did not specifically examine levels of analysis issues per se, simulations detailing leadership impacts on individual followers, team performance issues, organization networks, and/or the routinization of charisma inherently suggest a multi-level perspective. Based on valuable insights gained from these several leadership simulations, Hazy (2007) suggests a variety of potential benefits from additional examination of intermediate structures and aggregates (e.g., groups, teams, departments) through simulation, some of which include the notion that power structures may persist through time and survive the transience of individual agents, and/or leadership may serve to decrease the potential of interaction catastrophe. Minimally, aggregate-based investigations suggested by Hazy (2007) likely involve nonlinear properties (i.e., time passage and/or interaction catastrophe), which may be further complicated when a theory or target in question has a levels-based component as well (e.g., dyadic relationships, leader–team interactions). Since dynamic organizational issues (e.g., nonlinear) are not widely understood in organizational research Leadership and Group Decision 231

(Hulin & Ilgen, 2000), and levels-based leadership issues are not widely researched (Yammarino et al., 2005), dynamic, levels-based leadership research may be a classic specimen of ‘‘simple theory’’ (Davis et al., 2007) suitable for simulation research. Given this omission, we propose development of a computational model to investigate how various levels-based leadership theories might influence the nature and outcome of dynamic group decisions. The benefit of the computational model over traditional experimentation or field study relates directly to the computational model’s ability to study complex processes or interrelationships that occur overtime, as would be the case with leaders, their groups, and the decision-making process; after all, teams, and arguably groups, by definition interact dynamically, interdependently, and adaptively (Salas, Guthrie, Wilson-Donnelly, Priest, & Burke, 2005). From a more practical perspective, a traditional, experimental-based examination of multi-level leadership and group decision making would require enormous cooperation from leaders and their groups, in addition to the participation of a large number of subjects who span various levels and conditions. The resulting overall sample size and commitment to participation from organizations would be logistically difficult, not to mention fairly unrealistic, to achieve. As such, computational modeling provides a viable alternative, both methodologically and practically, to a multi-level investigation of this kind. Further, because this investigation of levels-based leadership theories and group decision making is a first, computational modeling can act as a form of ‘‘pilot testing’’ for theoretical specifications and relationships. Attempts to operationalize complex leadership and decision- making processes can point researchers to weaknesses in both theory and experimental potential (Schwab & Olsen, 2000). Through the preliminary use of a computational model, this study can be accomplished without misusing precious time and efforts of actual leaders and group members. This article provides a levels-based comparison study between four types of leadership that represent three different levels: individual, dyad, and group. These leadership types are examined across a dynamic group decision-making optimization scenario, addressing the overall research question: Can hierarchical (i.e., leader led) groups optimize group decisions? This article briefly reviews group decision literature, introduces four leadership models used in the computational model, presents assumptions predicating the simulation, and discusses results and implications for future research and managerial practice. 232 SHELLEY D. DIONNE AND PETER J. DIONNE

GROUP DECISION MAKING

Within the last several years, the use of groups and teams in organizations has increased dramatically; as such, group/team research has increased as well (Kerr & Tindale, 2004). As research on group/team decision making has progressed, much of the focus has remained at the individual level, such as stereotype formation, perceived group homogeneity, and cognitions about groups (Kerr & Tindale, 2004), or tended toward consensus decision models viewed at the group level of analysis (Hollenbeck et al., 1995; Davis, 1996). However, the interrelated behaviors, cognitions, attitudes, and actions of group members in decision-making processes inherently produce a complex, multi-level phenomenon, a point which very few researchers have addressed. Hollenbeck et al. (1995) were among the first researchers to explore a multi-level nature of decision making by examining the effects of hierarchical status within a group – in other words, leader-led groups. These researchers identified three primary characteristics of a hierarchical decision-making group: unequal status among members, distributed expertise, and a task where the outcome is a decision. Additionally, they identified three-core constructs central to the decision-making accuracy: team informity, staff validity, and hierarchical sensitivity. Team informity is the degree to which all members have information on the decision objects or are apprised of the relevant cue values associated with the decision dimensions. Staff validity is the degree to which a team’s lower-level members possess judgments predictive of the true state of the decision object. Hierarchical sensitivity is the degree to which a leader effectively weights team members’ judgments when making the final team decision. Non-core constructs – including dyadic sensitivity, individual validity, and environmental factors – represented a variety of other levels of analysis and were mediated by the group-level core constructs. Important theoretical implications from Hollenbeck et al.’s (1995) research include the following points: Team decision-making theory can be considered to be an adaptation of individual decision models (providing a strong multi-level basis for consideration); unlike consensus models, inputs into decisions are a critical starting point (as opposed to outcome prominence); and decision alternatives can vary along a univariate continuum. These authors captured both the multi-level implications of team decisions, especially those with leaders, and the complexity of the decision-making process through continuum-based decision alternatives. Hollenbeck et al.’s (1995) theoretical implications are particularly relevant Leadership and Group Decision 233 in simulation research, in that each group member can begin with an individual, personal utility curve, which can then be adapted overtime as a group decision is formed. As such, a simulation can capture the variation of decision alternatives along a continuum. Hollenbeck et al. (1995) consider yet another factor that could be relevant to a decision simulation – a leader’s judgment of the predictive validity of staff members’ judgments. The relevance and usefulness of weighting subordinates’ potential performance within a decision-making simulation highlight the critical importance of leadership to the overall process. The decision literature includes some precedence for considering a weighting process, which, beyond introducing the importance of leadership, emphasizes the relevance of multiple levels of analysis. Hollenbeck et al. (1995) asserted an inherent belief that a leader has dyadic relations with each team member when they chose to assign a weighting policy. Elsewhere in the decision literature, other theories address a weighting function as well. Specifically, the Social Judgment Scheme (SJS) model for groups (Davis, 1996) extends the discrete alternative consensus processes to include a weighted linear combination of member preferences. Weights are an exponential function of the distances between one member’s preferences and all other members’ preferences (Davis, 1996). This weighting scheme results in similar members having larger weights as compared to dissimilar members, who receive smaller weights. From a levels perspective, this is a difficult concept to pigeonhole into a specific level of analysis. There is certainly an individual-level component, because the weights are functions of, in part, an individual’s preference distance from that of the group’s preferences. Nevertheless, the resulting combination of weights also may reflect a group level, or subgroup level if there are in-groups and out-groups within the team. More research is needed in this area to investigate the levels implications of this theory. Models such as this one reflect the influence of ‘‘social sharedness’’ at the preference level, positing that elements shared among group members inordinately influence the group response (Kerr & Tindale, 2004; Tindale & Kameda, 2000). Thus, the alternative with the most ‘‘sharedness’’ among members is the model most often selected by the group (Kerr & Tindale, 2004). Although the level of analysis for which it is appropriate is not clear, the notion of weighting has relevance for simulations – namely, a weighting function can be modeled and applied to various levels, especially given that Hollenbeck et al. (1995) believe weighting may be relevant at the dyadic level, and other researchers (cf., Tindale & Kameda, 2000) note group-level relevance. Either way, this 234 SHELLEY D. DIONNE AND PETER J. DIONNE theory positions weighting as a fundamental aspect of decision making, especially in hierarchical teams, which therefore warrants inclusion in a group decision simulation. Similarly, the ‘‘judge–advisor systems’’ research stream examines how ‘‘advisors’’ may influence the final decision of a ‘‘judge’’ (Budescu & Rantilla, 2000; Sniezek, 1992; Sniezek & Buckley, 1995). This setting would be similar to a leadership scenario in which the leader has final decision- making authority, but consults (or, in some cases, does not consult) members of the workgroup for advice and information. In general, research has shown that advisors can influence a judge; the judge’s preference, however, tends to lean toward his or her own position (Harvey, Harries, & Fischer, 2000). As such, the leader represents a critical decision maker in a hierarchical group setting, although Kerr and Tindale (2004) note this preference toward one’s own opinion may often be a suboptimal strategy. From a levels perspective, this could be an individual-level phenomenon, or it could reflect a dyadic level, depending on the type of interactions between the leader and group members. If the judge completely relies on the advisors, then the possibility also exists of a group-level phenomenon in which the judge suspends individual authority and favors the group decision. In any of these potential levels-based situations, clearly the role of expertise (i.e., ‘‘advisor’’) within a group is important. In a group-based decision scenario, enhanced social power can accrue from the perceptions of expertise associated with a cognitively central member, or a member whose information is similar to that held by others in the group (Kameda, Ohtsubo, & Takezawa, 1997). Thus, high expertise combined with a similar ‘‘view’’ of the decision among group members will likely result in the ‘‘expert’’ member assuming a fairly powerful role in the decision-making process. Moreover, given that groups rarely discuss strategies, but rather exchange information and, more often than not, simply choose a single member’s judgment (Tindale, Filkins, Thomas, & Smith, 1993; Kerr & Tindale, 2004), the central expert could be in the position to solely determine the direction for the group. Even in a leader-rendered decision scenario, the leader may rely on an expert to advise him or her, especially if that expert has a similar ‘‘view’’ of the decision (Harvey et al., 2000), or if that expert has been correct in the past (Budescu, Rantilla, Yu, & Krelitz, 2003). Thus, although a leader may weight his or her own preference more strongly, support from an expert is likely to affirm the leader’s decision preference, once again placing expertise in a central role in group decision making. Leadership and Group Decision 235

Thus, group decision literature has presented several critical variables to consider in a group decision simulation – the element of personal utility curves, the existence of a decision continuum, and weighting functions that ultimately lead to perceptions of expertise. As noted previously, from a leadership perspective, several levels of analysis may be relevant when considering group decisions. However, many of these multi-level considerations have not been explicitly addressed in either the group decision-making literature or the multi-level leadership literature, and the implications of continuing to ignore multi-level theory and testing within an inherently multi-level phenomenon are considerable (Yammarino et al., 2005). To provide a more detailed levels-based view of leadership on which to overlay group decision making, we brieﬂy describe four leadership models at three different levels of analysis. We then link these leadership models to group decisions in a hierarchical setting.

LEADERSHIP AND LEVELS OF ANALYSIS

To test the practicality of decision making while employing diverse levels- based leadership theories, four leadership theories or concepts could be considered for simulation modeling: an authoritarian leadership style as depicted in the normative decision model (Vroom & Jago, 1988; Vroom & Yetton, 1973) to represent an individual level of analysis; individualized leadership (Dansereau et al., 1995) to represent a dyadic level of analysis; leader–member exchange (LMX) (Graen & Uhl-Bien, 1995) leadership to represent a dyadic-within-group level of analysis; and participative leadership (Koopman & Wierdsma, 1998) to represent a group level of analysis. These leadership models have been selected for inclusion due to their speciﬁc linkages and identiﬁcation with a particular level of analysis.

Individual Level of Analysis

The autocratic style within the normative model (Vroom & Yetton, 1973) varies in the degree of subordinate participation. In AI, the leader asks for no input and makes the decision without any subordinate involvement. This would represent an individual level of analysis of leadership, in that there is no interaction or inﬂuence by a subordinate in the decision-making process. In this case, the leader most likely assumes that time is critical, that he or she has the most expertise, or that the group will support his or her decision. 236 SHELLEY D. DIONNE AND PETER J. DIONNE

Similar to AI, AII is an autocratic style, albeit one in which subordinates may respond to specific requests for information from the leader (Jago & Vroom, 1980). Ultimately, however, the leader makes the decision. Because of this autocratic decision making, the significance of the influence of the subordinates is downplayed. As such, the AII style also reflects an individual level of analysis focused on the leader. From a hierarchical group decision perspective, this leadership style focuses more on the concept of ‘‘hierarchical’’ and less on the concept of ‘‘group’’ than any other leadership style. The decision is solely made by a leader, who intends that the group then follow his or her decision, although admittedly the willingness to follow the leader’s decision is not high when the group has no influence in the decision-making process (Jago & Vroom, 1980). In any event, this leadership style is the ultimate representation of a hierarchical relationship within a group decision task. Moreover, as noted by Kerr and Tindale (2004), this preference toward one’s own opinion may often be a suboptimal strategy.

Dyadic Level of Analysis

Individualized leadership (Dansereau et al., 1995) is a multi-level approach that views leaders and subordinates as forming relationships with one another independent of the relationships that either party forms with any other person. The basis for the interdependency is the exchange between a leader’s support for a subordinate’s self-worth and the subordinate’s performing in such a way as to satisfy the leader. Dansereau et al. (1995) assert that this interdependency represents a linkage between the follower as a person and the part of a leader that views the subordinate as a person. Moreover, Dansereau et al. (1995) suggest that interdependency develops over a time, where more established dyads display higher levels of interdependency. Certain factors – for example, dissimilarity in personality and cognitive structures, and limited interaction with a superior (Dansereau et al., 1995) – may have some bearing on the perception of leader support and satisfying subordinate performance on behalf of both leader and follower. From a levels perspective, this interdependency represents a dyadic perspective in that it depends on the leader’s view of the follower and the follower’s view of the leader. From a hierarchical group decision perspective, this leadership style may be linked to the notion of ‘‘judge–advisor systems’’ discussed previously. The leader can retain ﬁnal decision-making authority, but the judge’s Leadership and Group Decision 237 preference can be swayed by the expertise of particular members, who act as advisors. The more ‘‘rich’’ the dyadic relations, the more likely that an advisor can inﬂuence a judge’s decision. Thus, the weighting functions of SJS decision models have particular relevance for dyadic leadership scenarios.

Dyad-Within-Group Level of Analysis

LMX theory depicts the relationship between a leader and a subordinate as evolving through an exchange process overtime, where subordinates may have high-quality exchanges or low-quality exchanges with the leader (Graen & Uhl-Bien, 1995). Subordinates with high-quality exchanges provide increased effort and personal loyalty to their leader; in turn, leaders in these exchanges allow subordinates more control and influence. In contrast, subordinates in low-quality exchanges do not exceed normal work expectations and rely on more formal organizational exchange patterns to guide their interactions, so leaders are less likely to initiate or provide additional benefits or resources in exchange (Schriesheim et al., 2001; Graen & Uhl-Bien, 1995). From a levels perspective, this interdependency represents a unique dyadic perspective, in that the quality of leader–follower relations depends on the leader’s view of the subgroup to which the follower belongs (in-group vs. out-group). This creation of dyadic relations within subgroupings of followers is reflected in the dyad-within-groups classification. From a hierarchical group decision perspective, this leadership style also may be linked to the notion of ‘‘judge–advisor systems.’’ The leader can retain final decision-making authority, but the judge’s preference can be swayed by the expertise of particular members, who act as advisors. In all likelihood, these advisors represent the in-group; as such, their importance (weighting) may be more likely to be assigned at a group level, in that subgroups (i.e., in- and out-groups) may receive weightings via membership status. This would differ from the individualized leadership condition, in which weighting proceeds on a one-to-one basis – that is, by judging each member based on his or her unique relationship with the leader. In the LMX approach, a more ‘‘blanketed’’ approach may be employed, where the leader weights all members of the in-group similarly and, likewise, weights all members of the out-group similarly. Again, the weighting functions discussed in SJS decision models have particular relevance for this dyad- within-group leadership scenario. 238 SHELLEY D. DIONNE AND PETER J. DIONNE

Group Level of Analysis

Participative leadership is defined as joint decision making or shared influence in decision making by a leader and his or her respective followers (Koopman & Wierdsma, 1998), with most research viewing participative leadership as generally homogeneous within a workgroup (Somech, 2003). This view has several implications closely linked with the average leadership style (ALS) approach, a between-groups approach that assumes uniform behavior of the leader toward all group members (Vecchio, 1982), and homogeneity of perceptions on behalf of all group members as to represent a single entity (Graen & Cashman, 1975). Participative decision making has been linked to a number of benefits including increased quality of decisions (Scully, Kirkpatrick, & Locke, 1995). From a hierarchical group decision perspective, participative leadership may represent an extreme point in ‘‘judge–advisor systems.’’ In this leadership model, the judge recognizes that advisors possess excellent expertise and consequently equalizes the system to provide equal input authority to all advisors. The leader retains final decision-making authority, but opts to take the group’s decision as his or her own. Therefore, the emphasis on the weighting function of members is downplayed, in that the leader recognizes each member as being equally important.

METHOD

In the preceding section, key elements from group decision theory were shown to be essential for development of a hierarchical group-based decision simulation, and four leadership models were presented that linked speciﬁc aspects of a leader’s style to elements of a group decision-making process. This section introduces speciﬁc components of the simulation, and discusses the production run process involved in the simulation.

Components of Group Decision Making

The environment of a decision problem takes into account the structure of considered options, and it generally produces two types of structure: discrete or continuous. The discrete case (Kemeny & Snell, 1962) proposes a model of ordinal-ranked options, with more recent modiﬁcations employed to Leadership and Group Decision 239 examine degree of preferences (Cook & Kress, 1985). Consensus generally represents the minimization of distance between individual and group preferences (Beck & Lin, 1983; Cook & Kress, 1985). The continuous structure of a set of options is most widely operationalized to assume that every decision maker has an implicit utility function that aggregates multiple objectives (Zionts & Wallenius, 1976; Abdelaziz, Martel, & Mselmi, 2004). The group’s utility function is determined by the nature of the group. For example, a dictator produces a dominant utility function, whereas a cooperative group produces an aggregated decision utility, usually in additive form (Keeney & Raiffa, 1976; Torrance, Boyle, & Horwood, 1982; Abdelaziz et al., 2004). Based on the previous theoretical discussion linking levels-based leadership to continuous decision making, and because continuous structures lend themselves to comparison of levels-based concepts (i.e., a dictator at an individual level and a cooperative at a group level), a continuous structure was employed in the model, producing personal utility curves for all group members, including leaders. Personal decision utility curves were randomly generated in a piecewise linear fashion. Fig. 1 shows example of personal utilities curves for a single group. Consistent with the continuous structure of decision making, group decision utility curves were established as a weighted aggregated decision utility, compiled for each group in an additive form. The weighting factor is an ideal measure of an individual’s importance within the group. Individuals who are more important have a greater inﬂuence on the group utility curve. Fig. 2 shows the team utility curve aggregated from the eight personal utility curve samples presented in Fig. 1.

Dynamic Decision Component Based on interactions between group members, members could shift their personal decision points toward that of another member, or toward the decision point of maximum team utility, depending on a variety of individual and team factors. This continuous adjustment of personal utility more accurately reﬂects the interdependency of group members involved in a decision-making task. The interdependency and complexity of group interactions, and the resultant decisions, require the addition of several performance components to a group decision simulation to reﬂect the basis for what may be driving the willingness of leaders and group members to shift their decision points. 240 SHELLEY D. DIONNE AND PETER J. DIONNE

1 Functions 0 Utility

0 Individual

Sample 0

0 0 0 0.2 0.4 0.6 0.8 1 d (Decision Point) Fig. 1. Sample Personal Utility Curves for Group Members.

Components of Group Performance

Salas et al. (2005) note that, owing to its unique interactions and dynamics, modeling team performance is more complicated than modeling individual performance. Based on the previous theoretical discussion identifying individual and group components relevant for consideration in decision models, and based on suggestions and prior research presented by simulation experts, several performance components were included in this Leadership and Group Decision 241

6.5

Optimal Group Utility 6

5.5

5 Function

Utility

4.5 Group 4 Sample 3.5

3 0 0.2 0.4 0.6 0.8 1 d (Decision Point) Fig. 2. Sample Group Utility Curve. simulation. To guide research within the team simulation domain, Salas et al. (2005) produced a heuristic to aid researchers in determining which factors to include in team performance models. Given the similarity between the decision dynamics of groups and teams, and because decision optimization is a form of group performance, we have applied their heuristic to our concept of group performance as well, especially as it relates to a dynamic decision-making process. First and foremost, Salas et al. (2005) noted that the following components must be included in any team performance model, reﬂecting the importance and history of these variables in affecting various dimensions of performance:

Individual characteristics of team members, such as cognitive ability or personality; Team characteristics such as team type, structure, and size; Task characteristics such as task type and interdependency; Work structure, such as assignments and communication structure; The environment, especially regarding task load and time pressure; 242 SHELLEY D. DIONNE AND PETER J. DIONNE

Noise, which these researchers viewed as consisting of any unwanted sounds by a listener (their particular interest is in simulation games where noise may impede hearing, communication, and attention).

Because time and money constraints as well as less than ideal research conditions are present within organizations, Salas et al. (2005) recommended additional components that could be modeled to more accurately represent the realism of team performance, such as team diversity, and individual characteristics, such as expertise, motivation, and attitudes. Moreover, including team competencies (e.g., leadership, shared mental models, and adaptability) and team cohesiveness, climate, and team norms (difficult-to-model key areas where data are lacking) also could better represent the realism and complexity of team performance conditions. Following Salas et al.’s (2005) suggestions for modeling components of team performance, the ‘‘must be modeled’’ components were included either as parameters or as variables, depending on their relationship with decision making. Some of the ‘‘should be modeled’’ components were included to enhance the realism of the decision-making environment. Likewise, the ‘‘would like to model’’ component of leadership was included to advance our understanding of a leader’s influence on group decision making. Table 1 presents the model characteristics, organized using Salas et al.’s (2005) heuristic for team simulation models as a framework. The following sections detail the specific components included in the leadership and group decision- making model.

‘‘MUST INCLUDE’’ MODEL COMPONENTS

Individual Characteristics

Cognitive Ability Given the complexity pervasive in today’s team-based, fast-paced global organizations, the importance of hiring and retaining intelligent personnel may be increasing. Gottfredson’s editorial statement in the Wall Street Journal dated December 13, 1994 (reprinted: Gottfredson, 1997a) discussed, among other factors, that although differences in intelligence are not the only factor affecting performance in highly complex jobs, intelligence is often the most important factor. This statement was signed by 51 other premier researchers in the scientiﬁc study of intelligence, who were Leadership and Group Decision 243

Table 1. Model Components Based on a Team Simulation Heuristic.

Team Simulation Leadership and Group Component Mean Values Component Component Decision Model Ranges Around Recommendations Mean Values

Must include Individual characteristics Cognitive ability IQ 118 77 (111–125) Personality Openness 27 76.5 (20.5–33.5) Team characteristics Team type Professional employee Assumed group Team structure Individual, dyad, group Varies depending on level Team size 8 group members 8 70 (no variability) Task characteristics Task type Decision task Assumed Interdependency Individual, dyad, group Inﬂuence varies inﬂuence depending on level Work structure Communication Individual, dyad, group Interactions vary structure interactions depending on level Environment Time pressure Time-step iterations 10,000 70 (no variability) Noise Random noise 1% applied to probability 70 (no variability) of accepting new decision Should Include Individual characteristics Expertise Expertise, tenure, and Expertise: 0.9 Expertise: 70.10 actual importance of all (0.8–1.0) group members Tenure: 8 Tenure: 72 (6–10 years) Actual importance: weighted average of IQ, openness, expertise, and tenure Attitudes Perceived importance of Perceived importance: 0.5 70.5 (0–1.0) all group members Like to Include Leadership Autocratic, Interaction patterns and individualized, LMX, frequencies, and participative probability of accepting new decision varies depending on level

Source: Salas et al. (2005). 244 SHELLEY D. DIONNE AND PETER J. DIONNE representative of a wide spectrum of disciplines and perspectives (Gottfredson, 1997a). Gottfredson (1997b) also notes that general intelligence (g) has pervasive utility in work settings because of its ability to deal with complex information processing or, more generally, its ability to mentally manipulate information (Jensen, 1981). Recall that intelligence is not how much you know, but rather how well you execute complex mental behaviors such as effectively acquiring, updating, organizing, selecting, and applying information (Gottfredson, 1997b). As such, the more complex a work task, the more likely that higher general intelligence will facilitate its performance. Because decision making in a team setting involves a certain amount of complex information processing, especially in the light of the notion that teams are often assembled to handle problems too complex for individuals (Salas, Dickinson, Converse, & Tannenbaum, 1992), intelligence is included as a key individual-level variable within the model. Intelligence satisﬁes Salas et al.’s (2005) requirement for inclusion of a cognitive ability indicator in team simulation research. Intelligence parameters included in the model were obtained from Gottfredson (1997b), who reprinted the Wonderlic (1992) scores examining ‘‘position applied for’’ and ‘‘training potential.’’ To include general variability within the model, positions representing the 70th to the 91st percentile of the general population were selected for inclusion. These positions (nurses, sales account executives at the 70th percentile, to research analysts, attorneys at the 91st percentile) represent the type of position requiring advanced education; as such, they may indicate environments that are more complex and dynamic than environments in positions below these percentiles. Positions below these percentiles include clerical, skilled, and unskilled laborers and various manufacturing positions, which may not require the same educational levels that positions in the top 30 percentiles represent. As such, these lower-percentile positions may not experience the same complex, dynamic decision-making environment, and their participation in decision making may be more variable, ranging from no decision-making authority, to some decision-making authority, to equal decision-making authority. Therefore, as an initial investigation, individual intelligence levels as represented in the upper-percentile positions were modeled for group members, in an attempt to reﬂect the highest probability of a complex, dynamic work environment requiring group decisions. Leadership and Group Decision 245

Personality One of the Big Five personality factors, openness may be particularly relevant to decision making within teams. Individuals who rate high on openness are receptive to new ideas, approaches, and experiences (McCrea & Costa, 1990). Other researchers have linked high levels of openness to teams that look for new and creative ways of completing tasks (Saavedra, 1990), and that pursue ongoing learning and improvement (Hofmann & Jones, 2005). As such, openness was modeled due to its relationship with being willing to accept other team members’ or the leader’s utility for a decision. A team member who rates low on openness may not be receptive to new ideas or approaches, including another team member’s utility function for the same decision. Given that this receptivity to new approaches and ideas could determine a person’s willingness to move off his or her own decision utility, openness was included in the simulation. Using the inventory profile included with the NEO Five-Factor Inventory (Costa & McCrea, 1991), an openness range was determined for inclusion into the simulation. According to the NEO Five-Factor Inventory (Costa & McCrea, 1991), the 50th percentile score on openness is 27 (‘‘average’’ category for openness in both men and women). The range for openness selected for the simulation was 715 percentiles (35th–65th percentiles); this score range spans from ‘‘low’’ on openness (18–23) to ‘‘high’’ on openness (31–36), including ‘‘average’’ on openness (24–30). For the purposes of this project, a conservative approach to personality was taken, in that extreme scores (i.e., very low and very high) were not considered, providing a team profile of moderately reasonable members. Undoubtedly, teams sometimes include extreme individuals, and these individuals could certainly influence the decision-making process.

Team Characteristics

Team Type To reﬂect the complexity facing organizations today, and especially professional employees within organizations, the team type was assumed to reﬂect a hierarchically based decision group with distributed professional expertise. Hollenbeck et al. (1995) note that these kinds of teams exist in all types of organizations and may include managerial staffs, health care-based teams, research teams, and congressional committees. 246 SHELLEY D. DIONNE AND PETER J. DIONNE

Team Structure Depending on the level of analysis, team structure varied. Members of the individual-level group did not interact with one another in any way. Members of the dyadic-based group interacted on a one-to-one basis with the leader. Members of the dyad within group interacted on a one-to-one basis, but members of the ‘‘in-group’’ were more likely to engage in interactions with the leader than members of the ‘‘out-group.’’ Lastly, in the group-level scenario, members were equally likely to interact with one another, including the leader.

Team Size Based on previous literature, which revealed that teams containing more than 10 members often divide into subteams to accomplish tasks (Likert, 1977), team size was established as seven group members and one leader – in other words, an eight-person decision group. Eight-person groups have been used previously in dynamic system simulations regarding leaders and teams (Black et al., 2006).

Task Characteristics

Task Type The type of task established for the simulated groups was a decision task. That is, the groups did not manufacture or produce a product. As established previously, decision teams exist in all types of organizations (Hollenbeck et al., 1995).

Interdependency Interdependency was modeled depending on the level of analysis. At the individual level, no interdependency was modeled, as a leader solely controls the group decision process. At the dyad level, interdependency was modeled on a one-to-one basis. At the dyad-within-group level, interdependency was modeled on a one-to-one basis, but with a greater interdependency among members of the respective ‘‘in-groups’’ and ‘‘out- groups.’’ Finally, interdependency at the group level was modeled as a collective concept, where all members were equally interdependent on one another. Leadership and Group Decision 247

Work Structure

Communication Depending on the level of analysis, group communication varied. Members of the individual-level group did not communicate or interact with one another in any way, as the leader solely controls the decision. Members of the dyadic-based group communicated and interacted on a one-to-one basis with the leader. Members of the dyadic-within-group decision group communicated and interacted on a one-to-one basis as well, though members of the ‘‘in-group’’ were more likely to engage in communication with the leader than members of the ‘‘out-group.’’ Lastly, in the group-level team, members were equally likely to communicate and interact with one another, including the leader.

Environment

Time Pressure To simulate time pressure, model iterations were used to represent time passing, where higher iterations represented longer periods of time. These iterations provide some indication of the length of time needed to reach decision optimization. A discussion of time steps appears in a later section of this article describing production runs.

Noise

Random variation was included in the model to represent noise, or uncertainty in prediction. Because human behavior cannot be perfectly predicted, this randomness is included to model that inherent uncertainty.

‘‘SHOULD INCLUDE’’ MODEL COMPONENTS

Individual Characteristics

Expertise Research exists linking recognition of expertise and effective decision making (Bottger, 1984; Bottger & Yetton, 1988; Libby, Trotman, & Zimmer, 1987; Littlepage & Silbiger, 1992). Group performance on 248 SHELLEY D. DIONNE AND PETER J. DIONNE decision-making tasks has been linked to both the level of expertise and the ability to recognize expertise (Littlepage & Silbiger, 1992; Libby et al., 1987), and previous research has focused on the appropriate weighting of expertise within the group (Bottger & Yetton, 1988; Einhorn, Hogarth, & Klempner, 1977). Given that previous research indicates a link between level of expertise, and recognition of expertise in group decision-making effectiveness, these factors were included in the model. All group members, including the leader, were randomly assigned an expertise level ranging from 0 (no expertise) to 1 (high expertise).

Perceived Expertise The perceived expertise level was more complex, in that its calculation reflected an overall importance factor and considered the influence of a variety of other individual components. For example, a team member with high expertise can be perceived as important; however, given the complex nature of group decisions, other factors such as tenure and intelligence may be critical as well. A very bright individual with a long tenure on the team may have some insight useful to an expert who may be new to a team. Consider a financial expert who is new to a design team and is charged with the task of calculating the cost of a new product launch. An engineer who has 10 years tenure on the team and high intelligence may be able to provide insight about cost-cutting mechanisms, barters, and other measures that the team has previously employed, based on the ‘‘group memory’’ that comes with increased tenure and intelligence of a particular team member. Similar to the notion that quality is a multidimensional construct (Garvin, 1987; Mersha & Adlakha, 1992; Richard & Allaway, 1993), a person’s importance to the group can be likened to his or her overall quality rating within that group. Using a single attribute to evaluate importance to the group seems short sighted, given the use of multidimensional quality constructs available in decision sciences. Thus, a perceived importance factor that accounts for tenure, intellect, and expertise may be more indicative of the complexity involved in deciding the quality or importance of various group members. As such, perceived importance was modeled as a function of intelligence, expertise, and tenure; each person on the team, including the leader, then ‘‘determined’’ the perceived importance factor for every other person on the team. Perceived importance values ranged from 0 (not important at all) to 1 (most important). Within the simulation, perceived importance drives the dynamic outcome of the interaction between members, relying on the importance of symmetry Leadership and Group Decision 249 versus asymmetry to increase the willingness of a member to move toward another decision point. The notion of symmetry of perceived importance may be linked to findings from implicit egotism, which reveal that individuals who are similar in some form (e.g., intelligence, expertise, tenure) are more likely to be attracted (Jones, Pelham, Carvallo, & Mirenberg, 2004)to one another, regardless of the expectation of reciprocal liking, which is more related to attitudes and values within the similarity attraction theory (Jones et al., 2004). Moreover, implicit egotism has been linked to increased cooperation (Miller, Downs, & Prentice, 1998) and increased compliance with requests (Burger, Messian, Patel, del Prado, & Anderson, 2004). As such, members with symmetric perceived importance are more likely to move toward another symmetric member’s decision point. These factors have implications for group decisions. For example, two members with symmetric perceived importance (i.e., similar to each other) may regard each other’s decision utility as plausible, or worthy of value and consideration. An asymmetric perceived importance between two members (i.e., one member is perceived as important and one member is perceived as not important) may result in the highly important member discounting the plausibility of the decision utility of the low-importance member, as that member lacks the same intellect, expertise, and tenure with the team. As a consequence, members with asymmetric perceived importance are not likely to move toward the lesser member’s decision point.

Tenure Tenure with a group may represent a form of expertise in terms of ‘‘group memory.’’ Therefore, tenure with the group was assigned randomly on a scale from 0 to 10 years, with 5 years representing the mean tenure of the group.

Actual Importance This individualized measure assesses a group member’s actual importance to the group. To obtain this score, each member’s individual utility function is multiplied by his or her actual importance factor; this factor is itself computed as a function of intelligence, openness, expertise, and tenure, where the peak individual maximum is normalized to 1. This factor scales the group utility curve in the summation process of personal utility curves, setting the optimal group utility point. 250 SHELLEY D. DIONNE AND PETER J. DIONNE

‘‘WOULD LIKE TO INCLUDE’’ MODEL COMPONENTS

Leadership

Individual Level of Analysis To model the autocratic style, a leader simply makes a decision based on his or her personal utility on behalf of the group. In this approach, the point where the leader’s personal utility is optimized is the decision rendered for the group. The group members’ utilities are not considered in any way, other than in determining the group utility curve.

Dyad Level of Analysis To accurately represent an individualized leadership scenario, the model assumed that leaders had the highest levels of interdependency with subordinates with longer tenure. Interdependency was modeled as the number of interactions between the leader and the subordinate. Given that prior research has indicated stronger ties between satisfying performance and self-worth in relationships with greater tenure (Dansereau et al., 1995), increased exposure or interdependency with a subordinate may present more opportunities for the subordinate’s performance to satisfy the leader. Moreover, this satisfying relationship may increase the likelihood of the leader accepting the decision of the subordinate, because that subordinate has provided satisfying performance in the past. This increased likelihood was modeled by developing a preliminary set of probabilities for accepting decisions, which were based on the underlying theoretical notions of the types of relationships expected within individualized leadership. Table 2 shows these probabilities. Because dyadic relationships are independent of any other relationship within the team (Dansereau et al., 1995), we did not generate a matrix comparing other team members’ perceived importance. The theoretical independence of these relationships has interesting practical implications for team decision making. The leader interacts uniquely with each team member and does not consider the value (or lack of value) of any other team member during that interaction. In part, based on the established exchange relationship, the leader provides support for the follower’s sense of self-worth; the follower reciprocates by delivering satisfying performance. Leadership and Group Decision 251

Table 2. Probability of Accepting New Decision by Levels-Based Leadership.

Acceptance of New Decision Probability of Accepting New Decision (Beneﬁt Recipient) (by Leadership Condition)

Individual Subgroup Team Participative Individualized LMX LMX leadership leadership (in-group) (out-group)

No No No 0.0001 0.0001 0.0001 0.0001 No No Yes 1.00 0.95 0.95 0.95 No Yes No 0.01 0.025 0.03 0.03 No Yes Yes 1.00 0.99 0.9 0.9 Yes No No 0.00625 0.001 0.01 0.025 Yes No Yes 1.00 0.98 0.85 0.98 Yes Yes No 0.015 0.025 0.03 0.05 Yes Yes Yes 1.00 1.00 1.00 1.00

Dyad-Within-Group Level of Analysis Scandura, Graen, and Novak (1986) found a significant positive correlation between LMX quality and strength of a subordinate decision influence. Other research has revealed similar findings related to empowerment and control over the leader in high-quality LMX relationships (Keller & Dansereau, 1995; Schriesheim, Neider, & Scandura, 1998), indicating that subordinates in high-quality LMX relationships are likely able to exert stronger control over the leader (Schriesheim et al., 2001). As such, subordinates in high-quality LMX relationships are more likely to interact with and influence their leader to move toward the subordinate’s decision utility in the model, whereas subordinates in low-quality LMX relationships are less likely to interact with and influence their leader to move toward their decision utility. The basis for modeling high-quality LMX relationships versus low- quality LMX relationships was based on research by Dienesch and Liden (1986) and Liden, Wayne, and Stilwell (1993), who suggested that the more similar the dyad perceives itself (including attitudes and personality), the more likely and enhanced working relationship will result. This affective component is believed to play an important role in LMX development (Dienesch & Liden, 1986); thus, modeling of high- and low-quality LMX relationships was based on similarity of leader and subordinates. Leaders and subordinates who were similar on individual-level factors such as personality, intelligence, experience, and expertise were modeled with high-quality LMX relationships; dissimilar leaders and followers were 252 SHELLEY D. DIONNE AND PETER J. DIONNE modeled with low-quality LMX relationships. An element of ‘‘noise’’ was added to the model – that is, 1% of the dyads were randomly assigned to high- or low-quality LMX relationships – to better account for those relationships where dissimilar individuals have a high-quality relationship and similar individuals have a low-quality relationship. Given this theoretical perspective, a preliminary set of probabilities was developed, representing the probability of accepting an alternate decision as a result of an interaction either between a leader and a group member or between group members interacting with each other. These probabilities are presented in Table 2.

Group Level of Analysis Participative leadership was modeled to reﬂect the notion that all group members, including the leader, were equally likely to interact owing to their perceived importance factors being equal, which in effect negates the hierarchy. Although modeled as interactions among two individuals, the preliminary set of probabilities to accept or reject decision points more closely reﬂects this equalized condition; that is, group members in this condition are likely to ignore the person-aspect of the interaction and focus more on the outcome or decision utility gain for the overall group. With equal perceived importance, any one member has an excellent chance to sway another team member, including the leader. As illustrated by the probability of accepting a decision in a given interaction, this condition places less of an emphasis on self-gain and shows low concern for who belongs in which subgroup. These probabilities are presented in Table 2.

Production Runs

Monte Carlo The Monte Carlo technique employed in the computer simulation resembles techniques widely used in the physical sciences to simulate the structure and dynamics of molecules (Dionne, Ozisik, & Picu, 2005; Dionne, Picu, & Ozisik, 2006). It refers to the use of random sampling of distributions. For example, in molecular modeling, a system of molecules can be simulated by randomly varying the positions of the molecules in the system. The energetic interactions between molecules are mathematically modeled, and an overall energy level of the system can be determined. If the new position of a molecule produces a system at lower energy, the move is accepted; otherwise, the move is rejected, another molecule is selected at random, and Leadership and Group Decision 253 a new position randomly chosen. This typifies the idea of importance sampling, where the molecular states with a high probability are also the states that contribute the most to the system’s properties. Importance sampling of this type is commonly referred to as the Metropolis algorithm, after the researchers who first described this method (Metropolis, Rosenbluth, Rosenbluth, Teller, & Teller, 1953). This importance sampling process is repeated many times, and thermodynamic properties of the system can be determined by averaging certain quantities over the many molecular states generated. Similar to how molecular systems are simulated, the decision-making process of a group of people can be simulated using the Metropolis Monte Carlo technique. Instead of using importance sampling to minimize the energy level of a molecular system, however, importance sampling is applied to maximize the utility of a group of people making a decision. Instead of a system of molecules, we have a group of individuals, who, just like molecules, interact with one another. Instead of molecular positions in space, we have people at different decision points on a one-dimensional decision space and a corresponding one-to-one mapping to an individual’s utility function. The repeating process of selecting two people at random from the group, randomly choosing a new decision point, and then determining whether that new decision point is accepted by the individuals, based on the group utility function, is directly analogous to the scheme employed in molecular simulations. All data inputs consisted of synthetic data, generated from parameters guided by theory and established specifically for this simulation. These parameters reflect ‘‘sample characteristics’’ that are reported when more traditional empirical data analyses are conducted. Table 3 describes the computer simulation algorithm.

Sample Characteristics In the model, 10,000 eight-member teams were used to examine each leadership condition as well as an overall control condition. Each team experienced 10,000 time steps, which represented the number of interactions between group members. Importantly, a time step does not represent a formal meeting per se, but rather could represent an e-mail discussion, casual conversation, informal meeting, phone call, or any other type of interaction in which members consider their decision positions. More- over, interactions could last from seconds to hours, depending on the type of interaction; we were more interested in the decision outcome (i.e., movement toward the optimal decision) than in the length of time members were 254 SHELLEY D. DIONNE AND PETER J. DIONNE

Table 3. Computer Simulation Algorithm.

Start loop 1 (number of teams simulated) Assemble a group of people Randomly pick a leader Assign characteristics to individuals by randomly picking from a set distribution Characteristics include IQ, openness, expertise, and tenure Assign importance factors to everyone Assign how others perceive the importance of an individual (do this for all individuals) Randomly assign a utility function to every individual Initialize everyone’s memory of all others’ decision points Determine the initial group utility function (at time ¼ 0) Determine the maximum group utility and corresponding decision point Start loop 2 (number of one-on-one interactions simulated) Choose two individuals to interact at random Choose a new decision point that is between the two individuals’ current decision points Determine whether each individual will accept the new decision point Determine whether the individual utility function increases or decrease as a result Determine whether the combined utility function increases or decreases as a result Determine whether the overall group utility function increases or decreases as a result based upon change in the individual, pair, and group utility functions Determine whether this new decision point is accepted Add the values of both individuals’ utility functions at this new decision to the memory of both individuals interacting Both individuals share their memory of other people’s utility at this decision point Determine the new group decision Determine the new group utility function Determine the difference between the current group utility and the optimal group utility Repeat Loop 2 Repeat Loop 1 meeting. Although length of time is critically important in most group decisions, the number of interactions could be representative of a time period if one assumes that 100 interactions to reach an optimal decision take a signiﬁcantly less period of time than 10,000 interactions.

Production Runs 1–4 To serve as a type of ‘‘generic’’ control, an ‘‘ideal group’’ condition was implemented in which all eight members of the group had equal actual and perceived importance, accepted only new decisions that moved the team toward the optimal decision, ignored any personal utility, and did not experience any ‘‘noise’’ or randomness in their environment. This condition illustrated that in a perfect team setting, convergence to optimal could occur; thus, it illustrated that the components included in the model could Leadership and Group Decision 255 ultimately produce an optimal decision. Although considered ‘‘ideal,’’ this ‘‘control’’ condition represented Production Run 1, participative leadership represented Production Run 2, individualized leadership represented Production Run 3, and LMX represented Production Run 4. In Production Runs 1–4, members’ individual attributes – namely, IQ, openness, expertise, and tenure – were set to the overall mean level (see Table 1). Differences between leadership conditions were represented by the interaction pattern and frequency, as dictated by the appropriate levels- based theoretical lens. For example, in an LMX situation where an interaction occurs between a leader and an out-group member, the probability that the leader will move off his or her decision utility toward that of an out- group member’s utility is likely lower than the probability that the same leader will move toward a decision utility possessed by a member of the in-group. When interacting with a member from the in-group, the leader is more likely to consider the in-group member’s opinion and move along a decision curve toward that of the in-group member’s curve.

Production Runs 5–8 In this set of production runs, the results from each prior leadership condition were examined for veriﬁcation purposes. Within these production runs, members’ individual attribute mean scores were decreased by 3%, to verify that decreasing members’ IQ, openness, expertise, and tenure would reduce the chances of reaching the optimal decision. Creating less experienced groups with less effective problem-solving skills should negatively affect group decisions if the model components adequately represent underlying group decision factors. If these production runs do not produce decision curves farther away from the optimum, the applicability of the model would be seriously compromised, as it would indicate that we have not established key underlying factors. As in Production Runs 1–4, differences between leadership conditions were represented by the interaction pattern and frequency, as dictated by the appropriate levels-based theoretical lens. In other words, the actual and perceived importance of individuals, which produced potentially different interaction patterns and frequencies within leadership conditions, drove interactions within the levels-based leadership settings. Similar to the situation in Production Runs 1–4, participative leadership represented Production Run 5, individualized leadership represented Production Run 6, and LMX represented Production Run 7. The ‘‘control’’ condition was not relabeled and remained Production Run 1, as changes in means will not affect movement toward the optimal decision in this perfect condition, 256 SHELLEY D. DIONNE AND PETER J. DIONNE because movement toward the optimum is driven solely by the equalization between actual and perceived importance factors, both of which retain their value of 1. Although each production run represents a levels-based leadership condition, the autocratic leadership condition has no interactions and, therefore, is not subjected to a production run. Rather, the autocratic leader’s personal utility curve optimal is selected as the group decision and is represented by Production Run 8. The utility line for this condition reﬂects the difference between the average personal utility point for a sample of 10,000 leaders, who engage in no interactions with followers, and their optimal group decision point, calculated using the group members’ personal utility curves. Autocratic leadership also is unaffected by mean shifts. This decision point is solely driven by random processes; as such, this randomization averages to the same decision point regardless of member attribute changes. Therefore, the autocratic leadership condition was not relabeled and remains Production Run 8.

RESULTS

The production run curves in Fig. 3 depict the average distance between the group decision utility at any given time step and the optimal group decision utility at that same time step. The ideal ‘‘control’’ condition (Production Run 1) was the only condition in which a group could reach the optimal decision (converges toward zero). Production Run 2, participative leadership, followed an approximately similar decent (i.e., slope) toward optimization for the ﬁrst 2,000 interactions; after that point, the rate of decent toward optimization became less rapid, and generally plateaued after about 5,000 iterations. Interestingly, this was also the apparent plateau point for both Production Run 3 (individualized leadership) and Production Run 4 (LMX). This trend indicates that after the halfway point, decision groups fail to reach any more attractive decision across any leadership model. Production Run 3, individualized leadership, appeared to be the next most productive leadership model (after participative leadership) in moving a group toward the optimal decision. Individualized leadership enables an initially rapid decent toward the optimum; after approximately 1,000 iterations, however, the rate of decent becomes less rapid than that seen with participative leadership as it moves toward its plateau point. Leadership and Group Decision 257

1.1

1 Utility 0.9 P8 Group

0.8

0.7 P4 Current

0.6 P3 minus 0.5 P2

Utility 0.4

0.3 Group

0.2 P1

Optimal 0.1

0 0 2000 4000 6000 8000 10000 Current Time Step (Number of Interactions) Fig. 3. Production Runs for Levels-Based Group Decision Simulation. P1 ¼ Ideal Team (Control), P2 ¼ Participative Leadership (Group Level), P3 ¼ Individualized Leadership (Dyad Level), P4 ¼ LMX (Dyad-Within-Group Level), P8 ¼ Autocratic Leadership (Individual Level).

Similar to the curve for individualized leadership, Production Run 4 (LMX) displays an initially rapid decent toward optimization, but after approximately 1,000 iterations, the rate of decent becomes less rapid as the group moves toward its plateau value. Additionally, LMX appears less effective than individualized leadership and participative leadership in moving group members toward optimization, as indicated by the curve being farther away from the optimum. Production Run 8, autocratic leadership, appears to be the least effective leadership method for reaching the optimal decision. As noted by the line farthest away from the decision optimum, autocratic leadership initially (at time 0) represents the most optimal decision of any leadership model. However, its decision points are very quickly surpassed by other leadership styles, as members in these other leadership conditions descend toward the optimum. Fig. 4 compares Production Runs 1–4 and 5–8, illustrating what would happen to the decision utility curves if the personal abilities and expertise of 258 SHELLEY D. DIONNE AND PETER J. DIONNE

1.1

0.9 Utility P8 0.8 Group P7 0.7 P6

Current 0.6 P4

0.5 P3 minus P5 0.4

Utility P2 0.3 Group

0.2

0.1 P1 Optimal 0 0 2000 4000 6000 8000 10000 Current Time Step (Number of Interactions) Fig. 4. Production Runs for Levels-Based Group Decision Simulation Comparing a 3% Decrease in the Mean for an Individual’s Intelligence, Openness, Expertise, and Tenure. P1 ¼ Ideal Team (Control); P2 ¼ Participative Leadership (Group Level); P3 ¼ Individualized Leadership (Dyad Level); P4 ¼ LMX (Dyad-Within-Group Level); P5 ¼ Participative Leadership – 3% Decrease in IQ, Openness, Expertise, and Tenure; P6 ¼ Individualized Leadership – 3% Decrease in IQ, Openness, Expertise, and Tenure; P7 ¼ LMX – 3% Decrease in IQ, Openness, Expertise, and Tenure; P8 ¼ Autocratic Leadership (Individual Level). all members were decreased by 3%. As expected, the ‘‘control’’ condition (P1) and ‘‘autocratic leadership’’ condition (P8) curves remain exactly the same, as neither relies on perceived importance to generate interaction outcomes. Each condition remains the same distance from the computed team utility curve, so these curves are not relabeled. Similarly, although members of the participative leadership group underwent a 3% decrease in their individual cognitive and experience-based components, the change produced only a slightly less effective decision curve (P5 is slightly farther away from the optimum than P2). Given that this condition relies on equalized perceived importance, it is not surprising that even less desirable members can produce better decisions by remaining open to the notion that each member has the potential to contribute signiﬁcantly Leadership and Group Decision 259 to the group decision. Notably, even with the decrease in individual components, this leadership style remains superior to the other three leadership styles examined, as indicated by a decision curve closer to the optimal decision point (P5 is closer to the optimum than P3, P4, P6, P7, and P8). As expected, decreasing the cognitive skills and expertise of members produced decision curves that were farther from the optimum in both the individualized leadership condition (P3 remains closer to the optimum than P6) and the LMX condition (P4 remains closer to the optimum than P7). When dealing with less cognitive complexity and less experienced team members, both of these leadership conditions produced less optimal decisions, although individualized leadership remains a superior style, in terms of decision optimization, to LMX leadership (P6 is closer to the optimum than P7). Lastly, in spite of the decrease in individual components’ contributions to the four leadership styles, participative leadership, individualized leadership, and LMX remained superior leadership styles (in terms of decision optimization) relative to autocratic leadership. Autocratic leadership (P8) remains the farthest from decision optimization.

DISCUSSION

The levels-based leadership simulation has several interesting implications. Most notably, it suggests that a group-based leadership style is most effective for promoting decision optimization in hierarchical groups. The notion of equalized perceived importance, a group-based phenomena, seems to be an important contributing factor to decision optimization. Although weighting the potential contribution of team members appears to occur in group decision-making contexts (cf., Kerr & Tindale, 2004), leaders may want to consider strongly promoting a sense of equality within the group. This equality may lead to equalized perceived importance factors (i.e., equal weighting of members), which has been shown in this simulation to be an effective tool for promoting optimization in hierarchical decision groups. Although participative leadership fails to converge toward the optimal hierarchical group decision, it represents a more effective leadership style than any other style examined in the simulation. Moreover, even when ‘‘less capable’’ members are placed on a hierarchical team, simulation results indicated that participative leadership is still a more effective style in promoting decision optimization than other leadership conditions – including those where group members remained 3% ‘‘more capable’’ in 260 SHELLEY D. DIONNE AND PETER J. DIONNE terms of intelligence and expertise. Therefore, when faced with less than optimal member abilities, leaders of hierarchical decision groups may want to consider an overall focus on equality or perceived importance within the group. This focus on equality appears to contribute to decisions that are generally more optimal than those made under other dyadic or individualized styles, such as individualized leadership, LMX, and autocratic leadership. If equality is not present within a hierarchical decision group, decision optimization may be more effective if members possess the best possible individual abilities and expertise. As indicated in the simulation, both individualized leadership and LMX decision groups achieved better optimization when members possessed higher levels of ability and expertise. Although there were differences between the optimization values of these leadership styles, both conditions proved more effective than conditions where members possessed less ability and expertise, with the exception of participative leadership. As such, leaders should attempt to hire ‘‘the best and the brightest.’’ In such a scenario, even if a leader wants to form in-groups and out-groups within these hierarchical decision groups, they may still outperform less able group members under other dyadic- and individual-based leadership conditions. Although both individualized leadership and LMX were less effective in promoting decision optimization, it is important to note that the LMX leadership condition was less effective than the individualized leadership condition. This ﬁnding may be related to the notion that although individualized leadership maintains a hierarchy within the group, the leader’s focus on establishing one-to-one relationships with followers may have an effect analogous to the sense of equalization evident in participative leadership, albeit on a lower level of analysis. Although equalization of members and a leader is not present in individualized leadership, the idea that everyone is valued in some unique way by the leader may establish a sense of inclusiveness. This inclusiveness, which stems from one-to-one reciprocal relationships between each follower and the leader, may be an important factor in decision optimization at lower levels of analysis (i.e., dyads). Although not as powerful as equality in promoting decision optimization, in scenarios where participative leadership is not possible or not practiced, a leader who has developed strong individualized relationships with his or her followers may set an example that followers within the group then model. In other words, every member of the group has the potential to contribute something. Although the members’ potential to contribute may not be equal within the group, each follower nevertheless plays some role in the decision process. Leadership and Group Decision 261

This inclusiveness may promote a better-quality interaction within the hierarchically based group, thereby improving decisions relative to what can be achieved in situations where there is a clear sense of exclusion, like that found in an out-group scenario within an LMX environment. In an LMX setting, followers in the in-group may model the leader’s behavior and reject out-group followers’ ideas or suggestions, potentially reducing the overall quality of the decision or the decision-making process. When equality or group-level leadership is not practiced, the next best avenue for decision optimization may be a leadership style that recognizes each follower’s capacity and potential contribution, not solely the in-group followers’ capacity and potential contributions. The simulation also revealed that it takes considerable effort, even in the perfect team condition, to reach decision optimization. After approximately 10,000 interactions in which no team member accepts a decision that does not optimize the team utility and no team member acts in a self- interested way, the team will converge toward an optimal decision utility. An interesting finding from the simulation is that about halfway to 10,000- iteration point, the participative, individualized, and LMX leadership styles all reach their best optimization point. Although these points do not represent maximized decision points for these hierarchical groups, they do signal a point of diminishing returns – that is, any additional ‘‘time’’ spent on decision making beyond this point will not produce a better solution. This finding has significant practical implications. If hierarchical teams do not exist in the perfect team environment, then once an initial rapid decent is achieved toward the optimum, it may not make sense to devote much more time to reaching the optimal decision. Without some significant change or crisis (i.e., noise) inflicted on the process, hierarchical groups in participative, individualized, and LMX leadership conditions can expect to plateau and likely never reach the optimum. A skilled leader working within these conditions might recognize the initial plateau point and accept a group decision, thereby not wasting additional time to produce exactly the same decision at some later point. Finally, it would appear that a levels-based focus, given the group decision-making task, might make a significant difference in terms of a hierarchical group’s ability to optimize its decisions. In the simulation, alignment between the task requirement, group decision, and leadership level of focus (i.e., group based) produced the best optimization scenario. Therefore, when a leader needs to facilitate a group decision, he or she might consider employing a group-based leadership style to maximize the potential for optimization. 262 SHELLEY D. DIONNE AND PETER J. DIONNE

An interesting caveat to the preceding recommendation concerns the ﬁnding that an authoritative decision by the leader may produce the most optimal outcome when the group needs an immediate decision. As evidenced in both production runs, the authoritative leadership scenario reached the best decision during the initial interaction timeframe. Although the authoritative decision was generally far from optimal, it was closer to the optimum than the decisions made in all of the other leadership scenarios. This outcome suggests that in an emergency or urgent situation, the leader’s decision, rendered on behalf of the group, has the best chance for decision optimization at that moment. Of course, all other leadership conditions also quickly move toward more optimal decisions, so leaders may want to reserve this style for extreme situations.

Limitations

As illustrated in the differences between Production Runs 2–4 and Production Runs 5–7, a limitation of this study is the notion that different assumptions produce different results. Varying assumptions – for example, considering a different IQ range or mean for a given ‘‘sample,’’ including other personality factors such as extroversion or conscientiousness, or adding to or subtracting from the number of group members – has the potential to produce significantly different results. Therefore, our interpretations are limited to the conditions represented by the model, which naturally restricts the usefulness of the results. Although our assumptions were selected based on existing theory and practical recommendations, they are merely a starting point for further investigations. Another key limitation of this research arises from the fact that the data were simulated. As a consequence, verification in the form of experimental research or field studies must be the next step in the research process so as to develop external validity and generalizability of the results. The applicability of the current results for leaders and group decision makers needs to be viewed through a conservative lens.

Future Research

Inasmuch as leadership involves an influence process with at least one other person, we should at least consider the possibility that levels of analysis issues may be operating throughout the process, depending on the level of Leadership and Group Decision 263 interdependency or interaction. First and foremost, leadership research needs to continue, at every opportunity, to address the relevance of levels of analysis, both theoretically and methodologically. Moreover, as methodological improvements expand our potential to view dynamic processes, we should consider the dynamic and multi-level implications within leadership research. Computer modeling is one way to capture both the dynamic and multi-level perspectives within research. Although it offers a powerful ability to model complex systems, computer modeling is no panacea for advancing leadership theory. The call for greater consideration of levels-based leadership research was intended to enhance scholarly thought surrounding leadership as a levels-based construct, which then could guide more methodologically appropriate testing of theory. Although computer modeling and simulation may be part of the overall process, all good research begins with good theory. Future research may want to examine further, from a multi-level perspective, the effects of different leadership behaviors on decision optimization within groups and teams. For example, one promising area of investigation is the effects of annealing leader behaviors on group optimization of decisions. Annealing – a type of algorithmic search condition that tends to move decisions increasingly toward the optimum – is a concept studied within complex systems examining decision optima (Bar-Yam, 1997; Klein, Faratin, Sayama, & Bar-Yam, 2003). Although annealing becomes difficult to implement when self-interested behaviors of multiple members are introduced, further multi-level theoretical development of this concept applied to a hierarchically based decision groups may advance our understanding of both multi-level leadership and group member actions in decision making. Once multi-level leadership theory is established, simulations may be employed as an initial examination of the process of annealing leader behavior within a hierarchical decision group. Afterward, experimental and field testing can occur, taking advantage of the information gained from a simulated model to better focus the execution of a study with actual organizational members. Future research might also address at least one of the limitations previously mentioned. Varying the assumptions employed in our model might enhance our understanding of other aspects of leadership and group behavior. For example, simulating the development of team mental models might produce different results in terms of teams reaching optimal decisions more quickly. We made no assumptions regarding the power of team-based cognitions, but the multi-level nature of the concept and the idea that a 264 SHELLEY D. DIONNE AND PETER J. DIONNE leader could influence the development of team mental models are areas ripe for further study. This investigation could enhance both multi-level leadership research as well as research into team performance and decision making. Finally, in a more general sense, modeling dynamic theory offers potential in many areas of leadership research, as researchers have begun to detect and examine nonlinear concepts in areas such as leadership emergence (Guastello, 2007; Zaror & Guastello, 2000), where a swallowtail catastrophe model highlights the control parameters, stable states and unstable state involved in the emergence process (Guastello, 2007). Ideally suited for nonlinear modeling, simulation methods offer greater flexibility in the early examination of theory where complex interactions and processes are not adequately developed with traditional linear theories and methodological approaches. Future leadership research may want to consider the nonlinear implications of levels-based leader/follower relations, especially within the rapidly changing environments of global organizations.

CONCLUSION

Hulin and Ilgen (2000) note that attempts to understand dynamic organizational behavior more closely resemble a slide show rather than an edited movie. This analogy highlights the notion that the results of traditional experimental and correlational research are snapshots taken during time intervals imposed on organizational processes, often arbitrarily and without regard to a theory of organizational time. The conclusions drawn from this traditional process may therefore be accurate but not true, as the ‘‘cadence of social and behavior processes may not unfold linearly with respect to the clock-time continuum’’ (Hulin & Ilgen, 2000, p. 6). This article presented the results of a dynamic simulation of levels-based leadership’s inﬂuence on hierarchical group decision making. Like any methodology, simulation has its shortcomings, though the same can be said for traditional experimental and correlational methodologies as well. We suggest that the best investigations of complex and dynamic leadership phenomena will employ both simulation and traditional methodologies. When this research is combined with a strong grounding in theoretical levels, we can continue to advance our understanding of leadership within organizations in today’s complex environments. Leadership and Group Decision 265

ACKNOWLEDGMENTS

Signiﬁcant portions of this article were published in Leadership Quarterly, Vol. 19, Shelley D. Dionne, and Peter J. Dionne, Levels-based leadership and hierarchical group decision optimization: A simulation, pp. 212–234, Copyright Elsevier 2008.

REFERENCES

Abdelaziz, F. B., Martel, J. M., & Mselmi, A. (2004). IMGD: An interactive method for multiobjective group decision aid. Journal of Operational Research Society, 55, 464–474. Adner, R. (2002). When are technologies disruptive? A demand-based view of the emergence of competition. Strategic Management Journal, 23, 667–688. Bar-Yam, Y. (1997). Dynamics of complex systems. Reading, MA: Addison-Wesley. Beck, M. P., & Lin, B. W. (1983). Some heuristics for the consensus ranking problem. Computers and Operations Research, 10, 1–7. Black, J. A., Oliver, R. L., Howell, J. P., & King, J. P. (2006). A dynamic system simulation of leader and group effects on context for learning. Leadership Quarterly, 17(1), 39–56. Bottger, P. C. (1984). Expertise and air time as basis of actual and perceived influence in problem solving groups. Journal of Applied Psychology, 69, 214–221. Bottger, P. C., & Yetton, P. W. (1988). An integration of process and decision scheme explanations of group problem solving and performance. Organizational Behavior and Human Decision Processes, 42, 234–249. Budescu, D. V., & Rantilla, A. K. (2000). Confidence in aggregation of expert opinions. Acta Psychologica, 104, 371–398. Budescu, D. V., Rantilla, A. K., Yu, H. T., & Krelitz, T. M. (2003). The effects of asymmetry among advisors on the aggregation of their opinions. Organizational Behavior and Human Decision Processes, 90(1), 178–194. Burger, J. M., Messian, N., Patel, S., del Prado, A., & Anderson, C. (2004). What a coincidence! The effects of incidental similarity on compliance. Personality and Social Psychology Bulletin, 30, 35–43. Carroll, G., & Harrison, J. R. (1998). Organizational demography and culture: Insights from a formal model and simulation. Administrative Science Quarterly, 43, 637–667. Chattoe, E. (1998). Just how (un)realistic are evolutionary algorithms as representations of social processes? Journal of Artificial Social Science Simulation, 1(3), 2.1–2.36. Cook, W. D., & Kress, M. (1985). Ordinal ranking with intensity of preference. Management Science, 39, 218–226. Costa, P. T., Jr., & McCrea, R. R. (1991). NEO five-factor inventory. Odessa, FL: Psychological Assessment Resources. Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory testing in organizational behavior: The varient approach. Englewood Cliffs, NJ: Prentice-Hall. Dansereau, F., & Yammarino, F. J. (1998). The multi-level approaches to leadership: Introduction and overview. In: F. Dansereau & F. J. Yammarino (Eds), Leadership: The multi-level approaches (Part A, pp. xxv–xliii; Part B, pp. xxv–xliii). Stamford, CT: JAI Press. 266 SHELLEY D. DIONNE AND PETER J. DIONNE

Dansereau, F., Yammarino, F. J., & Kohles, J. (1999). Multiple levels of analysis from a longitudinal perspective: Some implications for theory building. Academy of Manage- ment Review, 24, 346–357. Dansereau, F., Yammarino, F. J., Markham, S. E., Alutto, J. A., Newman, J., Duman, M., et al. (1995). Individualized leadership: A new multiple-level approach. Leadership Quarterly, 6(3), 413–450. Davis, J. H. (1996). Group decision making and quantitative judgments: A consensus model. In: E. Witte & J. H. Davis (Eds), Understanding group behavior: Consensual actions by small groups (Vol. 1, pp. 35–59). Mahwah, NJ: Erlbaum. Davis, J. P., Eisenhardt, K. M., & Bingham, C. B. (2007). Developing theory through simulation methods. Academy of Management Review, 32, 480–499. Dienesch, R. M., & Liden, R. C. (1986). Leader–member exchange model of leadership: A critique and further development. Academy of Management Review, 11, 618–634. Dionne, P. J., Ozisik, R., & Picu, C. (2005). Structure and dynamics of polyethylene nanocomposites. Macromolecules, 38, 9351–9358. Dionne, P. J., Picu, C., & Ozisik, R. (2006). Adsorption and desorption dynamics of linear polymer chains to spherical nanoparticles: A Monte Carlo investigation. Macromole- cules, 39, 3089–3092. Dionne, S. D., Yammarino, F. J., Atwater, L. E., & Spangler, W. D. (2004). Transformational leadership and team performance. Journal of Organizational Change Management, 17(2), 177–193. Einhorn, H. J., Hogarth, R. M., & Klempner, E. (1977). Quality of group judgment. Psychological Bulletin, 84, 158–172. Fine, G. A., & Elsbach, K. D. (2000). Ethnography and experiment in social psychologist theory building. Journal of Experimental Social Psychology, 36, 51–76. Garvin, D. A. (1987). Competing on the eight dimensions of quality. Harvard Business Review, 65(6), 101–109. Gottfredson, L. S. (1997a). Mainstream science on intelligence: An editorial with 52 signatures, history and bibliography. Intelligence, 24(1), 13–23. Gottfredson, L. S. (1997b). Why g matters: The complexity of everyday life. Intelligence, 24(1), 79–132. Graen, G. B., & Cashman, J. (1975). A role making model of leadership in formal organizations: A developmental approach. In: J. G. Hunt & L. L. Larson (Eds), Leadership frontiers (pp. 143–165). Kent, OH: Kent State University Press. Graen, G. B., & Uhl-Bien, M. (1995). Relationship-based approach to leadership: Development of leader–member exchange (LMX) theory of leadership over 25 years: Applying a multi-level multi-domain perspective. Leadership Quarterly, 6, 219–247. Guastello, S. J. (2007). Non-linear dynamics and leadership emergence. Leadership Quarterly, 18, 357–369. Hanisch, K. A. (2000). The impact of organizational interventions on behaviors: An examination of models of withdrawal. In: C. L. Hulin & D. R. Ilgen (Eds), Computational modeling of behavior in organizations: The third scientiﬁc discipline (pp. 33–60). Washington, DC: American Psychological Association. Hanisch, K. A., Hulin, C. L., & Seitz, S. T. (1996). Mathematical/computational modeling of organizational withdrawal processes: Beneﬁts, methods, and results. In: G. Ferris (Ed.), Research in personnel and human resource management (Vol. 14, pp. 91–142). Greenwich, CT: JAI Press. Leadership and Group Decision 267

Harvey, N., Harries, C., & Fischer, I. (2000). Using advice and assessing its quality. Organizational Behavior and Human Decision Processes, 81, 252–273. Hazy, J. K. (2007). Computer models of leadership: Foundations for a new discipline or meaningless diversion? Leadership Quarterly, 18, 391–410. Hofmann, D. A., & Jones, L. M. (2005). Leadership, collective personality, and performance. Journal of Applied Psychology, 90(3), 506–522. Hollenbeck, J. R., Ilgen, D. R., Sego, D. J., Hedlund, J., Major, D. A., & Phillips, J. (1995). Multi-level theory of team decision making: Decision performance in teams incorporat- ing distributed expertise. Journal of Applied Psychology, 80(2), 292–316. House, R. J., Rousseau, D. M., & Thomas-Hunt, M. (1995). The meso paradigm: A framework for the integration of micro and macro organizational behavior. Research in Organizational Behavior, 17, 71–114. Hulin, C. L., & Ilgen, D. R. (2000). Computational modeling of behavior in organizations: The third scientiﬁc discipline. Washington, DC: American Psychological Association. Jago, A. G., & Vroom, V. H. (1980). An evaluation of two alternatives to the Vroom/Yetton normative model. Academy of Management Journal, 23(2), 347–355. Jensen, A. R. (1981). Straight talk about mental tests. New York: Free Press. Jones, J. T., Pelham, B. W., Carvallo, M., & Mirenberg, M. C. (2004). How do I love thee? Let me count the Js: Implicit egotism and interpersonal attraction. Journal of Personality and Social Psychology, 87(5), 665–683. Kameda, T., Ohtsubo, Y., & Takezawa, M. (1997). Centrality in socio-cognitive network and social inﬂuence: An illustration in a group decision making context. Journal of Personality and Social Psychology, 73, 296–309. Keeney, R., & Raiffa, H. (1976). Decisions with multiple objectives: Preferences and value trade offs. New York: Wiley. Keller, T., & Dansereau, F. (1995). Leadership and empowerment: A social exchange perspective. Human Relations, 48, 127–145. Kemeny, J. G., & Snell, L. J. (1962). Mathematical models in the social sciences. Blaisdell, NY: Ginn & Co. Kerr, N. L., & Tindale, R. S. (2004). Group performance and decision making. Annual Review of Psychology, 55, 623–655. Klein, K. J., Dansereau, F., & Hall, R. J. (1994). Levels issues in theory development, data collection, and analysis. Academy of Management Review, 19, 195–229. Klein, M., Faratin, P., Sayama, H., & Bar-Yam, Y. (2003). Protocols for negotiating complex contracts. IEEE Intelligent Systems, 18(6), 32–38. Koopman, P. L., & Wierdsma, A. F. M. (1998). Participative management. In: P. J. D. Doentu, H. Thierry & C. J. de-Wolf (Eds), Personnel psychology: Handbook or work and organizational psychology (Vol. 3, pp. 297–324). Hove, UK: Psychology Press/ Erlbaum. Lant, T., & Mezias, S. (1992). Managing discontinuous change: A simulation study of organizational learning and entrepreneurship. Strategic Management Journal, 11, 147–179. Libby, R., Trotman, K. T., & Zimmer, I. (1987). Member variation, recognition of expertise, and group performance. Journal of Applied Psychology, 72, 81–87. Liden, R. C., Wayne, S. J., & Stilwell, D. (1993). A longitudinal study on the early development of leader–member exchanges. Journal of Applied Psychology, 78(4), 662–674. Likert, R. (1977). The nature of highly effective groups. In: J. L. Gray & F. A. Starke (Eds), Readings in organizational behavior (pp. 100–129). Columbus, OH: Merrill. 268 SHELLEY D. DIONNE AND PETER J. DIONNE

Littlepage, G. E., & Silbiger, H. (1992). Recognition of expertise in decision-making groups: Effects of group size and participation patterns. Small Group Research, 23(3), 344–355. McCrea, R. R., & Costa, P. T., Jr. (1990). Personality in adulthood. New York: Guilford Press. Mersha, T., & Adlakha, V. (1992). Attributes of service quality: The consumers’ perspective. International Journal of Service Industry Management, 3(3), 34–45. Metropolis, N., Rosenbluth, A. W., Rosenbluth, M. N., Teller, A. H., & Teller, E. (1953). Equation of state calculations by fast computing machines. Journal of Chemical Physics, 21, 1087–1092. Miller, D. T., Downs, J. S., & Prentice, D. A. (1998). Minimal conditions for the creation of unit relationship: The social bond between birthdaymates. European Journal of Social Psychology, 28, 475–481. Repenning, N. (2002). A simulation-based approach to understanding the dynamics of innovation implementation. Organizational Science, 13, 109–127. Richard, M. D., & Allaway, A. W. (1993). Service quality attributes and choice behavior. Journal of Services Marketing, 7(1), 59–68. Rousseau, D. M. (1985). Issues of level in organizational research: Multi-level and cross-level perspectives. Research in Organizational Behavior, 7, 1–37. Saavedra, R. (1990). Beer sales and delivery teams. In: J. R. Hackman (Ed.), Groups that work (and those that don’t): Creating conditions for effective teamwork (pp. 361–381). San Francisco, CA: Jossey-Bass. Salas, E., Dickinson, T. L., Converse, S. A., & Tannenbaum, S. I. (1992). Toward an understanding of team performance and training. In: R. W. Swezey & E. Salas (Eds), Teams: Their training and performance (pp. 3–29). Norwood, NJ: Ablex. Salas, E., Guthrie, J. W., Jr., Wilson-Donnelly, K. A., Priest, H. A., & Burke, C. S. (2005). Modeling team performance: The basic ingredients and research needs. In: W. B. Rouse & K. R. Boff (Eds), Organizational simulation (pp. 185–228). Hoboken, NJ: Wiley. Salas, E., Stagl, K. C., & Burke, C. S. (2004). 25 years of team effectiveness in organizations: Research themes and emerging needs. International Review of Industrial and Organiza- tional Psychology, 19, 47–91. Scandura, T. A., Graen, G. B., & Novak, M. A. (1986). When managers decide not to decide autocratically: An investigation of leader–member exchange and decision inﬂuence. Journal of Applied Psychology, 71, 579–584. Schriesheim, C. A., Castro, S. L., & Yammarino, F. J. (2000). Investigating contingencies: An examination of the impact of span of supervision and upward controllingness on leader– member exchange using traditional and multivariate within- and between-entities analysis. Journal of Applied Psychology, 85, 659–677. Schriesheim, C. A., Castro, S. L., Zhou, X. T., & Yammarino, F. J. (2001). The folly of theorizing ‘‘A’’ but testing ‘‘B’’: A selective level-of-analysis review of the ﬁeld and detailed leader–member exchange illustration. Leadership Quarterly, 12, 515–551. Schriesheim, C. A., Neider, L. L., & Scandura, T. A. (1998). Delegation and leader–member exchange: Main effects, moderators, and measurement issues. Academy of Management Journal, 41, 298–318. Schwab, D. P., & Olsen, C. O. (1990). Merit pay practices: Implications for pay-for- performance relationships. Industrial and Labor Relations Review, 43, 237–255. Schwab, D. P., & Olsen, C. O. (2000). Simulating effects of pay-for-performance systems on pay-for-performance relationships. In: C. L. Hulin & D. R. Ilgen (Eds), Computational Leadership and Group Decision 269

modeling of behavior in organizations: The third scientific discipline (pp. 115–127). Washington, DC: American Psychological Association. Scully, J. A., Kirkpatrick, S. A., & Locke, E. A. (1995). Locus of knowledge as a determinant of the effect of participation on performance, affect, and perceptions. Organizational Behavior and Human Decision Processes, 61, 276–288. Sniezek, J. A. (1992). Groups under uncertainty: An examination of confidence in group decision making. Organizational Behavior and Human Decision Processes, 52(1), 124–155. Sniezek, J. A., & Buckley, T. (1995). Cueing and cognitive conflict in judge–advisor decision making. Organizational Behavior and Human Decision Processes, 62, 159–174. Somech, A. (2003). Relationships of participative leadership with relational demography variables: A multi-level perspective. Journal of Organizational Behavior, 24, 1003–1018. Stasser, G. (1988). Computer simulation as a research tool: The DISCUSS model of group decision-making. Journal of Experimental Social Psychology, 24, 393–422. Stasser, G. (1990). Computer simulation of social interaction. In: C. Hendrick (Ed.), Review of personality and social psychology: Research methods in personality and social psychology (Vol. 11, pp. 120–141). Newbury Park, CA: Sage. Stasser, G. (2000). Information distribution, participation, and group decision: Explorations with the DISCUSS and SPEAK models. In: C. L. Hulin & D. R. Ilgen (Eds), Computational modeling of behavior in organizations: The third scientific discipline (pp. 135–156). Washington, DC: American Psychological Association. Tindale, R. S., Filkins, J., Thomas, L. S., & Smith, C. M. (1993). An attempt to reduce conjunctive errors in decision-making groups. Presented at the Annual Meeting of Social Judgment and Decision Making, Washington, DC. Tindale, R. S., & Kameda, T. (2000). ‘‘Social sharedness’’ as a unifying theme for information processing in groups. Group Process: Intergroup Relations, 3, 123–140. Torrance, G. W., Boyle, M. H., & Horwood, S. P. (1982). Applications of multi-attribute utility theory to measure social preferences for health setting. Operations Research, 30(6), 1043–1069. Vecchio, R. R. (1982). A further test of leadership effects due to between-group variation and within-group variation. Journal of Applied Psychology, 67, 200–208. Vroom, V. H., & Jago, A. G. (1988). The new leadership: Managing people in organizations. Englewood Cliffs, NJ: Prentice-Hall. Vroom, V. H., & Yetton, P. W. (1973). Leadership and decision-making. Pittsburgh, PA: University of Pittsburgh Press. Wonderlic Personnel Test Inc.. (1992). Wonderlic personnel test and scholastic level exam: User’s manual. Libertyville, IL: Author. Yammarino, F. J., & Dansereau, F. (2002). The many faces of multi-level issues. Research in multi-level issues (Vol. 1). Oxford, UK: Elsevier Science. Yammarino, F. J., Dionne, S. D., & Chun, J. U. (2002). Transformational and charismatic leadership: A levels-of-analysis review of theory, measurement, data analysis, and inferences. In: L. L. Neider & C. A. Shreishiem (Eds), Leadership research in management (Vol. 2, pp. 23–63). Greenwich, CT: Information Age Publishing. Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. Leadership Quarterly, 16, 879–919. Yammarino, F. J., & Naughton, T. J. (1992). Individualized and group-based views of participation in decision making. Group and Organization Management, 17, 398–413. 270 SHELLEY D. DIONNE AND PETER J. DIONNE

Zaror, G., & Guastello, S. J. (2000). Self-organization and leadership emergence: A cross- cultural replication. Nonlinear Dynamics, Psychology, and Life Sciences, 4, 113–119. Zickar, M. J. (2000). Modeling faking on personality tests. In: C. L. Hulin & D. R. Ilgen (Eds), Computational modeling of behavior in organizations: The third scientiﬁc discipline (pp. 95–108). Washington, DC: American Psychological Association. Zickar, M. J., & Robie, C. (1999). Modeling faking good on personality items: An item-level analysis. Journal of Applied Psychology, 84, 551–563. Zionts, S., & Wallenius, J. (1976). An interactive programming method for solving multiple criteria problems. Management Science, 22, 652–663. Zott, C. (2003). Dynamic capabilities and the emergence of intra-industry differential ﬁrm performance: Insights from a simulation study. Strategic Management Journal, 24, 97–125.

COMPARING SIMULATION RESULTS OF LEADERSHIP STYLE IMPACTS ON EMERGENT VERSUS SPECIFIC TASK OUTCOMES AND REQUIRED SIMULATION MODEL COMPONENTS

Janice A. Black, Richard L. Oliver and Lori D. Paris

ABSTRACT

The clear speciﬁcation of leadership efforts spanning levels of analysis has lagged behind leadership research in general. Simulation modeling, such as agent-based modeling, provides research platforms for exploring these interesting issues. This chapter uses agent-based models, along with Dionne and Dionne’s (2009) choices of leadership styles, to examine the impact of those styles on the generation of an emergent group resource, context-for-learning (CFL), instead of the speciﬁc task outcome (group decision making) described by Dionne and Dionne. Consistent effectiveness is found across leadership styles for workgroups with high and slightly lower initial individual levels of a CFL. A second agent-based model includes the ability of agents to forget previous learned skills and

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 271–299 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008012 271 272 JANICE A. BLACK ET AL.

reveals a reduced effectiveness of all leadership styles. However, the effectiveness of the leadership styles differs between the two outcomes (the speciﬁc group task model and the emergent group resource model). Reasons for these differences are explored, and implications from the comparisons of the two models are delineated.

INTRODUCTION

As Dionne and Dionne (2009) accurately point out, specifying the correct level of analysis, whether at a specific level or for multiple levels, has long been called for in a variety of organizational research areas (Dansereau, Alutto, & Yammarino, 1984; Rousseau, 1985) and continues to be a critical component in leadership research (Gittell & Weiss, 2004; Yammarino, Dionne, Chun, & Dansereau, 2005). Many organizational phenomena, ranging from leadership studies (Yammarino et al., 2005) to intra- and inter-organizational coordination (Gittell & Weiss, 2004), are recognized as inherently involving multiple levels of analysis, but we agree with Dionne and Dionne that existing research is not sufficient in including and specifying what is happening at each level of analysis. Research is further confounded when multiple levels are involved in the generation of phenomena at any particular level of analysis (Black, Oliver, Howell, & King, 2006). However, simulation-based investigations and, in particular, agent-based modeling (ABM) help to address this issue of unclear specifications (Davis, Eisenhardt, & Bingham, 2007; Black et al., 2006)by providing explicit level identification and identification of cross-level influences. Simulations are excellent tools for examining and building upon existing theory and providing ways to explore system effects, particularly complex system effects (Black et al., 2006). Dionne and Dionne, in their current effort, correctly point out the benefits of simulation modeling in examining cross-level leadership phenomena and their impact on group-level dynamics, particularly group decision making. We wondered if the same four leadership styles used by Dionne and Dionne would provide similar results on the development of a context-for- learning (CFL), which is an emergent group resource, rather than a specific task outcome such as reaching a group decision. We are also interested in determining whether forgetting might be a critical component to include in cross-level social science models, as is suggested by some researchers Comparing Simulation Results 273

(de Holan, Phillips, & Lawrence, 2004; Kriegesmann, Kley, & Schwering, 2005), but not included in their discussion by Dionne and Dionne. To examine these issues, we ﬁrst summarize the results that Dionne and Dionne found in their simulation of a speciﬁc task outcome using their set of required and desired components in modeling. We then present a summary of the results that Black and Oliver found in their earlier CFL simulation work, which examined an emergent group outcome using their set of required and variable components (Black et al., 2006). We modify that work to include leadership style as suggested by leader–member exchange theory (LMX), which is a component not included in the earlier CFL model. We then compare the two modeling efforts and their results and add forgetting by an agent, which is a new component suggested as being critically important in social sciences efforts. We demonstrate the effect of the forgetting construct on the simulation work done by Black et al. (2006) and suggest it as an additional – and necessary – component to the modeling orientation used by Dionne and Dionne.

THE SPECIFIC GROUP TASK MODEL: DIONNE AND DIONNE

Dionne and Dionne examine a speciﬁc task outcome when they focus on group decision-making point optimization. They base their examination of this phenomenon on the statement of required versus desired components for team performance modeling proposed by Salas, Guthrie, Wilson- Donnelly, Priest, and Burke (2005). Dionne and Dionne operationalize these components described in the literature, noting how they apply to group decision making, leadership styles, and the associated implied cross-level manifestations. Along the way, these authors examine four leadership-based theories (autocratic leadership, individual and dyadic-based leadership interactions, and participative leadership), each of which addresses a particular single-level or cross-level inﬂuence structure. Dionne and Dionne choose autocratic leadership to represent leadership decision making at an individual level of analysis. Autocratic leaders make decisions for their workgroup without follower participation (Vroom & Yetton, 1973). The term group is used loosely, as more focus is placed on decision making that follows along hierarchical levels than on true group interaction and participation. In Dionne and Dionne’s work, individualized leadership and LMX theory are used at the dyadic levels of analysis. 274 JANICE A. BLACK ET AL.

According to these authors, individualized leadership is a multi-level approach whereby a particular leader and a particular follower form a relationship that is independent of other relationships they form in their group. LMX theory represents the development of the leader–follower exchange relationship over time. According to LMX theory, participation in the decision-making process depends on the quality of the relationship between the leader and the follower. Followers considered part of the in-group, have higher-quality exchanges with their leader, and enjoy more control and influence over the decision-making process than members of the out-group, who have lower-quality exchanges with their leader. According to Dionne and Dionne, the assignment to either an in-group or an out-group is based on how much extra effort an individual is putting forth in pursuing the group’s goals. Finally, participative leadership is used at the group level of analysis. Group members engage in joint problem solving and decision making with their leader, because the leader respects members’ knowledge and intelligence. Clearly specifying all of the components that go into a model is important for simulation modeling (Davis et al., 2007; Black et al., 2006). Dionne and Dionne specify that a group in their model consists of a group of intelligent professional workers who possess the necessary competencies needed for group decision making. They imply that this group is homogenous in its skill levels and has a high level of skill in these competencies. In their work, Dionne and Dionne had group members randomly influencing each other in a two-by-two fashion. Over time, the decision preferences of the group members changed due to this randomly paired influence. By contrast, in the autocratic leader’s group, group member attributes were not used to influence the group decision. Dionne and Dionne’s findings conclude that the final order of effective leadership styles in creating an optimal decision-making effort is (from most to least effective) participative, individualized, LMX, and autocratic. Thus, for the highly intelligent professional worker groups, the most effective or optimal decision making occurs with a participative decision-making style and the least effective decision making accompanies an autocratic leadership style. In their chapter, the results of Dionne and Dionne’s simulation are presented in a series of graphs. These graphs highlight the distance between the group decision point at a specific iteration versus the optimal decision point for that iteration. Thus better results are indicated by a line that approaches zero. In Dionne and Dionne’s study, all leadership styles resulted in graphs that had plateaus, but those plateaus were at different levels (as indicated earlier). Comparing Simulation Results 275

EMERGENT GROUP RESOURCE: BLACK, OLIVER, HOWELL, AND KING

In similar studies, Black and associates (Black, Oliver, & King, 2004; Black et al., 2006; Black, King, & Oliver, 2005) have developed a computational model that examines the social creation of a group-level resource known as the group context-for-learning (group CFL). In the 2006 study, Black and associates focused on an emergent group resource CFL versus Dionne and Dionne’s specific group task. In their work, Black and associates used groups with varying levels of initial competence in the required components for the CFL as well as groups with varying levels of competence in managerial and leadership skill levels. They, too, employed an extensive literature review to determine the inputs and operationalizations. Black and associates also used other modeling criteria (Lewin, Parker, & Birute, 1998) to demonstrate the reasonableness of their modeling efforts. In earlier work examining the effect of leadership styles on the development of the CFL (Black et al., 2004), it was found that the specific skill sets of the followers, the perceptions of the leader, and the leader’s style were critical components in determining efficient CFL development. The leadership styles outlined by Black et al. (2004) were similar to the set of styles used by Dionne and Dionne in their examination of a specific group task of decision making. Black and colleagues’ analysis included four leadership styles: autocratic leadership, conventional leadership (where there is some influence on the leader’s decision by group members), democratic leadership (very similar to Dionne and Dionne’s participative leadership style), and no formal leader (members’ contribution only). However, rather than completely eliminating all influences of group members on the leader, Black and colleagues chose to reduce that influence down to a minor influence. Leaders and followers also influenced one another in Black and colleagues’ work, but that influence was modified by how similar or different their CFL profiles were to one another’s. People with similar profiles readily accepted influences from one another, whereas those with dissimilar profiles clung tightly to their own perspectives but allowed minimal influences by those who were different from themselves. Influence was also determined by each individual’s personal expertise. Expertise was the main contributing influence factor for the creation of a group CFL. Expertise levels changed over time depending on the team’s and the individual’s CFL profiles. Black and associates, in two separate examinations of the influence of leaders on the development of CFL, found that for groups consisting of individuals who initially had high levels of the components of a CFL, 276 JANICE A. BLACK ET AL. leadership style seemed somewhat irrelevant (Black et al., 2004). Only in those groups with low-CFL components did leadership styles and leadership skill levels begin to matter significantly (Black et al., 2006).

DUPLICATION OF LEADERSHIP STYLES AND THEIR EFFECTS ON CFL

The CFL construct has four components, which create both a profile (a set of values for individual components) and an index (a single value). Stated separately for individuals, the profile includes discipline, stretch, trust, and support (Ghoshal & Bartlett, 1994). Discipline assesses the presence of clear performance standards, fast feedback, open communication, and management by commitment. Stretch includes having a shared ambition for the future, a collective identity, and a personal link between individual and corporate goals. Trust requires the presence of perceived equity in decision making, involvement of people when decisions impact them and individual competence or expertise. Support is composed of having access to organization resources, the autonomy to make decisions and guidance and help from management. Averaged together for the individual, these four values create the CFL index. At the group level, the averaging is weighted by two influence parameters: the formal position of the individual and the individual’s personal expertise levels (Black et al., 2004, 2005). In their model, Black and colleagues also included experiential learning so that agents within the model could learn across time. The team and the leader influence the amount of learning any individual does by virtue of the degree of similarity that exists between the individual and either the team or the leader. With this explicit feedback loop, which goes from the team to the individual, and the use of the individual in calculating new team levels in each iteration, Black and associates have created a dynamic model that coevolves at the individual and group levels. Cross-level influences at both the dyadic and team levels are in place. Thus the CFL dynamically changes over time at both the individual and group levels. In modeling leaders, Black and colleagues have used the competing values framework (CVF) developed by Quinn, Faerman, Thompson, and McGrath (2003). In this framework, four quadrants comprising two leadership roles are created along two axes: Control’-Flexibility and Comparing Simulation Results 277

Internal’-External foci. This structure results in a comprehensive set of behaviors that include both leadership and managerial skills (Bedeian & Hunt, 2006), but in which individuals experience a conflict (Quinn et al., 2003). For example, the Innovator and Broker roles compete or conflict with the Coordinator and Monitor roles. Likewise, the Producer and Director roles compete or conflict with the Facilitator and Mentor roles. A person engaged in these roles will tend to favor one set of behaviors over the other. Because of this tendency, Black and her colleagues modeled the actions taken by a leader as those actions indicated by their roles with the highest skill levels. As leaders take action, their actions form patterns that can be considered leadership styles. Black and her colleagues used the Aggressive Achiever pattern (Quinn et al., 2003) to model effective leaders. Leaders also experientially learn by using their CVF-associated skills. While Black and associates in their computational model of the CFL (Black et al., 2004) used similar leadership styles as Dionne and Dionne employ, they did not completely match those styles used by Dionne and Dionne. They also examined a wider set of team skill levels than the high level of IQ of the professional workgroup identified by Dionne and Dionne. In the current study, we modify the leadership styles to roughly match Dionne and Dionne’s operationalization of leadership styles, use a leader with a positive skill level profile, and, paralleling Dionne and Dionne, restrict the group to one with initially high levels of the components – in this case, high levels of CFL. Thus, for autocratic leadership, which Dionne and Dionne operationalized as consisting of leader input only, we modify the influence equation so that only the leader’s input counts. Because the CFL is a socially constructed resource, however, this extreme position is not at all reflective of the social interaction needed for its creation. To address this shortcoming, we relax this assumption to one where the leader’s perspective is the majority perspective and contributes 75% of the influence in the creation of the CFL. We model the individualized leadership style suggested by Dionne and Dionne as one where the workgroup and the leader contribute equally (50%–50%) to the CFL. Dionne and Dionne look at only leader influences on their group output. Because CFL is a group socially constructed resource, in this model group influences also affect individuals to the degree that individuals and groups are similar, as was found in earlier simulation work by Black et al. (2005, 2006). The participative leadership style suggested by Dionne and Dionne calls for a greater influence on the part of the individuals on the CFL. We will model this as the team members contributing 75% to the group CFL and the leader contributing only 25% to the group CFL. 278 JANICE A. BLACK ET AL.

In all three of these leadership styles, the influence creating the group-level construct is based on the leader’s position power and the personal power of individual team members (i.e., relative levels of personal expertise). In other words, within the proportion of influence allocated to individual team members, each member contributes in a weighted fashion based on his or her relative level of expertise. Dionne and Dionne also use expertise as an influence variable. The remaining leadership style that must be modeled is the LMX style. To add this style to the current simulation, we operationalize the influence relationships as follows: Forty percent of the influence is assigned to the leader, and 40% to the individuals who have scores on the stretch variable, which maps to that extra effort referred to by Dionne and Dionne as indicative of ‘‘in-group’’ membership. Thus a stretch score of 5 or higher qualifies that individual to be a member of the in-group. The remaining 20% of the influence is reserved for the rest of the individuals, if any are there. If none has a stretch score less than 5, then the weighting is 50% leader and 50% for the average across a set of followers. Stretch was used for this purpose because, in its initial presentation by Ghoshal and Bartlett (1994),it was described as including those components found when individuals put forth significant extra effort.

COMPARING THE GROUP TASKS AND THE EMERGENT GROUP RESOURCE MODELS

Recall that our goal is to follow the pattern set by Dionne and Dionne in their speciﬁc task model so as to determine the most effective leadership styles to use for our emergent resource. Given a leader with high managerial and leadership skill levels along with a group composed of individuals with initially higher levels of the components of a CFL (a high-CFL group), we hypothesize that the efﬁciency in developing CFL will be the result of the leadership style pattern as found by Dionne and Dionne. The steepest developmental path results in reaching the optimal level of a CFL in the fewest iterations and can be considered the most effective result. The shallowest path reaches the optimal level using the most iterations and can be considered the least effective result.

Hypothesis 1. Given a high-CFL group, all leader styles will enable a steep developmental path of a group CFL. Comparing Simulation Results 279

Hypothesis 2. Given a high-CFL group, a leader using a participative leadership style will enable the steepest developmental path of a group CFL. Hypothesis 3. Given a high-CFL group, a leader using an individualized leadership style will enable a developmental path of a group CFL slightly less steep than the participative leadership style from Hypothesis 1. Hypothesis 4. Given a high-CFL group, a leader using a LMX leadership style will enable the third steepest developmental path of a group CFL. Hypothesis 5. Given a high-CFL group, a leader using an autocratic leadership style will enable the shallowest developmental path of a group CFL. When Dionne and Dionne adjusted their values for their professionals’ IQ levels (i.e., adjusted skill levels), they found that the same pattern persisted. We also adjust for that level change and examine a workgroup with a slightly lower initial level of CFL. Past experience with CFL models has shown that the level of the group does make a difference in the number of iterations it takes to reach the optimal level of CFL. We expect then to duplicate Dionne and Dionne’s findings and find that the order of efficiency remains the same but that the entire developmental paths start from a lower initial value and will be shallower than the developmental paths of the high- CFL groups. Thus we propose the following: Hypothesis 6. Given an initially lower value of CFL for a workgroup, the developmental paths associated with all leadership styles will begin lower and be shallower than the corresponding developmental paths for a high-CFL group. Hypothesis 7. The order of steepness of the developmental paths will follow the pattern identified in Hypotheses 2–5.

FORGETTING AS A NECESSARY COMPONENT

Dionne and Dionne make an excellent case for adhering to research standards set by Salas et al. (2005). They note which items must be included in any team performance model and which items are merely recommended or desired to be included in modeling team efforts. However, one such required component is missing: When learning is either explicitly or implicitly 280 JANICE A. BLACK ET AL. a part of the model, a required component is forgetting (de Holan et al., 2004; Kriegesmann et al., 2005). Scholarly work has been conducted on learning and forgetting at both the individual level (Shafer, Nembhard, & Uzumeri, 2001) and the organizational level (de Holan & Phillips, 2004). Forgetting follows predictable relationships between time passed and incident of learning (Sikstrom, 2002). At the individual level, people certainly can and do forget both intentionally and unintentionally. People also can and do forget things that they have learned in the past and things that they are currently learning. Furthermore, forgetting can occur with both tangible learning (muscle-based learning such as riding a bike or other physical labor), intangible learning (verbal and interaction-based learning such as observing others or discussions), and experiences (Nembhard & Osothsilp, 2001). Learning experiences may be both gathered and dispersed at differing rates. Because socially constructed intangible resources (e.g., cultures, climates, esprit de corps, reputations, creativity) emerge from individuals as they interact and learn (Black et al., 2006; Miller, Zhao, & Calantone, 2006), for the present study we focus on experiential learning. In experiential learning, when a sufficient break has occurred, forgetting begins (Nembhard & Osothsilp, 2001). The amount of learning prior to the break and the length of the break affect the amount of forgetting. The literature includes both power function and recency models of forgetting, although the recency model provides the best fit when a detailed history of interactions is available (Nembhard & Osothsilp, 2001). Furthermore, the successful use of a recency model in an organizational process simulation (Shafer et al., 2001) increases our confidence that this model will work best in this situation. Recency models examine the length of time between the last learning experience and the current experience. Because a break in experiential activities is required for the initiation of forgetting, we assign a lag of three time periods (roughly representing three months) before forgetting will occur (Anderlohr, 1969; Globerson, Nahumi, & Ellis, 1998) and use the same slope associated with a traditional S-shaped learning curve to determine the amount of forgetting that will occur (Jaber & Kher, 2004). If a leader has taken action (i.e., the leader’s behavior has occurred and the follower noticed it) within the last three iterations, no forgetting will occur. If either the leader did not take action (no leadership behavior occurred) or the follower did not notice it within the last three iterations, then forgetting at the rate of last learning will occur. It is important to note that each agent has a personalized placement on that learning curve and, therefore, will learn as individuals from their personal starting points. For the emergent Comparing Simulation Results 281 group resource model, all individuals may experience forgetting. In summary, individuals may forget in any of the components of CFL and leaders can forget in the CFL or in their CVF-associated skills. When one forgets, one’s progress toward a goal slows. The main effect of including forgetting into the present model, then, is expected to be that the number of iterations that it takes on average to reach the optimal level of CFL will be increased.

Hypothesis 8. All developmental paths from Hypotheses 1 and 6, with forgetting as an integral component of the model, will start at the same level but will take more iterations (be shallower) to reach the optimal score of 7.

Hypothesis 9. The order of steepness of the developmental paths will follow the pattern set in Hypotheses 2–5.

METHOD

Computational modeling is one of the most popular tools of complex systems research (Carley, 1995), and ABM is a particularly appropriate means by which to study social phenomena (Anderson, 1999; March, 1991; Miller et al., 2006). When creating a case study, ABM data can be obtained in the real world and then verified by comparing ABM results with the real- world results (Jacobsen & House, 2001). Alternatively, virtual data that are constructed to meet specific requirements can be used to verify that the ABM and the theoretical logic behind the model are adequately represented and that there are no hidden sensitivities or system effects that need to be included (March, 1991; Miller et al., 2006; Yin, 2003). This chapter, similar to the Dionne and Dionne chapter, uses virtual data and examines the computational and theoretical models for hidden sensitivities and/or system effects. In other words, the data used are generated by the computer according to specific boundary conditions. Therefore, when we discuss a leader or a follower, we are referring to virtual leaders and followers – not actual people.

CHANGES TO THE MODEL

ABM for social systems requires the specification of the relationships between and among agents and the outcomes of the social system. While all else can be held ceterus paribus, in a virtual simulation such as ABM, some 282 JANICE A. BLACK ET AL. changes merely require changing the values of variables, whereas other changes actually require changing the computational model. This work requires changing the computational model in the following ways: 1. When calculating the value for the emergent group CFL, the weights are changed to reflect the particular leadership styles. Autocratic leadership is represented by 75% leader’s CFL þ 25% group members’ CFL (where each group member contributes according to his or her own individual contribution to total expertise). Participative leadership is represented by 25% leader’s CFL þ 75% group members’ CFL (where each group member contributes according to his or her own individual contribution to total expertise). Individualized leadership is represented by 50% leader’s CFL þ 50% group members’ CFL (where each group member contributes according to his or her own individual contribution to total expertise) LMX leadership style is represented by 40% leader’s CFL þ 40% group members’ CFL (using the average CFL of all group members whose stretch score is 5 or higher) þ 20% group members’ CFL (using the average CFL of all group members whose stretch score is less than 5). The values of the individual CFL index for the high-CFL group members will average greater than 5; the values for the lower-CFL group will average between 3 and 5. 2. We add the forgetting module: First, data are recorded for each separate component of the CFL and CVF – that is, whether the individual or leader took action, and therefore learned, in that area. After three iterations, the forgetting function computes a negative influence on separate CFL and CVF components if no learning occurred in that component area for three consecutive iterations. Forgetting is determined by (1 learning function) and uses the existing learning function already calculated in the simulation.

INPUT DATA FOR AGENTS

Input Data for the Leader

A leader is defined by his or her position power, CFL profile, CVF profile, and recency profile which is calculated from the first three iterations. The CFL profile is composed of 13 embedded resource elements that are Comparing Simulation Results 283 synthetically generated to meet the general group parameters for a particular hypothesis. The CVF profile comes from the values defining an Aggressive Achiever, as specified by Quinn et al. (2003).

Input Data for Followers

We synthetically generate the data for followers to meet the parameters of the high-CFL group. Following earlier work by Black et al. (2006), this means that the individual’s CFL indices will be greater than 5. Individual components of the CFL may be less than 5, but their average must be greater than 5. We also synthetically generate an individual’s expertise level. The remaining identifying values are the learning histories for each of the embedded resource elements for the CFL proﬁles and expertise levels.

Input Data for the Group

Group-level data are generated in the simulation and reﬂect power inﬂuence relationships and leadership style. The CFL index is the average of the CFL resources of discipline, stretch, trust, and support. The group CFL index is calculated for each of the four leadership styles. Following Black et al.’s (2005, 2006) lead, the CFL index learning factor is extrapolated via an experiential learning S-curve and ranges from –2 to þ2; it represents the forgetting aspect when negative.

INPUT DATA FOR RELATIONSHIPS (LEARNING/ FORGETTING RELATIONSHIPS) Follower-to-Follower Relationship (Same-Level Inﬂuence): Follower’s Experiential Learning/Forgetting

Learning occurs in 14 elements: 13 CFL elements and an expertise element. Any follower is limited in the total number of elements in which learning can occur. A follower will learn in the ﬁrst ﬁve to nine areas affected by noticed leader behaviors in order of the leader’s skill in the behaviors. For any component, if no learning occurs in three consecutive iterations, the follower will experience forgetting in that component at the same rate as the associated learning rate but in a negative direction. 284 JANICE A. BLACK ET AL.

Group-to-Follower Relationship (Cross-Level Inﬂuence): Group’s Inﬂuence within an Iteration

Individuals do not move in or out of a group. Because group size is limited to eight members, including any particular follower, it is assumed that all group members are aware of and influenced by all other group members. The CFL profiles of the individual follower and the group are compared. If the profiles are similar, the group and follower will equally contribute to the follower’s experiential learning factor. If the profiles are dissimilar, then the follower’s experiential learning factor is emphasized as a 75% contribution and the group’s contribution is reduced to 25%. The experiential learning factor is joined with the leader-directed learning factor to create the final amount of learning for either an embedded CFL resource element or the expertise element.

Leader-to-Follower Relationship (Cross-Level Inﬂuence): Leader-Directed Learning

Black and associates developed an influence matrix that linked the leader’s behaviors with the embedded resource elements of the CFL profile (Black et al., 2005). This influence matrix is used in the present simulation as well. However, this influence occurs only if the follower notices the leader’s behavior. This component of noticing a leader is stochastically generated. The noticing component is also affected by the maximum number of elements in which the follower can learn for a time period. This maximum number is also stochastically generated for each learner and follows the adage of 7þ/–2 (Miller, 1956). Once that maximum is reached, even if the follower notices the leader, leader-directed learning does not occur.

Leader-to-Leader Relationship (Same-Level Inﬂuence): Leader’s Experiential and Directed Learning

The leader’s learning in the CFL proﬁle’s embedded elements occurs in the same fashion as followers’ learning does. Leaders also engage in learning in the CVF proﬁle. The leader learns and/or forgets just as followers do. Comparing Simulation Results 285

RUNNING THE SIMULATIONS

A simulation run consists of generating synthetic leader and follower input data as required by the hypothesis under testing and then initiating multiple iterations of inﬂuence and learning/forgetting. Because the simulation offers multiple opportunities for stochastic choice elements (e.g., opportunities for learning and forgetting, followers’ noticing the leader, leader’s choice of actions to engage in) to come into play, multiple simulation runs are necessary. Thus a virtual experiment consists of 100 repetitions, using different synthetically generated data, of 80 iterations. We conducted at least one virtual experiment for each hypothesis being tested. To facilitate interpretation and because the variance is extremely low between repetitions, the averages of these runs are represented in the output graphs given later in this chapter. Each graph is referred to as a developmental path. The top or optimal level of the CFL index is 7.

RESULTS

After running the simulation, we looked at the group developmental paths to determine whether our hypotheses were supported. The initial examination consisted of plotting the developmental paths and comparing graphs. We then compared the developmental paths for the workgroup with an initially high level CFL index (the high-CFL group) and after lowering the variable criteria for CFL from an average individual CFL index that is higher than 5 to one that is between 3 and 5. We next compared the developmental paths of this lower initial group level (the lower-CFL group). We ﬁnished by reporting on the simulation results of both the initially high-CFL and lower-CFL groups but with the inclusion of forgetting in the model.

High-CFL Group Results

Fig. 1 shows the developmental paths for each leadership style with the initially high-CFL groups. The ﬁrst hypothesis calls for all four leadership styles to have a steep developmental path, but it is evident from the graph that this is not the case. The LMX style of leadership has a very shallow path that does not approach the optimal CFL value of 7 until very late in the iterations (about 75). The other three paths are within 0.5 unit of the index value of 7 by about the tenth iteration and would qualify as being 286 JANICE A. BLACK ET AL.

Fig. 1. All Leaders with High CFL Ratings, without CFL Forgetting. steep. The next four hypotheses address the order in which the developmental paths either reach the optimal level or plateau. If we use reaching the 6.5 level of a group CFL as a point of comparison, all leadership styles are equally effective. Hypothesis 2 called for the participative leadership style to have the steepest slope and to reach the optimal level or the highest plateau in the smallest number of iterations. This was not the case. In the emergent group resource, the participative style started at the highest level but had the third steepest developmental path; it reached its plateau in third position. Hypothesis 2 was rejected. Hypothesis 3 called for the individualized leadership style to be the second steepest developmental line. In general, this was the case. The path started off in third place but was second within a couple of iterations and ended up having the second highest plateau level. Hypothesis 3 was not rejected. Hypothesis 4 noted that the LMX leadership style should have the third steepest developmental path and plateau level. However, this was not the case: The LMX group experienced a drop within the early iterations and did not plateau until approximately 75 iterations. Hypothesis 4 was rejected. Comparing Simulation Results 287

Hypothesis 5 reflected Dionne and Dionne’s findings that the autocratic leadership style demonstrated the worst results in their specific group task model. This was not the case in our emergent group resource model. This leadership style had a very steep path, as it went from the worst position initially to the best position by the eleventh or twelfth iteration. Hypothesis 5 was rejected. In general, our emergent group resource model using the leadership styles as presented by Dionne and Dionne (slightly modified to account for the social construction aspects of the emergent group resource) found very different effects than the specific group task model presented by Dionne and Dionne. Our order of leadership effectiveness, as evidenced by obtaining the highest plateau level in the shortest number of iterations, was autocratic, individualized, participative, and LMX. Note that this order is markedly different from the order of leadership effectiveness in reaching group decisions found by Dionne and Dionne. Their order of effectiveness was participative, individualized, LMX, and autocratic. Their worst leadership style could reasonably be considered the best leadership style for the emergent group resource. The only path that was relatively similar was the developmental path associated with the individualized leadership style.

Lower-CFL Group Results

In our second set of results, we present the developmental paths of the four leadership styles with the initially lower-CFL group. Fig. 2 graphs the four leadership styles and their associated developmental paths for the CFL index. In this graph, all four leadership styles have lower initial group CFL indices. The graphs do not duplicate the slopes and curvature of the developmental paths associated with the initially high-CFL groups; indeed, they do not show any similarity either to the high-CFL group’s paths or to one another’s paths. With the high-CFL group, the initial order of placement mimicked the placement order found in Dionne and Dionne’s work. By contrast, with the lower levels of CFL indices, the order of initial placement also changed. With the lower-CFL group, the initial order of increase in developmental paths was autocratic, individualized, participative, and then LMX. This order remained the same throughout the iterations, with the exception of iterations 25 through 35, when the LMX leadership style inﬂuenced higher levels of emergent group CFL; after this point, the LMX leadership style again dropped below the level for participative leadership. This placement 288 JANICE A. BLACK ET AL.

Fig. 2. All Leaders with Lower CFL Ratings, without CFL Forgetting. order makes sense given that the base leader profile (that of the Aggressive Achiever) had not changed. The Aggressive Achiever profile more closely matches that of the high-CFL group. The overall effectiveness of the various leadership styles (autocratic, individualized, participative, and then LMX) was the same for both the high-CFL group and the lower-CFL group. Thus, whenever the leader had the most influence, the developmental paths reflected those higher levels. However, the lower-CFL group’s profiles also had an impact. Specifically, the developmental paths were not as steep and took much longer to plateau than did the paths for the high-CFL group. The plateau phase occurred at about iteration 45 instead of at the earlier 12 iterations. In the high-CFL group results, the participative, individualized, and autocratic leadership styles reached within 0.5 unit of the optimal level by iteration 10. With the lower-CFL group, a greater variation existed in terms of the number of iterations each leadership style took to reach that same effectiveness level. The autocratic leadership style was still very effective and reached the 6.5 level of developmental path at iteration 10. That style resulted in a developmental path that became much shallower at that point Comparing Simulation Results 289 and ultimately approached the optimal level and basically paralleled it beginning around iteration 45. The individualized leadership style had a developmental path that showed the same general pattern as the autocratic leadership style but was much less pronounced (in terms of slope) and not as efficient. The developmental path results with the individualized leadership style did not reach 6.5 until about iteration 30. It, too, began by paralleling the optimal level at iteration 45, albeit at a slightly lower level than the autocratic leadership style’s developmental path. The participative leadership style had a developmental path that was less steep than even the individualized leadership style’s developmental path. It did not have the same pronounced bend from a steeper to a more shallow path as did the other two paths. The developmental path associated with the participative leadership style was fairly constant in slope until it bent and began paralleling the optimal line at iteration 50. It reached the 6.5 level at about iteration 45. With the LMX leadership style, a shift occurred in the in-group membership, with a resulting leap upward of the developmental path at about iteration 25. This higher level was maintained for about 10 iterations; at that point, there was a shift downward, albeit not as low as the first upward shift. These shifts were due to the black-and-white nature of either being in the in-group or being in the out-group. This developmental path continued, with this group being identified the least efficient and effective leadership style. As a consequence of these results, Hypothesis 6 was not rejected. However, the paths never matched the order found by Dionne and Dionne, so Hypothesis 7 was rejected. Interestingly, the order of effectiveness for the leadership styles remained constant within a simulation, although they were different across simulations. This finding implies that to be effectively developed or accomplished, different resources may require different leadership styles even when the same people are involved in the group.

Forgetting-Included Results

We next examined the developmental paths when forgetting was included in the simulation. Forgetting was called to be a required part of any model that either explicitly or implicitly included learning (de Holan et al., 2004; Kriegesmann et al., 2005). As Fig. 3 illustrates, the developmental paths were less steep or shallower than when forgetting was not included. 290 JANICE A. BLACK ET AL.

Fig. 3. All Leaders with High CFL Ratings, with CFL Forgetting.

Whereas in Fig. 1, the autocratic, participative, and individualized leadership styles all reached the 6.5 level at about the same time (approximately 10 iterations), in Fig. 3 these same three leadership styles took 30 iterations to cross one another; only one style approached and passed the 6.5 level – namely, the autocratic leadership style. Hypothesis 8 called for starting at the same levels but taking longer for the developmental paths to reach the optimal level. Thus Hypothesis 8 was not rejected. Hypothesis 9 called for the order of the efﬁciency of the developmental paths to remain the same as outlined in Hypotheses 2–5. As Fig. 3 again demonstrates, this was not the case. The most effective leadership style remained autocratic leadership, and the second most effective style was individualized leadership. Nevertheless, the participative and LMX leadership styles resulted in developmental paths that crossed at roughly iteration 60. It was also interesting that, with forgetting, the participative style was very shallow, peaked at about iteration 40, and then began a slow decline. Hypothesis 9 was rejected. Fig. 4 clearly demonstrates the importance of including an element of forgetting in the models. As shown in this graph, the developmental paths Comparing Simulation Results 291

Fig. 4. All Leaders with Lower CFL Ratings, with CFL Forgetting. did not ever reach the 6.5 efﬁciency level discussed earlier. Indeed, there was great variation in the developmental paths. The autocratic leader still had the most effective developmental path; it began by following a neutral path but then had a positive orientation. The individualized path was fairly neutral and slightly degraded across iterations. The participative path was gently degrading along all 80 iterations. The LMX leadership style resulted in an initially much lower developmental path than the others; this path was also very slightly downward sloping, although not as much as the developmental path associated with the participative leadership style. Hypothesis 8 called for less steep or shallower paths. That certainly happened, but there were also distinct changes in the slope patterns with the inclusion of forgetting for this lower-CFL group, with only one leadership style resulting in a developmental path that was positively sloped. Hypothesis 8 was not rejected for the lower-CFL group. The ﬁnal effectiveness order remained the same across all sets of runs for this simulation, and that order continued to be different from Dionne and Dionne’s order. Hypothesis 9 was rejected. 292 JANICE A. BLACK ET AL.

DISCUSSION

The order of the effectiveness of leadership styles remained constant across workgroups with high-CFL or slightly lower-CFL profiles. The level of effectiveness varied dramatically across the levels of CFL profiles and with the inclusion of forgetting. This finding implies that the development of an emergent resource will be sensitive to the initial levels of its components and that forgetting is, indeed, a critical component when doing social sciences modeling. Whether the order of effectiveness is appropriate only for the CFL or, more broadly, for emergent resources still needs to be investigated. In this study, the effects of leadership styles differed for a specific group task versus a group emergent resource. The lone exception was the relative effectiveness of individualized leadership style, which remained the second most effective style for either output. In the specific group task, the participative leadership style for a group of intelligent professionals was the most efficient and effective; this was not the case for the group emergent resource, however. This difference makes sense in light of the findings from earlier studies that demonstrate the importance of task in participative leader groups (Bass & Shackleton, 1979; Murnigghan & Leung, 1976). The more important and complex the task, the more participative leadership behaviors should be used. In the group emergent resource model, despite the influence of the team on individuals, the autocratic leadership style was the most efficient and effective. This held true even when a necessary component of forgetting was included in the model. Clearly, the definition of autocratic leadership used in these studies is of great importance. For the specific task, autocratic leadership was defined as complete decision-making power on the part of the leader. A sole influence on the CFL was not specified because of the socially constructed and emergent nature of the resource, but the leader was given a majority of the influence. When this style was transferred to the emergent group resource model, autocratic leadership proved to be effective. This outcome may reflect that the leader was held constant across the simulation – that is, as an effective leader with high CFL. Thus allowing the leader’s CFL to vary might potentially alter the results of the simulation. Nevertheless, the model may accurately represent the need of subordinates to look for direction and be more reliant on their leader’s expertise when faced with the co-evolution of group members’ learning, gaining expertise, and occasionally forgetting. In examining why the participative and LMX leadership styles performed so differently when considering the group emergent resource than in the specific group task, we offer several observations. First, the group-level Comparing Simulation Results 293 influence on the individual persisted throughout the use of all leadership styles. This group influence may offset, enhance, or confound the influence of the leader (Dionne, Yammarino, Howell, & Villa, 2005). In the work of Dionne and Dionne, however, the cross-level impact occurred only through the leadership style. Contrast that with the cross-level impacts from the CFL model that included the leader’s style, the group results, and the individual’s own impact on self across time. It may be that adding these influences to Dionne and Dionne’s work would affect their results. This extension of their model would have the advantage of adding and looking at some of the issues of substitutes for leadership that were brought up in a recent Leadership Quarterly article (Dionne et al., 2005). Second, the leadership definitions used – given that they were based on only the amount of influence by the leader or follower – also affect the findings. All of the leadership styles could be considered a variation of the participative leadership style (Yukl, 2002). This linkage may imply that, instead of including a category for the leadership style, the model should use a continuous variable representing the amount of leadership influence. This may be a better way to conceptualize leadership involvement across specific tasks and emergent resources. Our third observation is that we assumed the leader’s CFL profile fell into the same generally high range as did the high-CFL group. However, earlier work has shown that the relative level of the focal resource held by the leader makes a difference (Black et al., 2006). This sensitivity to initial conditions may contribute to the results obtained with our model. The CFL components also contained elements of high-involvement work teams that would most likely be beneficial in any situation in an organization – namely, trust, support, discipline, and stretch. These contextual elements were not a component of Dionne and Dionne’s model. Likewise, Dionne and Dionne’s work did not indicate whether the IQ levels used for the professional group characteristic were shared by the leader. Furthermore, when our model put the leader in charge of a group with a lower level of CFL, the sensitivity of the emergent group resource to the leader’s style and personal level of CFL became readily apparent. A fourth observation is that the specific lower level of the LMX developmental path may be attributable to the categorization process used by the leader. In this model, the categorization process was a binary one: An individual either was part of the in-group or was not. This may not be the case in Dionne and Dionne’s model. If greater leeway was given, then the abrupt decline in levels of CFL in the first few iterations might not occur. This drop indicated that those individuals who were part of the in-group 294 JANICE A. BLACK ET AL. were not those individuals with the highest levels of CFL. Perhaps a model that, instead of substituting effort for expertise, modifies the effort by the level of expertise would more accurately reflect which criteria leaders actually consider when using an LMX leadership style. Unlike the other leadership styles, the LMX style does not use all of the group members’ experience and expertise to the fullest. With all of the other leadership styles, all members have a chance of contributing to the outcome.

IMPLICATIONS

Research and Researchers

Simulation models are proving to be fertile tools for research. Simulations allow the researcher to conduct organizational research without requiring the intrusiveness of actual organizational studies. They are also less costly tools for organizational research, as the same model can be used repeatedly, with minor changes being made as needed. Simulation models also eliminate inadvertent human biases that can affect traditional organizational research because of the need to specify explicitly the relationships and influences under consideration. They also retain the ability to explicitly include biases of concern for the researcher. Even so, despite their ability to eliminate inadvertent biases, simulation models do not automatically respond to the subtle nuances that exist in the organizational environment. For example, individual perceptions greatly affect leadership effectiveness in organizational settings. Leaders may be deemed effective only if their leadership style fits the implicit leadership theories held by their followers, or their effectiveness may fluctuate when followers transfer previous experiences to the current setting (Lord & Maher, 1991; Ritter & Lord, 2007). These types of nuances remain to be included in models and their obvious and hidden system results examined. Although simulation research has certainly provided us with specifications and alternative views, many questions remain unanswered. The findings in this chapter and in the chapter by Dionne and Dionne might differ significantly if contextual influences such as culture, gender, and minority group status were added to the models. Exploration of the ability of the leader to adapt to situations might also yield key insights. A model that allowed the leader to change leadership style as the situation warranted would be interesting to examine. Such a model should allow the leader to adjust his or her style to reflect follower needs and expertise. This flexibility Comparing Simulation Results 295 may be especially important with forgetting, as a leader may need to vacillate between more directive, supportive, and participative styles as the subordinate learns and subsequently forgets skills and abilities. These two simulations allow us to see that different resources, organizational processes, and people may all affect the effectiveness of any particular leadership style. It does appear that the individualized leadership style, while consistently coming in second in the simulation results, may actually be a more overall effective style of leading and is worthy of further investigation. Further, these theoretical models need to be compared to field activities; such a comparison might be achieved through the sharing of the knowledge and models with managers and human resources professionals.

Managers and Human Resources Professionals

In general, simulations such as those described in this volume may be useful tools for human resources managers to determine the best leadership styles for training and developing employees. They may also be useful as diagnostic tools to determine the potential developmental times of groups. Even if these models are only theoretically developed, practitioners may find it informative to enter their specific conditions and examine options for improving their training and team-building effectiveness cycles. After being tested with such real-life situations, their results and the matching of deviations from the simulation results can then be used to further clarify the boundaries of the theories and concepts included in the models (McKelvey, 2002). The specification of the concepts and their operationalization required for the simulation effort enables specifics about the workplace to be examined in virtual experiments. Examination of the emergent group resource model simulations further reinforces the importance of using different leadership styles and behaviors to enhance both the leader’s effectiveness and the follower’s effectiveness. Others have noted that when an effective leadership style is used, followers’ expectations, motivation, and attitudes improve, in turn increasing performance and decreasing absenteeism and turnover (Howell & Costley, 2001). The simulation also demonstrates the organizational elements that allow managers to create environments that are conducive to learning.

High-CFL Group

A high-CFL environment requires that the leader set clear performance standards that are linked to corporate goals, equality, access to resources, 296 JANICE A. BLACK ET AL. and open communication, all of which enable subordinates to be involved in the decision-making process. Research ﬁndings emphasize that participative leadership generally enhances a leader’s effectiveness when followers are professional, have a high degree of job competence, have a desire to develop, and have autonomy on the job (Dorfman et al., 1997). According to our simulation results, when a high-level group is in this type of environment, autocratic and individualized leadership behaviors become important, as does participative leadership, especially where some forgetting is occurring. When forgetting is occurring, there is a clear need for the leader to provide direct input and make decisions on behalf of the group and to establish direct relationships with each individual group member. Participative leadership will be much less effective when subordinates are forgetting elements of their tasks and learned experiences. The leader may need to reiterate clear performance standards and provide direction based on each individual subordinate’s needs.

Low-CFL Group

In an organization that has a low CFL index, autocratic, individual, and participative leadership styles are still essential for effectiveness. However, when leading a low-CFL group where forgetting is occurring, an effective leader will focus on autocratic and individualized leadership styles and behaviors. Using a participative style at this juncture could have a negative impact on follower effectiveness. Leaders also need to focus on the reasons behind the development of the low-CFL environment and emphasize the need to create an environment where performance standards are communicated and linked to corporate goals, where employees have access to resources and are treated fairly and equitably, and where communication ﬂows between management and employees and among employees. Emphasis needs to be placed on training and development to enhance effectiveness.

CONCLUSION

These simulations address the importance of examining leadership styles and behaviors at different organizational levels and with followers at various levels of ability and expertise. Different leadership styles may result in more efficient and effective group emergent resources and specific group task developmental paths. It appears that the effectiveness of such styles Comparing Simulation Results 297 may be constant within a category (emergent resource or specific task). Just as lowering the IQ of professionals lowers the efficiency of reaching an optimal decision point in Dionne and Dionne’s simulation, so, too, does lowering the initial CFL index levels of a group. However, no matter how the expertise level is manifested, forgetting affects the efficiency of the developmental path of the emergent resource in our simulation. Further- more, because forgetting has been identified as a necessary requirement for modeling explicit or implicit learning, it may also be an important component to include in the specific group task model described by Dionne and Dionne. In both cases, hidden system effects, such as sensitivity to initial conditions, were elucidated from the theoretical composition of the models when the simulations were run. This points to the importance of using simulation-based research for many organizational topics – two of which are the specific group task modeling done by Dionne and Dionne and the group emergent resource modeling of Black and associates. The focus of the present volume is hierarchical leadership. Simulation results can clarify theoretical thinking about cross-level effects and tracing system effects inherent in such work. The need for specification found in modeling – while limiting in that it is not a complete system representation – is critical in clarifying when and where cross-level influences occur and how such influences affect the focal outcome. It is exciting to be part of this new area of research, which holds such promise in helping us understand cross-level research and in doing so in highly ethical ways without affecting the lives of people in a work environment. Granted, such isolated modeling does need to be validated through fieldwork and should not be left as the only form of research on these topics. Further, it does need to be part of a larger research cycle (McKelvey, 2002). Nevertheless, given the incremental nature of scientific investigation, it is an increasingly legitimate part of organizational research, and especially of cross-level and hierarchical-oriented research. We commend Dionne and Dionne for their work, which will serve as a crucial component of organizational research heading into the future.

REFERENCES

Anderlohr, G. (1969). What production breaks cost. Industrial Engineering, 1, 34–36. Anderson, P. (1999). Complexity theory and organization science. Organization Science, 10, 216–232. 298 JANICE A. BLACK ET AL.

Bass, B. M., & Shackleton, V. J. (1979). Industrial democracy and participative management: A case for synthesis. Academy of Management Review, 4, 393–404. Bedeian, A. G., & Hunt, J. G. (2006). Academic amnesia and vestigial assumptions of our forefathers. Leadership Quarterly, 17(2), 190–205. Black, J. A., King, J. P., & Oliver, R. L. (2005). Simulation of emergence of context-for-learning. In: R. Sanchez & A. Heene (Eds), Advances in applied business strategy: Volume 3: Resources, stakeholders, and renewal (pp. 275–304). Oxford, UK: Elsevier Pergamon. Black, J. A., Oliver, R. L., Howell, J. P., & King, J. P. (2006). A dynamic system simulation of leader and group effects on context for learning. Leadership Quarterly, 17(1), 39–56. Black, J. A., Oliver, R., & King, J. P. (2004). Effects of member, participative, balanced and autocratic leadership power relations on development of context-for-learning. Presented at Western Academy of Management, Anchorage, AK, April 2004. Carley, K. (1995). Computational and mathematical organization theory: Perspectives and directions. Computational and Mathematical Organization Theory, 1, 39–56. Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory testing in organizational behavior: The varient approach. Englewood Cliffs, NJ: Prentice-Hall. Davis, J. P., Eisenhardt, K. M., & Bingham, C. B. (2007). Developing theory through simulation methods. Academy of Management Review, 32, 480–499. de Holan, P. M., & Phillips, N. (2004). Remembrance of things past? The dynamics of organizational forgetting. Management Science, 50(11), 1603–1613. de Holan, P. M., Phillips, N., & Lawrence, T. B. (2004, Winter). Managing organizational forgetting. MIT Sloan Management Review, 45–51. Dionne, S. D., & Dionne, P. J. (2009). A levels-based leadership simulation: Insights regarding group decision optimization. In: F. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Dionne, S. D., Yammarino, F. J., Howell, J. P., & Villa, J. (2005). Substitutes for leadership, or not. Leadership Quarterly, 16(1), 169–193. Dorfman, P. W., Howell, J. P., Hibino, S., Lee, J., Tate, U., & Bautista, A. (1997). Leadership in Western and Asian countries: Commonalities and differences in effective leadership processes and substitutes across cultures. Leadership Quarterly, 8(3), 233–274. Ghoshal, S., & Bartlett, C. A. (1994). Linking organizational context and managerial action: The dimension of quality and management. Strategic Management Journal, 15, 91–112. Gittell, J. H., & Weiss, L. (2004). Coordination networks within and across organizations: A multi-level framework. Journal of Management Studies, 41(1), 127–154. Globerson, S., Nahumi, A., & Ellis, S. (1998). Rate of forgetting for motor and cognitive tasks. International Journal of Cognitive Ergonomics, 2, 181–191. Howell, J. P., & Costley, D. L. (2001). Understanding behaviors for effective leadership. Upper Saddle River, NJ: Prentice-Hall. Jaber, M. Y., & Kher, H. V. (2004). Variant versus invariant time to total forgetting: The learn– forget curve model revisited. Computers and Industrial Engineering, 46, 697–705. Jacobsen, C., & House, R. J. (2001). Dynamics of charismatic leadership: A process theory, simulation model, and tests. Leadership Quarterly, 12, 75–112. Kriegesmann, B., Kley, T., & Schwering, M. G. (2005). Creative errors and heroic failures: Capturing their innovative potential. Journal of Business Strategy, 26(3), 57–64. Lewin, R., Parker, T., & Birute, R. (1998). Complexity theory and the organization: Beyond the metaphor. Complexity, 3(4), 36–40. Comparing Simulation Results 299

Lord, R. G., & Maher, K. J. (1991). Leadership and information processing: Linking perceptions and performance. Boston, MA: Unwin Hyman. March, J. G. (1991). Exploration and exploitation in organizational learning. Organization Science, 2, 71–87. McKelvey, B. (2002). Model-centered organization science epistemology. In: J. A. C. Baum (Ed.), Blackwell companion to organizations (pp. 752–780). Oxford, UK: Blackwell. Miller, G. A. (1956, March). The magic number seven þ/2: Some limits on our capacity for processing information. Psychological Review, 81–97. Miller, K. D., Zhao, M., & Calantone, R. J. (2006). Adding interpersonal learning and tacit knowledge to March’s exploration–exploitation model. Academy of Management Journal, 49(4), 709–722. Murnigghan, J. K., & Leung, T. K. (1976). The effects of leadership involvement and the importance of task on subordinates performance. Organizational Behaviors and Human Performance, 17, 299–310. Nembhard, D. A., & Osothsilp, N. (2001). An empirical comparison of forgetting models. IEEE Transactions on Engineering Management, 48(3), 283–291. Quinn, R. E., Faerman, S. R., Thompson, M. P., & McGrath, M. R. (2003). Becoming a master manager: A competence approach (3rd ed.). Hoboken, NJ: Wiley. Ritter, B. A., & Lord, R. G. (2007). The impact of previous leaders on the evaluation of new leaders: An alternative to prototype matching. Journal of Applied Psychology, 92(6), 1683–1695. Rousseau, D. M. (1985). Issues of level in organization research: Multi-level and cross-level perspectives. Research in Organizational Behavior, 7, 1–37. Salas, E., Guthrie, J. W., Jr., Wilson-Donnelly, K. A., Priest, H. A., & Burke, C. S. (2005). Modeling team performance: The basic ingredients and research needs. In: W. B. Rouse & K. R. Boff (Eds), Organizational simulation (pp. 185–228). Hoboken, NJ: Wiley. Shafer, S. M., Nembhard, D. A., & Uzumeri, M. V. (2001). The effects of worker learning, forgetting and heterogeneity on assembly line productivity. Management Science, 47(12), 1639–1653. Sikstrom, S. (2002). Forgetting curves: Implications for connectionist models. Cognitive Psychology, 45, 95–152. Vroom, V. H., & Yetton, P. W. (1973). Leadership and decision-making. Pittsburg, PA: University of Pittsburg Press. Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. Leadership Quarterly, 16, 879–919. Yin, R. K. (2003). Case study research: Design and methods (applied social research methods). Thousand Oaks, CA: Sage Publications. Yukl, G. A. (2002). Leadership in organizations (5th ed.). Upper Saddle River, NJ: Prentice- Hall. This page intentionally left blank

MAKING IT PRACTICAL: SIMULATION, NATURALISTIC DECISION MAKING, AND COMPLEXITY IN TEAM PERFORMANCE

Jessica L. Wildman and Eduardo Salas

ABSTRACT

There has been a lack of focus on multi-level issues within leadership research. Dionne and Dionne (2009) address this gap in the research by presenting a Monte Carlo simulation examining leadership at four levels of analysis within a group decision-making context. While their work makes a strong contribution to the sciences of leadership, group decision making, and team complexity, many aspects of the research demonstrate potential for great expansion and improvement. Toward this purpose, this commentary discusses and provides suggestions regarding the topics of computer simulation in team research, group decision-making theory, and the modeling of team complexity. It is intended to stimulate continued critical thinking and more innovative, practical, and carefully designed research efforts.

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 301–319 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008013 301 302 JESSICA L. WILDMAN AND EDUARDO SALAS

INTRODUCTION

The research endeavor presented by Dionne and Dionne (2009) brings to light several critical issues in leadership research. The central point of their work is that leadership research has failed to adequately investigate level-based theory. These authors insightfully take advantage of computational modeling – a generally underused method in leadership research – to address this issue. Their Monte Carlo simulation examines four leadership levels of analysis (individual, dyadic, dyadic-within-group, and group) in a group decision- making setting that model numerous components of group performance. Their work is commendable, given that it makes a strong data-driven contribution to the current understanding of multi-level leadership while simultaneously presenting a complex and multidimensional perspective of group performance. However, there is always room for improvement in any research endeavor. As commentators, our purpose is to build and expand on the valuable ideas presented by Dionne and Dionne. To this end, we focus on three central topics gleaned from their chapter. First, we look at computer simulation beyond the Monte Carlo technique of computational modeling. The ﬁeld of simulation encompasses a broad set of research tools outside of the example provided by Dionne and Dionne, and this commentary aims to describe the many advantages simulation can provide multi-level team research. Additionally, we provide recommendations regarding how to choose an appropriate simulation method based on the purpose and nature of the research question. Second, we focus on the topic of group decision making, again taking a broad perspective by expanding the focus to include naturalistic decision making (NDM) in complex environments. Moreover, we discuss how computational modeling could be used to further NDM research. Third, we look at the towering iceberg that is team complexity and how the model presented by the authors represents only the very beginning of the possibilities for team complexity research. The discussion of all three topics is approached with the intent of generating more meaningful and practical research. We conclude our commentary with a few ﬁnal remarks.

THERE’S MORE TO SIMULATION THAN MONTE CARLO: WIDENING THE SPOTLIGHT

Dionne and Dionne utilize a specific type of computational modeling known as the Monte Carlo technique in their exploration of leadership and group Making It Practical 303 decision making. This technique applies importance sampling to a set of synthetic data in an effort to maximize the utility of a group of people making a decision. While this is a very useful method for simulating data and provides intriguing findings as used by the authors, it is not the ‘‘end all and be all’’ of simulations. Simply stated, a simulation can be defined as any virtual environment that represents or replicates a real-world task (Garris, Ahlers, & Driskell, 2002; Klabbers, 2001). This representation does not necessarily have to be an exact duplication of the referent task, but it must maintain the same essential characteristics to ensure it is useful for purposes such as research and training. A simulation does not necessarily need to use synthetic data or a set of mathematical algorithms to model human performance. Accordingly, simulation provides a broad set of research tools beyond the scope of what is presented by the authors. Simulations range from low-fidelity role-playing exercises to high-fidelity full-motion simulations, encompassing both computational modeling techniques and virtual environments. This section briefly describes the different types of simulation, along with the advantages and disadvantages of each, and then makes some suggestions regarding how to appropriately choose a simulation method.

Computational Modeling

Computational modeling refers to a ‘‘loosely interrelated set of research tools that have been developed to address questions about the function of complex systems’’ (Ilgen & Hulin, 2000, p. 4). Zachary, Campbell, Laughery, Glenn, and Cannon-Bowers (2001) define two broad categories of existing human performance modeling technology: (1) models are complete formulations that attempt to describe or predict aspects of human competence or performance and (2) modeling techniques are computational formulations that have been used to build models of human competence or performance. Modeling techniques differ from models in that they are the general-purpose mathematical tools used to develop models of individual, group, and organizational performance. Computational modeling techniques are generally very data-driven, as they use synthetic data rather than human participants. These techniques generally involve some sort of ‘‘if-then’’ statements that are computationally tested using algorithms. Computational modeling can be used to simulate virtually any system or situation, such as complete organizations, the phenomenon of faking in personality measures, and the complex interaction between teams, leadership, and decision making. For example, neural 304 JESSICA L. WILDMAN AND EDUARDO SALAS network and catastrophe modeling procedures are useful for modeling the nonlinear behavior of dynamic systems, making them particularly useful for complex, dynamic team research (Hanges, Lord, Godfrey, & Raver, 2002). Computational modeling includes these techniques as well as numerous others, such as the Monte Carlo technique, Petri Nets, and timeline analysis. For a more extensive description of several computation modeling techniques, see the work of Hanges et al. (2002). Not only can computational modeling be used to represent complex organizations and team behavior, but it can also be used to model individual team member behavior. One particular type of modeling, known as agent- based simulation, utilizes computer-programmed intelligent agents in roles that normally would be performed by other human beings (Nogueira & Raz, 2006). Simply stated, this technique creates, via computer algorithms, virtual team members that behave in the way real human are expected to behave. Agent-based simulation bridges the gap between completely synthesized data generation and simulations that use human participants, in that it provides the unique opportunity to use a combination of intelligent agents and human participants in a single task environment. The small amount of research conducted in this area has examined the utility of agent-based technology in a military team context (e.g., Heinze, Goss, & Josefsson, 2002; Lewis, Sycara, & Payne, 2003). Computational modeling has several general advantages as a research tool. Most obviously, using synthesized data eliminates the need to identify, recruit, and use human participants. This approach can make research situations that were previously impossible or difficult owing to participant issues feasible. Generally, using computational data can reduce the effort and cost of team research. However, computational modeling itself can often be complex to perform, and requires an expertise with the underlying technology, so reduction of effort may be dependent on the expertise of the user. Another advantage of computational modeling is that it serves as a research strategy that can be used to model large sets of theoretical claims simultaneously (Seitz, 2000). This characteristic makes it an ideal method for testing complex theories holistically (as illustrated by Dionne and Dionne), for testing theories in which the passage of time is necessary (Munson & Hulin, 2000), or for testing multiple levels of analysis at once, which may be extremely difficult otherwise. However, just as the data-driven nature of computational modeling makes it well suited for research situations that are too complex or difficult to study using more traditional methods, so it also creates limitations. Because every variable and factor in the model is experimentally controlled Making It Practical 305 to some extent, results obtained from computational models are difficult to generalize to real-world populations. The results of computational models apply only to field situations characterized by the exact same set of environmental factors that were included in the model. Even small changes in the computational model could potentially result in drastically different results. Dionne and Dionne insightfully acknowledge this limitation in their research, stating ‘‘different assumptions produce different results . . . Therefore, our interpretations are limited to the conditions represented by the model, which naturally restricts the usefulness of the results.’’ Not only does this low level of reliability make computational modeling more difficult to generalize, but it also makes this sort of modeling ill suited for theoretical development. Given these limitations, computational modeling is best suited for complex research questions in which a very specific situation, with strong theoretical backing, is in need of empirical support.

Virtual Environments

At the other end of the spectrum, the umbrella of simulation includes physical and representational virtual environments such as PC-based simulations, game-based simulations, synthetic task environments (STEs), and virtual reality. These simulation methods differ from computational modeling in that they are generally designed to be used by human participants in a laboratory setting, although, as mentioned previously, intelligent agents could also be used. The primary goal of these simulations is to replicate, to varying degrees, a particular task environment of interest, which then serves as a test bed for subsequent experimentation. Virtual environments range in their levels of fidelity. Hays and Singer (1989) define simulation fidelity as ‘‘the similarity between the training situation and the operational situation which is simulated’’ (p. 50). They further explain that this similarity can be measured in terms of two dimensions: (1) the physical characteristics of the simulation (i.e., the look and feel of the equipment and environment) and (2) the functional characteristics of the simulation (i.e., the functional aspects of the task and equipment). Some high-fidelity simulations, for example, can be quite realistic in both function and physical appearance, such as fully functional airplane cockpit mock-ups or virtual reality environments. High-fidelity simulations generally attempt to replicate as exactly as possible the work environment, equipment, and systems (Cooke & Shope, 2005). Such simulations are best suited for research that does not need tight 306 JESSICA L. WILDMAN AND EDUARDO SALAS experimental control, because they are highly realistic, multifaceted, and complex settings allow for more uncontrolled environmental variation. Moreover, high-fidelity simulations are often the most generalizable because they most closely replicate the transfer setting in both a physical and functional sense. Applying this idea to the work of Dionne and Dionne, using a high-fidelity simulation would have resulted in a very different study. For example, the authors could have been interested in leadership and decision making in the context of airline crews, and accordingly used a physically and functionally realistic commercial flight simulation to replicate both the task and the environment as closely as possible. Although this approach would have resulted in a specific and applicable understanding of how different levels of leadership influence the decision making of real airline crews, it would have been nearly impossible to include and control the large number of variables that were incorporated in the Monte Carlo simulation. STEs occupy a midpoint on the fidelity continuum, as they attempt to reduce the complexity of the task setting while maintaining the functional and cognitive fidelity of the task requirements. Generally, STEs are lower in physical fidelity, and they may not even directly replicate the referent task. Instead, they are designed to reproduce the core behaviors and cognitions of interest. For example, Cooke and Shope (2005) developed an STE that replicates the experience of a team piloting an unmanned aerial vehicle (UAV). This virtual environment is not a very accurate representation of the environment in which a UAV piloting team would perform or the physical make-up of the equipment, but it requires the same thoughts and behaviors used in the operational task. It also reduces the complexity of the system in an effort to reduce the amount of training necessary to perform the UAV task while retaining the behavior of research interest. Going back to the airline crew example, an STE developed to represent airline crew decision making would most likely not look exactly like the inside of a commercial airplane, and the participants might not engage in an exact replica of the piloting task. Nevertheless, the simulated task characteristics would be designed to elicit the desired behaviors and cognitions from the members of the crew. Accordingly, STEs are best suited for research in which the physical fidelity is secondary to the task goal, the task does not need to be exactly replicated to produce usable results, and the researcher desires some level of environmental control or variation. Low-fidelity simulations are the most abstract representations of the intended referent system, in both a physical sense and a functional sense. Examples of low-fidelity simulations include role-playing exercises and PC- based simulations. Making It Practical 307

The prevailing belief regarding fidelity is that higher fidelity is always associated with better outcomes, such as higher transfer of training or better research findings (e.g., Forsyth, 1990). This is not necessarily true, however. Bowers, Salas, Prince, and Brannick (1992) reviewed several studies and concluded that low-fidelity simulations can, indeed, be useful in studying team performance, and at a relatively low cost. Low-fidelity simulations have a higher level of experimental control: Although this characteristic may sacrifice the generalizability of the findings, it increases the rigor of the hypothesis tests (Bowers et al., 1992). In response to the common criticism that low-fidelity laboratory research is artificial and not generalizable, Driskell and Salas (1992) contend that laboratory settings are critical for understanding team performance in that they allow researchers to test hypotheses about the real world. In this way, theories of the real world can be developed and rigorously tested, and eventually applied in a more practical setting. Due to this characteristic, simulations with lower levels of fidelity are well suited for basic research endeavors and for initial tests of newly developed theories.

Choosing the ‘‘Right’’ Simulation Method

As the previously discussed examples clearly illustrate, different types of simulation, with different levels of fidelity, are best suited for very different purposes. Simulations of all types can be very useful for training, for experimentation, or for modeling complex behavior (Klabbers, 2001). However, all simulation techniques are not created equal, just as all research situations are not the same. In the end, one technique is not better than the other, and all simulations with all levels of fidelity hold a rightful place in research. It is most important that the particular simulation method is chosen for the level of fidelity and control necessary for the specific purpose at hand. Three key points determine whether a simulation technique will be successfully implemented: The simulation technique must be matched to its specific purpose. The simulation technique must mimic conditions of practical interest. The simulated model must be theoretically supported. Unfortunately, while the chapter by Dionne and Dionne contributes greatly to the science by providing robust data-driven support for a model of leadership and team decision making, it shows potential for improvement in each of these areas. For several reasons that will be described shortly, 308 JESSICA L. WILDMAN AND EDUARDO SALAS their study falls short of ideal in regard to the articulation of how and why the particular simulation technique was chosen. First and foremost, all simulations should be chosen carefully to match the purpose of the user, or the result may not serve the intended purpose. It is unstated, but must be assumed, that the purpose of Dionne and Dionne’s research is to better understand how leadership influences group decision making. There is no mention of a real-world organization or any training purpose, so basic research seems to be the best assumption of the authors’ purpose. Still, it is unknown why the researchers are interested in this particular combination of theoretical processes (i.e., participative vs. authoritarian leadership and linear decision making) or why they chose to pursue a data-driven computational modeling technique rather than a more generalizable simulation method. In the face of the lack of information provided, the research constructs seem to have been chosen randomly, possibly based on interest, whereas the simulation technique was chosen perhaps because it models complex systems without requiring convoluted experimental designs and large numbers of participants. While this approach can clearly produce interesting and significant findings, it is not the most systematic approach to choosing a simulation method and might potentially result in a mismatch between methodology and purpose. Not only should the simulation technology be carefully matched to the research purpose, but also, regardless of purpose and chosen simulation method, all simulations should mimic conditions of practical interest if at all possible. The ultimate goal of organizational and team research is to contribute to the successful performance of real-world organizations and teams within those organizations. If simulations do not shadow situations of interest to practitioners, then the research will not be of benefit to the community. In the case of Dionne and Dionne’s work, it would have been ideal (of course) if they had related their model of performance and findings to a real-world organizational setting in which leadership and decision making occur in the same manner as their computational model demonstrates. The authors hint at the importance of practical application when they ask, ‘‘How can practicing leaders understand the appropriate use and application of particular leader styles and behaviors?’’ Yet, their chapter does not dive any deeper, lacking descriptions such as how practicing leaders could actually benefit from this research or implement the discussed leadership styles. The authors could have furthered the practical impact of their work by providing direction for real-world team leaders in group decision-making settings and perhaps outlining some basic implementation tips for participative leadership based on their positive findings. Making It Practical 309

Additionally, it is important that the simulated conditions are developed based on strong theory. As mentioned previously, many simulation techniques – in particular, those in the realm of computational modeling – are very data-driven. This orientation results in findings that are highly unreliable and can change dramatically in response to small tweaks in the system modeled. Therefore, to ensure that the findings hold meaning across applications, the simulation must be grounded on sound theoretical bases that are carefully chosen for the specific purpose at hand. Dionne and Dionne acknowledge the limitations of simulated data, and qualify that ‘‘although . . . assumptions were selected based on existing theory and practical recommendations, they are merely a starting point for further investigations.’’ The authors describe in detail the leadership and decision-making theories used to develop their model, but do not spend much time describing why they chose these particular theories. Specifically, they choose to focus on four leadership conditions (i.e., individual, dyadic, dyadic-within-group, and participative) and introduce a theory of group decision making based on a linear utility curve. While these theories are well developed, inherently interesting, and worthy of research, the authors could augment their theoretical bases by describing practical reasons for choosing these theories over comparable alternatives such as transformational leadership or NDM. Spring-boarding off of this concept, we now explore one of the comparable theoretical options just mentioned: NDM.

DECISION MAKING ISN’T ALWAYS LINEAR: NATURALISTIC DECISION MAKING

The approach Dionne and Dionne take toward group decision making is a linear one, focusing on personal utility curves and the influences of others. This linear approach to decision making assumes that decision makers are struggling to choose between a finite set of choices, weighing the pros and cons of each choice against the others, and that there is an optimal choice to be made. Yet field experience (e.g. Klein, 1998) has shown that this process is not necessarily how real professionals make decisions. For example, imagine a team of fire fighters responding to a residential house fire where there have been reports of trapped individuals. It is unlikely that upon arriving at the scene the firefighting team will sit down and brainstorm all of the possible approaches to putting out the fire, compare 310 JESSICA L. WILDMAN AND EDUARDO SALAS the advantages and disadvantages, weigh each team member’s opinion based on his or her expertise and tenure, and then choose the corresponding optimal option. More likely, an experienced leader will immediately assess the situation upon arriving at the scene and within seconds put in motion a plan of action to extract the trapped individuals and contain the fire based on prior experience and expertise. As this example illustrates, not all decisions involve a salient cognitive struggle between choices – sometimes a decision maker is so skilled and experienced in the situation that he or she arrives at the ‘‘correct’’ choice without ever consciously considering another option. This type of decision making, known as recognition-primed decision making, builds off of Rasmussen’s (1993) levels of routine performance. After repeated practice and exposure to similar decision-making situations, the decision maker begins to recognize patterns within the situation as they occur, and performance can eventually become automatic (Rosen, Salas, Lyons, & Fiore, 2008). This complex type of decision making differs drastically from that which Dionne and Dionne describe in their research. Additionally, decision makers are not always attempting to reach the optimal decision, as in the simulation presented by the authors. Instead, they may settle for the first feasible option that satisfactorily solves the problem at hand. As in the firefighting example, optimization may not be necessary, or even desirable, in all decision-making situations. Thus it is quite apparent that in many real-world environments, group decision making is abundantly complex, and potentially quite different from the model presented by Dionne and Dionne. Decision makers with varying levels of experience often operate in dynamic, uncertain, and fast-paced environments, trying to make decisions that have real and immediate consequences for them and those around them (Zsambok & Klein, 1997). Accordingly, if researchers such as Dionne and Dionne are to apply their experimental findings to real teams in real situations, they must develop models of decision making based on how decision making actually occurs in the field (Salas, Guthrie, & Burke, 2007). This is not to say that utility-based decision making does not take place in real organizations, but rather it represents only one of many decision-making situations that could benefit from simulated research. To develop real-world models of decision making, researchers can benefit from the literature on NDM. This literature focuses on how professionals use their experience to make decisions in field settings (Rosen et al., 2008). Traditional theories of decision making often lack predictive validity in non- laboratory settings because they do not take into account the effects of the Making It Practical 311 various contextual factors that accompany decision making in the real world. The NDM movement grew out of a desire to remedy this lack of predictive validity. As a result, the theories of NDM take into account the multidimensionality of decision making in highly complex environments by including the various factors that can influence decision making in field settings. Specifically, Orasanu and Connolly (1993) delineated eight critical task and setting factors that can affect the way real-world decision-making occurs:

1. Ill-structured problems (rather than well-structured, clearly defined, laboratory-created problems) 2. Uncertain, dynamic environments (rather than static, unchanging, predictable simulated environments) 3. Shifting, ill-defined, or competing goals (rather than clear and stable predefined goals) 4. Action/feedback loops (rather than one-shot decisions that are unrelated to the decisions surrounding them) 5. Time stress (rather than unlimited time constraints) 6. High stakes (rather than laboratory situations with no real consequences) 7. Multiple players (rather than individual decision making) 8. Organizational goals and norms (rather than decision making that happens with no outside influences)

As described by Dionne and Dionne, the linear decision-making task in their study was quite far from including all of the aforementioned factors. In fact, the only factor clearly included in their model is multiple players, given that the authors were examining the effects of different levels of leadership on group decision making rather than individual decision making. Otherwise, the decision-making process was computationally simulated in a linear fashion, which by definition means it was a predictable, stable, well- structured problem with a clear optimal solution. Therefore, while these authors’ work is quite valuable and contributes to our scientific understanding of leadership in linear decision-making environments, it lacks something in the form of results that can be generalized to other decision makers – specifically, those performing in dynamic situations with contextual factors such as an ill-structured problem or uncertain end goals. The authors’ work does, however, serve as a strong point of departure for further research beyond this linear perspective of decision making, and NDM provides the next steps for this expansion. 312 JESSICA L. WILDMAN AND EDUARDO SALAS

Just as computational modeling was used to simulate linear decision making, so it could be used to model the multiple task and setting components involved in NDM. Combining our discussion of simulation techniques with our discussion of NDM, computation modeling techniques represent a potential alternative to laboratory experiments as a method for investigating complex decision making in the hopes of understanding how decision making happens in the ﬁeld. Rosen et al. (2008) posit that the core of NDM theory resides in the inclusion of two factors: the role of expertise and the role of context. Dionne and Dionne did include the role of expertise as an individual characteristic component, making their model of linear decision making impressively comprehensive and realistic. In addition to expertise, the model could have included contextual variables such as competing goals, time stress, and high-stakes consequences so as to more closely represent decision making as it happens in more complex contexts (some of these variables are discussed in the following section). In this way, computational modeling could be used to produce strong data-driven support for the theories of NDM.

MODELING TEAM COMPLEXITY: JUST THE TIP OF THE ICEBERG

As previously stated, the chapter by Dionne and Dionne makes a tremendous contribution to the science of teams by modeling team performance in a very complex and multidimensional manner. These authors’ model included an expansive number of components, such as cognitive ability, personality, team structure, expertise, interdependency, and tenure. Their simulation is a solid attempt at modeling a theory of team performance holistically by including all of the myriad aspects that are present in the real world. By including these numerous variables, Dionne and Dionne were able to draw tentative conclusions regarding the effects of the various factors on the quality of group decision making under differing leadership conditions. However, as expected, the model was nowhere near comprehensive or representative of every possible team performance situation. The complexity of team performance is so expansive that is virtually impossible to model all possible components at once. Luckily, recent emerging research has given us several interesting possibilities for new investigations, so there is no reason to give up our lofty goal of fully understanding every aspect of team performance. Making It Practical 313

The model of team performance put forth by Dionne and Dionne provides a strong basis for endless additional explorations in the world of team complexity. This section provides several (but not all of the possible) illustrative examples of components that could potentially be included when modeling team complexity.

Individual Characteristics

Dionne and Dionne include two individual characteristics in their model: cognitive ability and personality. Several other characteristics not included in the model could also potentially have a noteworthy impact on decision making. For example, Dionne and Dionne include the Big Five factor ‘‘openness’’ as their only personality construct, citing the work of Salas, Guthrie, Wilson-Donnelly, Priest, and Burke (2005) as the reasoning behind the choice of components included in their model. While Salas and colleagues do, indeed, suggest that the personality trait of openness is essential for individuals in command positions (i.e., leaders; p. 192), it is not the only personality variable they mention. The suggestions made by Salas et al. (2005) are intended to be only the first steps in modeling team performance. Recent research has demonstrated that several other personality constructs, such as conscientiousness and extraversion, are related to team performance (e.g., Bell, 2007). In the context of Dionne and Dionne’s study, it is possible that extraversion could have played a critical role in the relationship between leadership and decision making. Specifically, extraverted team members would be more likely to interact with others, including the team leader, thereby increasing their potential influence on the group decision. This variable could have even potentially interacted with other variables such as expertise and team tenure, causing those factors to have a significant impact on decision making only when the team member is extraverted and, therefore, comfortable with interpersonal interactions. While this example is completely hypothetical in nature, it illustrates the basic need for variables to be chosen (and excluded) based on existing theory and areas of interest in actual organizations. Outside of personality, several other individual-level constructs could potentially be included in the authors’ model of team performance. Specifically, Dionne and Dionne include cognitive ability, but choose for their purposes not to include any other individual characteristics, such as motivation or goal orientation. Depending on the transfer setting, these individual characteristics could play a big part in the team members’ 314 JESSICA L. WILDMAN AND EDUARDO SALAS decision-making processes. Focusing on motivation, if team members are not motivated to reach any particular decision, there is a much lower chance that they will actively participate in the decision-making process, and consequently there is a lower chance that unmotivated team members will actively attempt to influence the decisions of the team leader. Perhaps the opposite occurs as well: Team members who are highly motivated to reach a decision because it has a direct impact on something they personally value may put extra effort into influencing the group decision. Dionne and Dionne could enrich their findings by including other individual-level characteristics that influence team performance, especially if their end goal is to apply the findings to practicing leaders and teams that will undoubtedly have varying levels of motivation.

Multicultural Issues

One of the biggest up-and-coming areas of team research focuses on the effects of cultural diversity in teams. As the world’s technology improves and work becomes more complex, geographical boundaries are becoming obsolete and teams are becoming ever more prevalent in the workplace. Not only is the use of teams increasing, but collaboration across national and cultural boundaries is also becoming more prevalent with the advent of global virtual teams, multinational corporations, and the current global political climate. Research has shown that collectivism is related to team performance in field settings (Bell, 2007), and findings such as this suggest that collaboration across cultural boundaries could influence the functioning of a team. Sutton, Pierce, Burke, and Salas (2006) recently presented an extensive list of cultural variables, including collectivism that they believe could potentially affect teamwork. This new avenue of research in team performance provides a huge opportunity for modeling complex team systems. Dionne and Dionne did not include culture as a component in their model, but doing so could have resulted in very interesting and practical results. For example, assume they were interested specifically in the effects of different levels of leadership on the decision-making process of NATO teams. Would these teams, which are composed of leading figures from various North Atlantic nations, reach decisions differently than a team composed of members from just one nation? What sort of cultural differences between nations would have the most noticeable impact on decision making? These are highly relevant questions to ask given the current global issues, and modeling the decision-making Making It Practical 315 process of a multicultural team could produce highly valuable results for NATO teams as well as many other multicultural team settings. This is only one of many practically applicable issues that could potentially be explored by computationally modeling multicultural team performance.

Environmental Stressors

Dionne and Dionne include two characteristics of the environment in their model: time pressure and noise. Time pressure was represented through model iterations, with less iterations representing less time and more iterations representing longer periods of time. Noise was included in the model as random variation. While it is laudable that the authors included such environmental variables in their model of team performance, these two concepts in no way capture all of the possible environmental influences on team performance. Take the example of time pressure: As represented by Dionne and Dionne’s model, time pressure was not defined as expected. Salas et al. (2005) describe time pressure as ‘‘the time restriction required for task performance’’ (p. 204). Time pressure, in this sense, is not just the simple passage of time during a decision-making episode, but rather the existence of a deadline or external time-based pressure that actually imposes a psychological impact on the decision maker. The firefighting team example given earlier is a perfect illustration of a situation in which the decision makers are under a very salient and stressful time pressure situation. In such an environment, the demand to act quickly to save lives may have a very salient impact on the decision-making process. Decision makers may be highly aware that they have only minutes to make life-or-death decisions, and they may react to this pressure both physiologically and psychologically. This is an aspect of time pressure that the authors did not account for, and further research could reveal very interesting and practically applicable results. Time pressure represents only one possible environmental stressor that could be modeled. Ambitious researchers could also examine the effects of many other environmental factors, such as high-stakes consequences, unfamiliar terrain, or audible noise such as gunfire, to name but a few.

Team Cognition

Researchers have long examined the behaviors of teams, such as communication, coordination, and decision making, as well as the attitudes 316 JESSICA L. WILDMAN AND EDUARDO SALAS of teams, such as team cohesion, satisfaction, and team efficacy. Owing to an abundance of empirical and theoretical literature, we know much about how teams act and feel together. The study by Dionne and Dionne is a perfect example of behavioral research, as it delves into how interactions between a team leader and team members can influence overall team decision making. What we do not know as much about, however, is how teams think together. Salas and Fiore (2004) suggest team cognition – that is, the way teams think together – is another critical component of team performance. As the most recent addition to the team performance puzzle, team cognition is the next frontier in the quest to fully understand how teams operate. Without team cognition, researchers are not capturing the full complexity inherent within team functioning. The concept of team cognition intuitively goes hand-in-hand with the main tenets of Dionne and Dionne’s work, as thinking as a team is a necessary step before making a decision as a team. One of the most commonly investigated concepts within team cognition is shared mental models, described as the ‘‘sharedness’’ between team members’ mental representations of objects, actions, situations or people; the relationships between these concepts; and the way in which the team members organize the information (Bierhals, Schuster, Kohler, & Badke-Schaub, 2007; Edwards, Day, Arthur, & Bell, 2006). Including cognitive variables, such as shared mental models, as a component of the team performance model, would aid in Dionne and Dionne’s goal to accurately and completely model team complexity. The authors could further expand their model to examine how interactions between team members and team leaders influence the development of shared mental models, and how shared mental models in turn influence group decision making. Team cognition represents just one of many great opportunities for future team research to build off of the foundation provided by the Dionne and Dionne.

CONCLUSION: MAKE IT PRACTICAL

The chapter by Dionne and Dionne took leadership and group decision- making research several steps forward. Their Monte Carlo simulation combined the advantages of computational modeling with the complexity of team performance, and with the multi-level nature of leadership in group decision making. Our commentary has expanded on the strengths of their work regarding computer simulation techniques, group decision-making theory, and modeling of the complexity of team performance. It has also Making It Practical 317 pointed out the weaknesses and gaps within this research endeavor and made suggestions for new and exciting avenues to be explored by other researchers. Our hope is that other scientists interested in these topics will build on the foundation created by the authors’ findings to further our understanding and find additional ways to integrate the science into real organizational settings. Ultimately, our aim is to encourage researchers, in team studies and other areas alike, to critically examine the purpose behind their empirical explorations. Undoubtedly, the end goal of every research endeavor is to improve the quality of human life and engender real differences in performance. Every model of organizational performance – whether of leadership, decision making, or team complexity – should bring tangible benefits to real people in real organizations. Researchers should be asking questions inspired by field situations and issues of critical interest to practitioners, and then carefully designing and conducting studies with the goal of reaping the most applicable benefits possible. So we ask you, researchers, to ‘‘make it practical’’: Team research should benefit real teams.

ACKNOWLEDGMENT

This work was supported by the Office of Naval Research Collaboration and Knowledge Interoperability (CKI) Program and ONR MURI Grant #N000140610446. The views expressed in this work are those of the authors and do not necessarily reflect official Navy policy.

REFERENCES

Bell, S. T. (2007). Deep-level composition variables as predictors of team performance: A meta- analysis. Journal of Applied Psychology, 92, 595–615. Bierhals, R., Schuster, I., Kohler, P., & Badke-Schaub, P. (2007). Shared mental models: Linking team cognition and performance. CoDesign, 3, 75–94. Bowers, C. A., Salas, E., Prince, C., & Brannick, M. (1992). Games teams play: A method for investigating team coordination and performance. Behavior Research Methods, Instruments, & Computers, 24, 503–506. Cooke, N. J., & Shope, S. M. (2005). Synthetic task environments for teams: CERTT’s UAV- STE. Boca Raton, FL: CLC Press. Dionne, S. D., & Dionne, P. J. (2009). A levels-based leadership simulation: Insights regarding group decision optimization. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald Publishing Group. 318 JESSICA L. WILDMAN AND EDUARDO SALAS

Driskell, J. E., & Salas, E. (1992). Can you study real teams in contrived settings? The value of small group research to understanding teams. In: R. W. Swezey & E. Salas (Eds), Teams: Their training and performance (pp. 101–124). Norwood, NJ: Ablex. Edwards, B. D., Day, E. A., Arthur, W., & Bell, S. T. (2006). Relationships among team ability, composition, team mental models, and team performance. Journal of Applied Psychology, 91, 727–736. Forsyth, D. R. (1990). Group dynamics. Pacific Grove, CA: Brooks/Cole. Garris, R., Ahlers, R., & Driskell, J. (2002). Games, motivation, and learning: A research and practice model. Simulation & Gaming, 33, 441–467. Hanges, P. J., Lord, R. G., Godfrey, E. G., & Raver, J. L. (2002). Modeling nonlinear relationships: Neural networks and catastrophe analysis. In: S. G. Rogelberg (Ed.), Handbook of research methods in industrial and organizational psychology (pp. 431–455). Malden, MA: Blackwell. Hays, R. T., & Singer, M. J. (1989). Simulation fidelity in training system design: Bridging the gap between reality and training. New York: Springer-Verlag. Heinze, C., Goss, S., Josefsson, T., et al. (2002). Interchanging agents and humans in military simulation. Artificial Intelligence Magazine, 23, 37–47. Ilgen, D. R., & Hulin, C. L. (2000). Computational modeling of behavior in organizations: The third scientific discipline. Washington, DC: American Psychological Association. Klabbers, J. H. G. (2001). The emerging field of simulation and gaming: Meanings of a retrospect. Simulation & Gaming, 32, 471–480. Klein, G. (1998). Sources of power: How people make decisions. Cambridge, MA: MIT Press. Lewis, M., Sycara, K., & Payne, T. (July 14, 2003). Agent roles in human teams. AAMAS-03 workshop on humans and multi-agent systems, Melbourne, Australia. Munson, L. J., & Hulin, C. L. (2000). Examining the fit between empirical data and theoretical simulations. In: D. R. Ilgen & C. L. Hulin (Eds), Computational modeling of behavior in organizations: The third scientific discipline (pp. 69–83). Washington, DC: American Psychological Association. Nogueira, J. C., & Raz, T. (2006). Structure and flexibility of project teams under turbulent environments: An application of agent-based simulation. Project Management Journal, 37(2), 5–10. Orasanu, J., & Connolly, T. (1993). The reinvention of decision making. In: G. Klein, J. Orasanu, R. Calderwood & C. E. Zsambok (Eds), Decision making in action (pp. 3–20). Norwood, CT: Ablex. Rasmussen, J. (1993). Deciding and doing: Decision making in natural contexts. In: G. Klein, J. Orasallu, R. Calderwood & C. E. Zsambok (Eds), Decision making in action: Models and methods (pp. 158–171). Norwood, NJ: Ablex. Rosen, M. A., Salas, E., Lyons, R., & Fiore, S. M. (2008). Expertise and naturalistic decision making in organizations: Mechanisms of effective decision making. In: G. P. Hodgkinson & W. H. Starbuck (Eds), The Oxford handbook of organizational decision making: Psychological and management perspectives (pp. 211–230). Oxford, UK: Oxford University Press. Salas, E., & Fiore, S. M. (2004). Why team cognition? An overview. In: E. Salas & S. M. Fiore (Eds), Team cognition (pp. 3–8). Washington, DC: American Psychological Association. Salas, E., Guthrie, J., & Burke, S. (2007). Why training team decision making is not as easy as you think: Guiding principles and needs. In: M. Cook, J. Noyes & Y. Masakowski (Eds), Making It Practical 319

Decision making in complex environments (pp. 225–232). Burlington, VT: Ashgate Publishing. Salas, E., Guthrie, J. W., Wilson-Donnelly, K. A., Priest, H. A., & Burke, C. S. (2005). Modeling team performance: The basic ingredients and research needs. In: W. B. Rouse & K. R. Boff (Eds), Organizational simulation (pp. 185–228). Hoboken, NJ: Wiley. Seitz, S. T. (2000). Virtual organizations. In: D. R. Ilgen & C. L. Hulin (Eds), Computational modeling of behavior in organizations: The third scientiﬁc discipline (pp. 19–32). Washington, DC: American Psychological Association. Sutton, J. L., Pierce, L., Burke, C. S., & Salas, E. (2006). Cultural adaptability. In: E. Salas (Series Ed.), C. S. Burke, L. Pierce, & E. Salas (Vol. Eds.), Advances in human performance and cognitive engineering. Vol. 6: A prerequisite for effective performance within complex environments (pp. 143–173). Amsterdam: Elsevier. Zachary, W., Campbell, G. E., Laughery, K. R., Glenn, F., & Cannon-Bowers, J. A. (2001). The application of human modeling technology to the design, evaluation and operation of complex systems. In: E. Salas (Series Ed.), C. A. Bowers, N. Cooke, J. E. Driskell, & D. Stone (Vol. Eds.), Advances in human performance and cognitive engineering research (Vol. 1, pp. 201–250). Kidlington, Oxford, UK: Elsevier Science. Zsambok, C. E., & Klein, G. (1997). Naturalistic decision making. Mahwah, NJ: Lawrence Erlbaum Associates.

This page intentionally left blank

SINS OF OMISSION AND ENVY: REDEMPTION AND SALVATION THROUGH LEVELS OF ANALYSIS

Shelley D. Dionne and Peter J. Dionne

ABSTRACT

In critiquing our levels-based group decision simulation, Wilderman and Salas (2009) suggest that more descriptive decision models and more sophisticated simulation techniques would improve the practicality of our model. Black, Oliver, and Paris (2009) employ an agent-based model within an emergent task context to examine a leader’s influence on group context for learning and discuss differences in key findings. Although we admit to sins of omission regarding contextual decision theory, we highlight the practicality of our model and contrast this quality with the generalizability of higher-fidelity simulations. Additionally, we admit to sins of envy in that both critiques offer an exciting glimpse into the future of group decision research.

INTRODUCTION

Just as teams in organizational settings face shifting and increasing complexity, environmental uncertainty, and time constraints, organizational

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 321–333 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008014 321 322 SHELLEY D. DIONNE AND PETER J. DIONNE research follows a nearly identical trajectory. Examining old problems in new ways and attempting to map new techniques onto existing problems are two of the many strategies employed by organizational teams to solve problems. But these approaches are not limited to organizational teams: The same problem-solving strategies can be employed by research teams as well. Our levels-based leadership simulation (this volume) was designed to combine existing theories of leadership, team process, and levels of analysis, and map these concepts onto a more recent methodology emerging across various sciences – namely, that of simulation. Both commentaries on our work, by Wilderman and Salas (2009) and Black, Oliver, and Paris (2009), represent ‘‘all that is right’’ with continued exploration of existing problems in new ways. Additionally, as complexity in our environment increases to reveal problems that are barely understood, these authors also provide suggestions for mapping new techniques on new problems. In subsequent sections, we will address the questions and concerns raised by both commentaries. First and foremost, however, let us state clearly that we are grateful for their comments and time invested in improving our research. Did we commit sins of omission in developing and conducting our Monte Carlo decision-making simulation, as pointed out by both commentaries? Very likely, guilty as charged. We will discuss some of the omission concerns presented by both commentaries and hopefully atone for those sins by providing a clearer explanation of our focus. More importantly, what we are guilty of at this point in time is sins of envy, as Wilderman and Salas as well as Black and colleagues offer various suggestions and demonstrations related to what the future of team decision research has to offer. Wilderman and Salas discuss the benefits of moving away from prescriptive decision models (such as we used) and instead focusing on the more recent descriptive decision models – specifically, naturalistic decision making (NDM) – in an attempt to better address fluid environments often faced by teams today. Black and colleagues provide a demonstration of the power of agent-based modeling as related to emergent and evolutionary changes within teams. What we will attempt to do here is highlight how a more specific levels-based focus within both of these paradigms can continue to move decision-making research and applications forward. We begin by clarifying our position related to the specificity of what we did and did not test, and then follow with unifying research suggestions within a specific levels of analysis framework. Sins of Omission and Envy 323

SINS OF OMISSION

Wilderman and Salas (2009) assert that the purpose of our research was not stated and, therefore, must be to better understand how leadership affects group decision making. True, those relationships were tested in the model. Nevertheless, as stated in our earlier chapter, the purpose of our study was more specifically to conduct a levels-based comparison of leadership impacts on group decisions. We believe the levels-based comparison of leadership is fundamentally different from examining leadership impacts on group decision-making net of a levels of analysis framework. As a consequence, we devoted the opening of our chapter to highlighting the idea that most research in leadership has been conducted without a levels of analysis framework (Yammarino, Dionne, Chun, & Dansereau, 2005). Prior levels of analysis research has indicated several problems associated with ignoring a levels of analysis framework in theoretical development and methodological tests (cf., Yammarino et al., 2005). Without clear levels of analysis specification in theory development, appropriate testing of relationships can be compromised, and a potential for misalignment exists between theory and data. Despite the seriousness of this potential flaw, most leadership research ignores this danger (Yammarino et al., 2005). As such, it would seem that as a field, we conduct sins of omission on a regular basis. Perhaps the irony of the Wilderman and Salas commentary is that in critiquing why we chose the leadership and group constructs that we did, they, too, may have followed the predominant pattern of leadership research (Yammarino et al., 2005) and ignored our emphasis on levels of analysis as a theoretical focus. Without specifying the importance that a theoretical levels of analysis perspective has on leadership variables, it might seem that variables were randomly selected. In reality, our leadership selections were driven by the notion that particular levels of analysis may be specifically related and appropriately linked to various leader behaviors. Hence, because we were interested in comparing the impact of levels-specific leadership behaviors on group decision-making efficacy, we needed to model these more clean-cut, accessible levels-specific behaviors. Additionally, we believe that, while accessible, autocratic, dyadic, and participative leader behaviors and exchanges represent viable, realistic leader approaches found operating in organizations today. Again, our research question was related to how a levels-based approach to leadership might affect the ability of groups to make decisions. Because our research comprised a comparison study, we also needed to include leader actions that represented 324 SHELLEY D. DIONNE AND PETER J. DIONNE various theoretical levels of analysis to conduct the comparison. We did not consider the levels of analysis as (solely) a methodological focus; instead, we believed that some information may be gained from examining the levels- based context of leadership and its impact, if any, on group decision making. As Yammarino et al. (2005) concluded, there is strong evidence that leadership research continues to view levels of analysis as a methodological tool, rather than a tool that can direct theoretical assumptions and relationships. We attempted to embody this theoretical view, wherein levels of analysis provide a framework in which to examine how leader actions may differentially influence group performance. Rather than developing a transformational leadership theory about leading teams, our focus was on understanding whether leader behaviors linked to a particular level of analysis may or may not be better suited for managing an interactive decision process. We viewed the levels of analysis framework as a theoretical factor, not (solely) a methodological tool. Should an autocratic approach be used to dictate a decision for a group, or something more democratic? Is a focus on one-to-one relationships within each team more effective in promoting group decision making than allowing an in-group to influence the decision? These questions directed the theoretical inclusion of leadership behaviors in the simulation, as the level in which the leadership ‘‘resided’’ was considered an important variable with potential impact on group decision making. Thus, as stated earlier, our purpose was to explore the effects of various levels-based leadership behaviors on a group decision process, and those levels-based behaviors were the guiding factors for leadership selection. Regarding our use of a decision model devoid of context, Wilderman and Salas note that we offer few practical implications of our findings as a result of this sin of omission. They assert that a high-fidelity simulation often produces the most generalizable result due to close replication of the transfer setting. While we agree with this statement in principle, we also believe that generalizability and practicality are not necessarily the same concept. We agree that high-fidelity simulations such as flight simulators produce highly generalizable results to pilots in real cockpits, but these results are not easily applied to groups and individuals outside of the cockpit and, therefore, are not particularly practical. As such, we believe that the more practical (i.e., useful) model is likely the more simple model. Our findings regarding leadership styles apply across a variety of decision- making teams, including legislative bodies, design teams, health care teams, civic organizations, church organizations, and many other groups that use leaders and group members in their decision-making process. Sins of Omission and Envy 325

Wilderman and Salas’s critique of our use of a rational, mathematical decision model is understood, as more descriptive decision models represent increased complexity and higher fidelity. Increasing complexity and fidelity may result in greater generalizability, albeit most likely in settings that closely mirror the simulation. However, lessons learned from research in complex systems indicate that although the capability exists to devise extremely complex decision-making models, these models may be too complex to be useful for scientific inquiry (Miller & Page, 2007). This view is not new to the scientific field, as evidenced by various quotations from Albert Einstein and Carl Sagan regarding the need for simplicity in science, and by Wolfram’s (2002) premise that complexity may best be explained by simple notions. This sentiment has more recently been asserted within organizational behavior research by Yammarino and Dansereau (2009), who advocate using the smallest number of constructs possible along with an explicit levels of analysis focus. We need to reiterate a sentiment addressed in our earlier chapter: We view our model as a form of ‘‘pilot testing’’ the influence of specific levels of analysis-based leadership styles and behaviors on a group decision-making process. We agree with Wilderman and Salas that more sophisticated simulation models can add significantly to the generalizability of our findings. Nevertheless, in the interest of maintaining simplicity within our exploration of a dynamic process, we began with a more basic model of leader behavior and group interactions. And, to support our earlier advocacy of simplicity as a scientific approach, we believe that our simple model produced practical results that can be used by a variety of decision-making teams. Moreover, our results may mirror outcomes found in more complex decision-making models. For example, Wilderman and Salas offer the example of a firefighting team arriving at a fire scene and being directed by an experienced leader – an appropriate course given that the members of the firefighting team have little time to debate the pros and cons of decision optimization (a dimension of our decision-making model). Our research, with its lower complexity, supports that model as well, in spite of optimization considerations. At time 0, or immediacy, the best decision-making model, no matter how experienced, intelligent, and interdependent the team members, is the autocratic decision of the leader. Regardless of the decision model employed, the practical outcome remains the same: In an immediate crisis, the best decision is an autocratic one by the leader. As the crisis unfolds, and as more time passes and greater complexity emerges, we believe that the recommendation changes to favor a more participative leader, drawing on the strength of his 326 SHELLEY D. DIONNE AND PETER J. DIONNE or her team members. In all likelihood, the descriptive decision-making models may suggest a similar course of action. Could a contextual, descriptive model provide more insight into the reasons why autocratic decisions in ‘‘crisis’’ may be better? Or the reasons why at other times, team decisions are better? Possibly, although increasing complexity may again make straightforward deciphering of relationships more difficult. However, we do agree that as capability permits, these complexities can be introduced into decision-making models in the future, if they are layered in carefully as to facilitate isolation of key relationships and associations. Wilderman and Salas note that inclusion of a technique that captures the uncertainty of the environment, shifting of goals, and high stakes with real consequences would have improved our model. Moreover, they note that employing higher-fidelity simulations, such as flight simulators, would have produced a very different study. To reiterate a point made earlier, although our model was not high in fidelity, we believe that some of our results may mirror what happens on an in-flight emergency. At the initial realization of a serious and catastrophic in-flight emergency, in all likelihood an autocratic leader (i.e., pilot) makes a split-second decision (e.g., eject). By contrast, in a situation unfolding over a longer duration to reveal increasing complexity, a likely scenario may involve more and more ‘‘experts,’’ such as flight control, airplane manufacturers, and engineers, who are included in the decision process for guidance and decision-making assistance. In other words, the decision-making process becomes more participative, which is the model our results suggested worked best in longer time steps. One simply has to examine history to see this pattern emerge in certain situations. Consider, for example, the Apollo 13 mission. During this ill- fated flight, several astronauts and engineers from both NASA and various manufacturers worked together to solve carbon dioxide filtering issues, a crippling power consumption problem, and software limitations (Cass, 2005). More recently in 2005, a JetBlue airliner had its landing gear locked in a sideways position and spent hours in the air over California. The extra time permitted involvement of JetBlue’s maintenance facility, Airbus and Messier-Dowty engineers, and control tower and airport personnel from Long Beach and Los Angeles in the decision-making process (Scott, 2005; Morin, Pfeifer, & Garvey, 2005). In both cases, the longer time frame enabled an important, yet successful participative decision- making process. As suggested by Wilderman and Salas, the weakness of our model revolves around our optimization view, which is mainly outcome focused. Sins of Omission and Envy 327

By comparison, a naturalistic decision-making model can provide the ‘‘what’’ (i.e., outcome) and greater insight into the ‘‘how’’ through its examination of how participants in the evolving situation tap expertise and differing mental capabilities, feedback loops, and other aspects. It may be that more information or higher fidelity produces better generalizability, but fundamentally the decision outcomes may not be significantly different in the various situations. This notion points to the practicality of a simpler model, albeit not the generalizability per se. Increasing complexity involves increasing detail, which in turn can produce models that are not ‘‘transparent’’ enough to offer clear relationships between assumptions and outcomes (Miller & Page, 2007). Additionally, Maldonato (2007) notes there is likely no best way, as yet, to view the decision-making process. Thus, as we were conducting a preliminary investigation into the appropriateness of viewing leadership and group decision making from a theoretical levels-based perspective, we made the decision to start with a model of lower complexity. While we sacrificed generalizability, we may not have completely sacrificed the practicality or usefulness of our findings. Ultimately, we agree with Wilderman and Salas that increasingly sophisticated methodologies enable a more in-depth exploration of the complexity of a dynamic environment. Although we committed a sin of omission for leaving environmental complexity out of our simulation, future research will certainly include more sophisticated decision-making models such as those suggested by Wilderman and Salas. Moreover, these authors did note that there is likely value in conducting low-fidelity simulations to inform basic research, which may be where levels-based leadership research resides at the moment, given its scarcity in the published literature (Yammarino et al., 2005). Thus, the study of contrasts and comparisons in decision-making models and of similarities and differences in methodological techniques is just one of the many future research avenues ripe for potential exploration.

SINS OF ENVY

The work of Black et al. (2009) also represents an exciting development within decision research. Their agent-based model highlights some of the flexibility of virtual simulation techniques not addressed within our model. Most notably, Black and colleagues have produced a model that allowed for the emergence or social creation of a group-level resource, context for learning (CFL). The CFL consists of four components – discipline, stretch, 328 SHELLEY D. DIONNE AND PETER J. DIONNE trust, and support – that create both a profile and an index (i.e., averaged amount of each dimension). At the group level, the CFL index varies according to the weighting of two influence parameters, formal position and the personal expertise level of the individual (Black, King, & Oliver, 2005). By inserting a feedback loop that enables the individual to calculate new team levels for learning within each iteration, Black and colleagues have developed a model that allows for learning over time. From a levels of analysis perspective, the team and the leader influence the amount of learning via similarity assessments between the individual and either the team or the leader. This information is then used by the individual to learn, and the newer value is used in calculating new team levels in each iteration. Thus, Black and colleagues have specifically included multi-level influence within their model and produced a dynamic model that coevolves at the individual and group level. The inclusion of forgetting also increases the fidelity of the model, as prior research has indicated that forgetting is a critical aspect of learning models (de Holan, Phillips, & Lawrence, 2004; Kriegesmann, Kley, & Schwering, 2005). With their focus on an emergent task rather than a specific task, Black and colleagues found different results regarding the efficacy of leadership styles on CFL. Specifically, autocratic leadership was the most highly regarded leadership in the creation of CFL, regardless of whether the group ranked high or low on CFL, or whether forgetting of CFL occurred among leaders and/or group members. In our model, the value of an autocratic leader was limited to situations where immediacy drove decisions, but as time continued, this leadership style became ineffective. In CFL, where teams are aiming for equity and autonomy by definition (Black et al., 2009), it would seem that the least likely choice for creating or even supporting this environment would be an autocratic leader. Black and colleagues’ finding is interesting in that half of the CFL dimensions involve equity in decisions (trust) and autonomy in decision making (support), yet leaders who did not embody these characteristics the majority of time (i.e., autocratic leaders) were still more effective in developing a CFL. In their review of group decision-making research, Kerr and Tindale (2004) noted that a preference toward one’s own opinion may be a suboptimal strategy. Additionally, leadership experts rank autocratic leadership as one of the more ineffective leadership strategies in practice, particularly in a situation where groups and teams want equity, autonomy, and commitment (Jago & Vroom, 1980). While the tenuous nature of emerging characteristics may require strong leadership, over time, as the team learns and members achieve greater Sins of Omission and Envy 329 expertise (the main contributing factor for the creation of group CFL), it would seem that autocratic leadership would decrease in effectiveness. Interestingly, the increasing slopes leading to the highest plateaus indicated that this was not the case in Black et al.’s (2009) research. This unusual result may indicate that autocratic leadership as a concept is not suitable for creating CFL, as it lacks a particular social interaction necessary in CFL and, as such, cannot be modeled appropriately [see Black et al.’s reply (2009)]. Although we believe that the concept of autocratic leadership may be mismatched with the creation of CFL and, as such, the interpretation of the autocratic result should proceed with caution, there may be another avenue for explaining these results. The answer to this unique finding may lie in within-group analyses, where patterns within groups influence the group’s CFL more substantially in emergent tasks than in specific tasks.

REDEMPTION AND SALVATION THROUGH LEVELS OF ANALYSIS

The expanded view of levels of analysis offered by Dansereau, Yammarino, and colleagues (Dansereau, Alutto, & Yammarino, 1984; Dansereau, Yammarino, & Kholes, 1999) includes the theoretical feasibility of significant within-level exploration and development of constructs. In emergent and learning-based contexts, it may be that examining and/or proposing specific patterns or trends within a level of analysis has relevance for understanding how leaders and teams adapt and change. For example, a within-and- between perspective on teams may shed some light on the significant and sharp changes seen in the LMX condition within Black et al.’s simulation (2009). Within the LMX model, they used stretch scores (a CFL dimension) to drive who was considered a member of the in-group (higher values) versus the out-group (lower values), which enables specific examination of how team patterns of CFL may substantially affect a leader’s ability to create high CFLs. Patterns of CFL dimensions within the high- and low-CFL groups, beyond that of a stretch dimension, may contribute to creation of CFL in LMX leadership. Do all groups within the in-group exhibit the same pattern of CFL dimension values (i.e., homogenous), or is there likely to be a more heterogeneous pattern of CFL dimension values? Could this heterogeneity account for the unusual fluctuations in the LMX path, in that differing alignment of CFL dimensions, even within an in-group, can cause tension 330 SHELLEY D. DIONNE AND PETER J. DIONNE and initial confusion, resulting in a decreasing CFL? A within-levels examination of CFL patterns may produce interesting associations. Likewise, patterns of group CFL may differentially influence the autocratic leader’s ability. Moreover, in the model constructed by Black et al. (2009), both individuals and leaders were learning over time. As a consequence, intra-individual changes may substantially change the development of a CFL as well. Specific patterns of learning within persons may differentially contribute to the development of the four major dimensions within CFL. Perhaps, the autocratic style becomes more group oriented, moving toward consultative and participative styles as the leader learns more about how to work with people. Both Wilderman and Salas (2009) and Black et al. (2009) have offered exciting views of how group decision-making research and team development might be advanced. Both provide a sophisticated view of the value of contextual factors and unique suggestions to incorporate more dynamic contextual factors. In the spirit of advancing levels of analysis-based research in leadership and group development studies, the question is asked, where can we go from here? We strongly believe that increasing the fidelity of decision research, as suggested by Wilderman and Salas, and exemplified by the work of Black, Oliver, and Paris, is a critical step in improving our understanding of the complexity of group decision making, and of the leader’s impact on this process. However, we take the position that the next steps must involve a careful layering of context on processes that theoretically test and methodologically examine the specific impact of a levels of analysis perspective on leading decision-making groups. Currently, some group decision-making multi-level theories [or cross-level theories from a Rousseau’s perspective (1985)] suggest that a leader influences a team, or a leader influences a context, or a context influences a leader (Kerr & Tindale, 2004). Both Wilderman and Salas, and Black and colleagues, have developed and discussed several of these multi-level influences, which are not new within decision research (Kerr & Tindale, 2004), but rather provide evidence of the importance of examining between-level influences. Augmenting this traditional view of levels of analysis is a more careful examination of within-level factors and influences, which supports even greater capabilities in understanding and explaining complex associations and relationships within teams or organizations. Because there is agreement on the dynamic nature of the environment from all contributors in these chapters, within-environment exploration is destined to produce interesting and critical findings, especially in terms of how strategy relates to decision making. Sins of Omission and Envy 331

Additionally, as Black and colleagues note, leaders and team members are learning and possibly changing over time as well. Thus, the environment is not the only dynamic factor to be considered. A within-level exploration of leaders, teams, and the environment also may pinpoint more critical episodes of change, growth, influence, or failure within the decision process. For example, are there any patterns in an individual’s development or adaptation that point to more effective leadership, or more effective membership within a team? Black and colleagues have begun an interesting exploration of learning and forgetting, and their research suggests countless avenues that would enable further exploration of within-level (and between level) associations and relationships. Similar to the potential that within-person theories may have in terms of decision-making interpretation, within-group-level factors may influence a group decision process as well. Wilderman and Salas’s interest in cultural demography and group decision making is well suited to promote a better theorization regarding how within- and between-group cultural patterns may affect group decisions. Likewise, examining functional expertise within and between teams may provide interesting insights. Do teams with homogeneous functional backgrounds produce better decisions than teams with heterogeneous functional backgrounds? Do super groups (i.e., all experts) produce better decisions than mixed groups (i.e., some experts, some novices)? With specific levels-based theory, we can test the appropriateness of viewing a team’s success as being related to differences between teams or differences within teams. These types of findings have significant implications for structuring decision teams and setting expectations for timelines to success.

CONCLUSION

Although simulation was the tool with which we dynamically examined leadership’s impact within the group decision-making process, the key factor explored in our study was the role that levels of analysis played in developing the simulation. We were not merely interested in leadership and decision making. Rather, we were interested in evaluating whether levels of analysis played any role in determining the appropriateness of the leader’s style and behavior in leading teams to better decisions. Simulation was merely a means by which to examine many types of leaders and followers interacting and making decisions. Simulation is not, and likely never will be, the most important means for examining group decision environments. However, as Wilderman and Salas 332 SHELLEY D. DIONNE AND PETER J. DIONNE

(2009) and Black et al. (2009) note, we can make improvements in simulation techniques and methodology, and improve the construct development surrounding leadership and decision making. Both of these suggestions have great potential to influence the direction of group decision-making research. We agree that this view of the future is certainly exciting, but caution that a key factor in developing our understanding of multi-level phenomena is increasing the specificity of our theories to reflect the complex nature of multi-level relationships. Ultimately, we should also aim to increase the frequency with which we include this multi-level specification in group decision making and leadership research.

REFERENCES

Black, J. A., King, J. P., & Oliver, R. (2005). Simulation of emergence of context for learning. In: R. Sanchez & A. Heene (Eds), Advances in applied business strategy: Volume 3. Resources, stakeholders, and renewal. Amsterdam: Elsevier. Black, J. A., Oliver, R. L., & Paris, L. D. (2009). Comparing simulation results of leadership style impacts on emergent versus speciﬁc task outcomes and required simulation model components. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Cass, S. (2005). Apollo 13, we have a solution. IEEE Spectrum On-line, 04-01. Available at http://www.spectrum.ieee.org/apr05/2697 Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory testing in organizational behavior: The varient approach. Englewood Cliffs, NJ: Prentice-Hall. Dansereau, F., Yammarino, F. J., & Kholes, J. (1999). Multiple levels of analysis from a longitudinal perspective: Some implications for theory building. Academy of Manage- ment Review, 24, 346–357. de Holan, P. M., Phillips, N., & Lawrence, T. B. (2004). Managing organizational forgetting. MIT Sloan Management Review, Winter, 45–51. Jago, A. G., & Vroom, V. H. (1980). An evaluation of two alternatives to the Vroom/Yetton normative model. Academy of Management Journal, 23(2), 347–355. Kerr, N. L., & Tindale, R. S. (2004). Group performance and decision making. Annual Review of Psychology, 55, 623–655. Kriegesmann, B., Kley, T., & Schwering, M. G. (2005). Creative errors and heroic failures: Capturing their innovative potential. Journal of Business Strategy, 26(3), 57–64. Maldonato, M. (2007). Undecidable decisions: Rationality limits and decision-making heuristics. World Futures: Journal of General Evolution, 63(1), 28–37. Miller, J. H., & Page, S. E. (2007). Complex adaptive systems: An introduction to computational models of social life. Princeton, NJ: Princeton University Press. Morin, M., Pfeifer, S., & Garvey, M. (2005, September 22). Disabled airliner creates 3-hour drama in skies. Los Angeles Times, p. A-1. Available at http://articles.latimes.com/2005/ sep/22/local/me-jetblue22 Sins of Omission and Envy 333

Rousseau, D. M. (1985). Issues of level in organizational research: Multi-level and cross-level perspectives. Research in Organizational Behavior, 7, 1–37. Scott, J. (2005). JetBlue A320 emergency landing. Available at http://www.aerospaceweb.org/ question/planes/q0245a.shtml Wilderman, J. L., & Salas, E. (2009). Making it practical: Simulation, naturalistic decision- making, and complexity in team performance. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Wolfram, S. (2002). A new kind of science. Champaign, IL: Wolfram Media. Yammarino, F. J., & Dansereau, F. (2009). A new kind of OB. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Yammarino, F. J., Dionne, S. D., Chun, J. U., & Dansereau, F. (2005). Leadership and levels of analysis: A state-of-the-science review. Leadership Quarterly, 16, 879–919.

This page intentionally left blank

PART V ENVIROSCAPES

This page intentionally left blank

ENVIROSCAPES: A MULTI-LEVEL CONTEXTUAL APPROACH TO ORGANIZATIONAL LEADERSHIP

Richard Reeves-Ellington

ABSTRACT

Organizational studies fail to examine organizations in terms of the several environments in which they operate, both internally and externally. That is, studies tend to focus on climate, or time, or trust, or leadership. This chapter builds on academic research that discusses organizational environments in ways that show all of these environments are important for organizational understanding, especially for organizational leadership. In particular, this chapter offers a paradigm of understanding organizational leadership realities through multi-level understanding of the organizational environments of climate, knowledge, ethnos, and time. The chapter ﬁrst discusses ﬁve enviroscapes – climate, knowledge, ethos, time, and leadership. Each of these enviroscapes has two phenotypes – business and commerce. Each of these enviroscapes, with its concomitant phenotypes, is used differently at multiple levels of management and leadership by senior managers, middle managers, and entry-level managers. The scope of organizational reach, in terms of global, regional, and local levels of analysis, provides additional context for the use of

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 337–420 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008015 337 338 RICHARD REEVES-ELLINGTON

enviroscapes. After a review of the theoretical bases for each enviroscape, the chapter applies appropriate theory and models to an extended time case study of land purchase in Indonesia.

INTRODUCTION

Understanding of leadership remains elusive, with some leadership researchers questioning the importance of leadership (Antonakis, Cianciolo, & Sternberg, 2004; Day & Lord, 1988; Eden & Leviatan, 1975; Hunt & Larson, 1973), even though concepts of leaders and leadership have existed throughout history in historical, literature, anthropology, and business studies (Bass, 1990; Bligh, 2001; Campbell, 1973; Machiavelli, 1990; Reeves- Ellington, 1999a; Reeves-Ellington & Anderson, 1997a; Sandys, 1965) and more than 100 years of leadership research continues to yield insights. Perhaps the reason that many have given up on this endeavor is that, even after adopting a generally accepted principle of leadership, the nature of the influencing process – and its resultant outcomes – that occurs between a leader and followers leads to a focus on leaders rather than leadership. This line of research suggests that organizational leadership flows from leaders who are defined by their dispositional characteristics and behaviors (Antonakis et al., 2004). Arguably, the validity of James McGregor Burns’ (1978) comment that ‘‘leadership is one of the most observed and least understood phenomena on earth’’ comes from the focus of leadership being on leaders rather than on organizational social constructs of leadership. The predominant strain of leadership research (Antonakis et al., 2004) suggests the existence of several schools of thought seeking to identify leaders and the causes of individuals being identified as leaders (Lowe & Gardner, 2000). Each of these schools features substantial scholarship that appears to build off the other schools’ work (Avolio, Waldman, & Yammarino, 1991; Bass, 1990; Bass & Avolio, 1994; Dansereau, Graen, & Haga, 1975; Fiedler, 1971; Katz, Maccoby, Gurin, & Floor, 1951; Lord, Foti, & Philips, 1982; Mann, 1959; Stogdill, 1948). Each of these schools also works under the unspoken assumption that leaders drive leadership. Further, in the evolution of leadership scholarship, each school seems to reach a research crisis stage when unresolved conflicts arose between data and theory (Antonakis et al., 2004; Lowe & Gardner, 2000). Adapting Burns’ language, these crises seem to support the contention that leaders are poorly understood. Arguably, underlying unresolved and unexamined assumptions lead to a lack of understanding of leader theories and cause their periodic crises. Enviroscapes 339

These leadership studies and schools appear to revolve around three underlying tautologies that limit the evolution of organizational leadership studies: The first tautology is that organizational leadership is vested in specific individuals and is dependent upon the nature of these individuals. To some, vested leadership is warranted through the identification of traits that are necessary in a leader (Mann, 1959; Zaccaro, Foti, & Kenny, 1991). Trait-based research is now exploring the genetic underpinnings of leader identification (Arvey, Zhang, Avolio, & Kreuger, 2006). One could argue that this line of research is returning to the fore, based on the new school of leadership scholarship (Bass, 1998; Bass & Avolio, 1994; Conger & Kanungo, 1987). The second tautology is that organizational leaders require positions of power (Bass, 1990; Etzioni, 1964; French & Raven, 1968; Reeves- Ellington & Anderson, 1997a). Organizational leaders’ power can originate in organizational contexts, structures, position, and information (Brown & Lord, 2001; Conger & Kanungo, 1987; Dansereau et al., 1975; Fiedler, 1964, 1971; Lord et al., 1982; Uhl-Bien, Graen, & Scandura, 2000). The third tautology is that within hierarchical organizations, leaders are at the top of the hierarchy (Hunt, 1991, 2004). Within this framework, other leaders might exist farther down the hierarchy, but without the strategic leader at the top of the organization, lower-level leaders are ineffective. Reliance on the three tautologies provides strong mid-level theories (Bass, 1990; Yukl, 2006) but fails to satisfactorily integrate multi-level leadership issues and to contextualize organization environments in a cohesive theory or paradigm (Chemers, 2000; Hogan & Kaiser, 2005; Van Vugt, Hogan, & Kaiser, 2008). When restricting the issue to leadership of organizations, with only a secondary interest in leadership in organizations, and shifting from the nature of leadership to the context that forms and structures specific leadership needs, a revised definition of leadership is needed. Leadership research must focus on organizational needs when addressing leadership in organizations, as they are defined by the primary environmental needs of the organization. A definition is needed that accommodates the contextual and level issues in ways that avoid entrapment in the tautologies discussed previously. The following definition of leadership is based on the work of Zaccaro et al. (1991): Leadership ensures that organizational strategic intent is grounded in realities of external and internal environments, that external and organizational realities are aligned, that these realities are transformed 340 RICHARD REEVES-ELLINGTON from data into wise organizational strategies, and that these strategies guide organizational and human resources toward the strategic objectives of the organization. Leadership forms are one of several organizational environments that create structure and purpose. This definition of leadership creates the need for new approaches or synthesis for leadership understanding. Building on Hunt’s (1991) requirements for a new synthesis, this chapter offers a paradigm of understanding organizational leadership realities through multi-level understanding of the organizational environments of climate, knowledge, ethnos, and time. Guided by these categories, a comprehensive, operating paradigm for understanding complex organizational leadership is developed. Three multi- level organizational premises are assumed along the way (Mackenzie, 2004; Pawlowsky, 2001):

1. Organizations are open systems – they are interdependent with their environments. 2. Interdependence is the core of organizations. 3. Organization environments are highly contextual.

Building on the work of Mackenzie (2004) and Rescher (1996), one can postulate that multi-level analysis of organizational leadership has (1) three management levels (senior, middle, and junior, plus an informal external management level), and (2) three levels of analysis (global/national, regional, and local). Understanding the leadership levels of management and levels of analysis requires understanding the contexts of organizational enviroscapes of climate, knowledge, ethos, and time, which are postulated to be universal to all organizational types but speciﬁc to each organization. This latter speciﬁcity requires an understanding of the enviroscapes that inform both internal organizational environments and the geopolitical contexts in which these environments are examined. Five genotypes of environment (enviroscapes) (Reeves-Ellington, 1995b; Reeves-Ellington & Anderson, 1997a) are proposed to determine leadership contexts and bound organizational leaders (Mintzberg, 1979, 1983) in one leadership environment:

Climate, which has two phenotypes – business and commerce (Reeves- Ellington & Anderson, 1997a) Knowledge, which has two phenotypes – business and commerce (Reeves- Ellington & Anderson, 1997a) Values, which have two phenotypes – distrust and trust (Reeves- Ellington, 2004b) Enviroscapes 341

Time, which has two phenotypes of business and commerce (Reeves- Ellington, 2007b) Leadership, which has three phenotypes – strategic, business, and inspirational (Bennis, 2007; Reeves-Ellington, 1999a; Reeves-Ellington & Anderson, 1997a; Van Vugt, 2006)

By interacting with the structuring of enviroscape genotypes of climate, values, time, and knowledge, the three leadership phenotypes inﬂuence both their organization and surrounding environments (Reeves-Ellington & Anderson, 1997a). While all three leadership phenotypes may be found throughout an organization, each predominates at one of the three organizational levels of management: senior, middle, or junior. Senior leaders consist of personnel located in the home ofﬁce who have strategic business leadership responsibilities. Mid-level leaders are located in either regional headquarters or mid-level country organizations. Junior managers are predominately nationals of the country in which the organization is situated. Organized understanding of organizational enviroscapes requires an organized study of how each enviroscape genotype builds on the others. Organizational climate suggests value focus and limitations. The organizational time phenotype’s dominance provides support for climate and value preferences. More broadly, organizational knowledge provides the intellectual glue for congruity of climate, values, and time. Organizational leaders must provide the data, knowledge, and wisdom to develop and execute strategies that support the preferred organizational enviroscapes of climate, values, time, and knowledge (see Fig. 1). While one can develop an elegant research model and theory from an abstract study of enviroscapes, organization researchers need to address a basic question: What is the reality being studied? From a research perspective, the concept of reality begs the question of whose reality and how it is determined. Hunt (1991) suggests that three creators of organizational knowledge may exist: academicians, practitioners, and consultants. Implied in this notion is the idea that each of these ‘‘created’’ realities is likely independent of the others. Nevertheless, theory and resultant practice should treat the same constructed realities that are important to the organizational practitioner. A good research theory should overcome independent organizational realities, avoiding the creation of interdependent realities through the use of a well-constructed theory that is grounded in action research (Antonakis et al., 2004; Argyris & Schoen, 1978; Carlile & Christensen, 2005; Reason, 1994; Reeves-Ellington, 2004a, 2007a). 342 RICHARD REEVES-ELLINGTON

Fig. 1. Order of Research to Understand Enviroscapes.

Consultants, practitioners, academicians, and other identiﬁed, vested parties create a shared reality through shared narratives that are framed in a shared theoretical framework. A successful leadership paradigm must also dip deeply into a variety of research streams (Bennis, 2007; Reeves-Ellington, 1999a; Van Vugt, 2006). For an integrated sense of organizational reality, their enviroscapes of context, values, time, and knowledge must be contextualized and tested in identiﬁed geopolitical contexts that include global, national, and regional perspectives. Only in this way do organizational realities function effectively. To achieve organizational effectiveness, the enviroscape of leadership applies learning gained from the other enviroscape elements (see Table 1). The remainder of this chapter provides a general theoretical and conceptual discussion of enviroscapes, each of which has two primary thrusts: a business phenotype and a social phenotype. First the organizational cognitive dimensions are set by climate, which is then informed through the collection of data, which is in turn converted into knowledge and disseminated strategically in the form of organizational wisdom. The dissemination of knowledge and the subsequent actions (Reeves-Ellington, 1995b; Reeves-Ellington & Anderson, 1997a, 1997b) caused by it are delineated by the ethos of trust and distrust (Reeves-Ellington, 2004b). Finally, the enviroscape of time sets the pace of organizational business and social actions (Reeves-Ellington, 2007b). One can then examine leadership Enviroscapes

Table 1. Organizational Leadership Level Responsibilities.

Commerce Business

Organizational Global Regional Local Global Regional Local levels organization organization organization organization organization organization Senior Inspirational Managerial Inspirational Strategic Managerial Managerial Middle Managerial Managerial Inspirational Managerial Managerial Managerial Junior Managerial Managerial Inspirational Managerial Managerial Managerial

343 344 RICHARD REEVES-ELLINGTON types (inspirational, strategic, and managerial) and their applications within the three organizational levels. The chapter concludes with an extended case study of an investment start- up in Indonesia during the 1960s and 1970s. This case, which is presented in an ethnographic format, is intended to provide a practitioner perspective on the more general theory, and to demonstrate how the environmental phenotypes of business and commerce play out in complex settings. The ﬁrst case scenario provides a general picture of the business investment and concomitant commerce leadership interactions required for success in the case context. The second case scenario (land purchase) explores the local leadership required to implement business and social intentions and satisfy needs at the company, investment board, and local village payer level. The analysis of both scenarios contextualizes leadership in the enviroscapes of climate and knowledge in global/national, regional, and local settings.

ENVIROSCAPES

Much of the research into each of the environments discussed here is focused on the speciﬁc subject of climate (Ashkanasy, Wilderrom, & Peterson, 2000), knowledge (Dierkes, Anatal, Child, & Nonaka, 2003; Easterby-Smith & Lyles, 2005), ethos (Lewicki, McAllister, & Bies, 1998), time (Bluedorn, 2002), or leadership (Bass, 1990). The literature is largely silent on how organizations use and are formed by these environments and how each of these environments might align to the others. This chapter addresses both issues. Of speciﬁc importance to aligning each environment organizationally is the concept of a business phenotype and a commerce phenotype for each enviroscape. The business primarily addressed here is what constitutes business and what is done to achieve success. The commerce phenotype focuses on how personal interactions are done within and across organizational formal boundaries. Each phenotype provides a thread for keeping organizational environments aligned and functions in meaningful ways.

ENVIROSCAPE: CLIMATESCAPES

As Hanna (1988) has aptly observed, ‘‘Every organization is perfectly designed to get the results it gets.’’ The basis of organizational design is the operational enviroscapes faced by the organization (Reeves-Ellington & Anderson, 1997a), which are managed by leaders who work within specific Enviroscapes 345 leadership constructs. Arguably, the starting point for understanding organizational enviroscapes is climates. Within this section, the current placement of climate discussions in the realm of psychology is shifted to a positioning within a structural and organizational model. The model discussion involves the roles of business and commerce in the construction of organizational climate. This chapter assumes that climate is very different from culture (Denison, 1990; Payne, 2001), but disagrees with earlier assumptions about the basis of that difference. As Schein (2001) points out, organizational climates are immediately perceived by people through organizational structures, emotionality of employees, and the ways in which organizational actors express themselves. Certainly, no one quarrels with the notion that Apple and Microsoft operate with different climate executions (Reeves-Ellington & Anderson, 1997a). The question is whether the climate constructs actually differ from one another or whether they are simply different executions of the same climate construct. Much of the organizational psychology literature makes the former assumption, whereas the structural literature makes the latter assumption. Further, the psychological literature focuses primarily on climate constructs that are determined by individual organizational actors’ perceptions and research measures that identify individual climate perceptions and then compiles these more-specific constructs to obtain more general climate constructs (Kunda, 2007; Schneider & Reichers, 1983). A third line of research seeks to examine and build organizational models (Argyris, 1958; Reeves-Ellington & Anderson, 1997a). The first two constructs might define climate as individuals’ perceptions, cumulatively presented as measured by quantitative methods. The first construct investigates how organizations build climate conditions that are responsive to members’ needs. The driver in this construct is the collective needs of organizational members. Early research in climate related to social psychology research in the study of leadership and groups (Flieshman, 1953; Lewin, Lippit, & White, 1939). McGregor (1960) refined this strand of organizational climate through his examination of the role of managers in the creation and management of organizational climate. Within this strain of organizational climate research, the focus is on within-group aggregates of psychological climate, including which conditions these aggregates need to satisfy to qualify as measures of organizational climate (James, Choi, & Ko, 2007). James et al. (2007) discuss in detail the types of research activities and tools necessary to pursue this theoretical line of thought. In particular, this psychological strain of climate theory might claim that psychological climate and organizational climate are one and the same. 346 RICHARD REEVES-ELLINGTON

A second climate construct, steeped in statistical measurement tools, focuses on the individual members’ perceptions of organizational climates (Schneider & Bartlett, 1968). This work has been parsed into three specific approaches in ways that codify previous conceptual frameworks and definitions (James & Jones, 1974). Following James and Jones, Joyce and Slocum (1984) build on their earlier work by postulating that organizational climates are defined by similarity of perceptions. In contrast, Woodman and King (1978) question the validity of measurement instruments that claim to quantify organizational climate. Schneider (1985) concludes that the concept of climate has become accepted and less studied. However, he recognizes the need to resolve both measurement and organizational issues that remain outstanding. Climate constructs that start with individual actors, whose attitudes are aggregated and then used to define organizational climates, beg the question of the purpose for which these climates exist. A third climate construct builds on organizational models by using taxonomy, context, and structure (Argyris, 1958; Guion, 1973; Moran, 1992; Owens, 2004). This line of research broadens the discussion from strictly psychological climates to organizationally defined climates (Litwin & Stringer, 1968; Tagiuri & Litwin, 1968). Later studies (Drexler, 1977; Lindahl, 2006; Powell & Butterfield, 1978) introduced climate variations that depend on organizational structure and levels. Moran (1992) integrated the perceptual and structural approaches to climate in his postulation that organizational climate arises from the inter-subjectivity of members as they interact within organizationally constructed contexts. By comparison, this chapter defines climate as the initial building block for an organizational design that frames all other operating environments (Burton, Lauridsen, & Obel, 2004) – namely, trust/distrust, time, and knowledge. Organization climates are a highly contextual bedrock that sets limits on how risk is defined, responsibility expressed (Litwin & Stringer, 1968), and actions taken. These notions are expressed in the two climate phenotypes of business and commerce.

Climatescape Phenotypes: Business and Commerce

The present approach to understanding climate relies on the work of the more structural theorists but would carry this argument a step further. All modern organizations operate within two climate phenotypes: business and commerce. The business climate defines the parameters of how the organization does its business, whether it be a for-profit or not-for-profit Enviroscapes 347

Table 2. Climatescape Phenotypes of Business and Commerce.

Business Commerce

Focused Complicated and compromised Amoral Morally charged Task completion – product Ongoing social interaction – process Efficiency in achieving stated outcomes Relationship effectiveness Well-defined boundaries Diffuse boundaries venture, tends to practice risk avoidance, and centralizes authority, The commerce phenotype focuses on individual and group interactions, requires higher risk taking, and works in a defused authority environment. All organizations are assumed to require both parameters, but dominance of one is expected to influence the patterns of the other (Krackhardt & Kilduff, 1990) (see Table 2). The discussion starts with business and provides a summary of the key concepts of organizational business in terms of its focus on profit and loss, brand building, and market domination and problem resolution. The purpose of organizational business is to exclude activities and actors that do not further the organization’s strategic intents. The discussion of organizational commerce focuses on organizational actors practicing inclusion of constituencies that are marginal yet important to their organization. By working within the existing possibilities and opportunities, they operate within a milieu of sharing and mediation of events that will help the business in ways that are not immediately obvious.

Climatescape Phenotype: Business

The mainstream institutional position in business is the classical liberal view, which holds that business should be pursued as a purely economic activity. Adam Smith’s (2007) enlightenment view, which sees economic activity as a pure product of competition among individuals, persists even today in popular business myths. The linked Weberian themes of capitalism, Protestantism, and individualism live on as pervasive ideological motifs for modern organizations. Within this construct, economic gain is, or should be, the first and overriding consideration of business. This view, in turn, has led to popular myths related to the amorality of business and discussions of its moral neutrality in an enlightened pursuit of progress (Chesher, 1992). Achieving the purest rational management to maximize overall business return means studiously diminishing the commerce phenotype by avoiding 348 RICHARD REEVES-ELLINGTON its more ambiguous and compromising considerations. The dominance of the business phenotype among organizations was solidified in the mass- production era of industrialism, which largely focused on producers and consumers, rather than customers and suppliers (Reeves-Ellington, 1995b). This phenotype relies on an I–It paradigm to avoid social entanglements: We (the producers) provide goods for them (the consumers), whom we do not know personally except through scientific study. The producers furnish goods or services to people chiefly for the company to thrive. The exchange relationship is – to borrow a term from Sherry, McGrath, and Levy (1995) – a monadic one in which there is no external partner: The gift is given to the self. Thus external customers exist to serve the interests and success of the producers and purveyors. Joseph Campbell describes this modern type of organization: ‘‘an economic–political origination. Their ideas are . . . of the secular . . . in hard and unremitting competition for material supremacy and resources’’ (1973, p. 401). The most efficient producers tend to drive the social out as being irrational and inefficient. This Western logic of business can be seen in Calvin Coolidge’s famous remark, ‘‘The business of America is business.’’ Peter Drucker again articulated this notion in the post–World War II era: ‘‘Business performance comes first – it is the aim of the enterprise and the reason for its existence’’ (1955, p. 16). Levitt stipulated it more forcefully: ‘‘Altruism, self-denial, charity, and similar values are vital in certain walks of life. But for the most part, those virtues are alien to economics. Further, the governing rule in industry should be that something is good only if it pays. Otherwise, it is alien and not permitted’’ (1958, p. 48). Levitt looks on businesspeople who are concerned with social responsibility as dangerous. Socially oriented activity contaminates the aim and diverts attention from profit making and stock enhancement, which are seen as business’s ultimate and purest purpose. This separate ethic of work and its effects have been thoroughly institutionalized under business organizations and professions, and is generally viewed positively (Wiley, 1998). Intrusions from the wider society have become highly politicized issues. People who work in a dissociated manner from their moral community and its traditions experience fears, anxieties (Heinze, 1991), isolation, and powerlessness, and they have limited access to information and energy beyond themselves (Klimo, 1988). There has been a lack of commerce support for business-driven organizations as organizational actors, with the wider public viewing all organizations, large or small, as occupying a profane moral status. Enviroscapes 349

Some have asked if this narrow approach is sufficient for modern organizations (Kanungo & Mendonca, 1996; Ohmann, 1989). More specifically, Ohmann suggests that a new religion might be in order. In reality, it is not a new religion that is needed, but rather an organizational climate phenotype that increases awareness and the ability of organizational actors to integrate the social with the economic. Such a phenotype shifts attention away from a solely business focus on the pursuit of rational efficiency to a more socially laden conception of commerce that is grounded in human interaction (Reeves-Ellington, 1995b).

Climatescape Phenotype: Commerce

Compared to purer conceptions of business, the idea of commerce is older – it predates the rise of Western capitalism. It is bundled with additional meanings because it had been merged institutionally and embedded in earlier social forms and contexts of exchange. Commerce’s meanings still linger in its contemporary connotations. It can never be morally neutral because of the active presence of persons who are continually responding with interpretations and decisions. More than just economic exchange takes place in commerce. This phenotype is closer to the ground, involving not only goods, currencies, and instruments, but also contextual and tangential information and all kinds of social contact. Commerce encompasses social traffic in things valued and cannot long exist without an I and a Thou (Reeves-Ellington & Anderson, 1997a). When responsible business interactions occur in practice, most often they involve commerce actors who are situated marginally to the centers of authority in either business or state institutions. These marginal actors operate closer to the other institutional centers of the immediate social and physical environments. In exchanges between organized systems of different size, complexity, and resources, the circumstances are asymmetric, and socially responsible business is always organizationally compromised. What responsibility can be exercised rests on the cultural relationships that the local parties can improvise with one another. What is general about specific exchange relations? Commerce actors occupy themselves with opportunities, rather than problems. They interact continuously and pragmatically in the present, always with an eye toward maximizing future choices and benefits. They necessarily mediate and interpret between systemic levels, and they often must resist, avoid, work 350 RICHARD REEVES-ELLINGTON out, or reinterpret mandates and limitations imposed by governing and outlying systems. This occurs especially when they perceive local relationships to be at risk as a result of poor fit with the larger entities’ demands. If they cannot avoid imposed constraints, successful commerce actors creatively reinterpret and avoid overt dissonance with their larger systems. The early years of Apple offer a case study in commerce. Umberto Eco captures the initial Apple spirit by metaphorically describing the company as ‘‘Catholic, with ‘sumptuous icons’ and the promise of offering everybody the chance to reach the Kingdom of Heaven by following a series of easy steps’’ (Marshall, 1997, p. 145). During the start-up period of Apple, Steve Jobs brought needed technical personnel into the company. It was these people who helped Apple obtain adequate financing (Garsten, 1994) and improved marketing. All of this was accomplished by the end of 1977 (Halliday, 1983). The company was run within a technical business climate until 1981, when the company suffered two failures: the Apple III and Lisa computers. It was the Macintosh, launched in 1984 by an Apple with a less certain future, that revolutionized the personal computer market. Both the machine and its introduction were unique and were conducted primarily within a commerce climate. The print ads suggested that the Macintosh was the computer ‘‘for the rest of us’’ (Scott, 1991). The ‘‘rest of us’’ were inclusively self-defined by themselves, with Jobs as visionary leading the way. The artifacts, resonating with the American counterculture of the 1970s, playfully identified this contingent as outlaws and revolutionaries. The Macintosh building in Cupertino, California, flew the pirate flag as an apt symbol for the entire operation. During an interview with Terry Gross on National Public Radio (February 23, 1996, WSKG), Jobs clearly expressed his advocacy of the idea of the wider social integration of business in his comments about Apple during his tenure there as CEO. Jobs claimed that the business operated on the premise that Apple’s duty was to provide customer services that people wanted, not that Apple wanted. According to Jobs, people at Apple did not work based on a hierarchy of knowledge but rather by recognizing equality of knowledge. Apple hired people or invited people with knowledge into the company so that those employees could tell the organization what to do. Jobs’ responsibility was to create an environment in which people could work and feel good about themselves. In other words, Jobs tried to integrate Apple into the lives of the company’s customers and employees. As Jobs learned, running or working in an emerging, integrative company has both organizational risks and personal risks. When people live in a Enviroscapes 351 business-paradigm organizational environment and attempt to create a commerce organizational phenotype that contrasts with the business one, the venture entails major risks, both organizationally and personally.

Climatescape Summary

All organizations have two interactive climates: business and commerce. The business phenotype focuses on getting tasks done, completing those tasks with efﬁciency, and staying within the well-deﬁned organizational boundaries. These activities are amoral in the sense that business is amoral in what it wants to accomplish. At the same time, commerce is complicated and oftentimes compromised by various organizational and related actors. Actions are morally charged, are ongoing interactions that demand effective human relationships.

ENVIROSCAPE: KNOWLEDGESCAPE

Knowledge provides the organization with sense making (Bontis & Choo, 1998; Weick, 1995) for both business and commerce climates (Berger & Luckmann, 1991; Elkjaer, 2005). All organizational knowledge is learning and context dependent for both climates (Gardner, 1999). Each climate has differing needs that, in turn, require different methods to obtain useful knowledge that will deliver desired outcomes (Gioia & Pitre, 1990; Senge, 1990). The inquiry into organizational knowledge is intended to resolve both organizational business issues and those commerce issues that involve both internal actors and the wider community. This section discusses how knowledge is created for both business and commerce phenotypes through the collection of information that is transformed into knowledge as required by the organization. This section also explains how organizational information and knowledge provide a basis for members to predict others’ desires and expectations (Choo, 2002) for business and commerce behavior. Finally, it demonstrates that all successful intra- and inter-organizational knowledge and information exchanges involve learning from a mixed set of enviroscapes that translates into newly synthesized knowledge sets. These sets focus on commonalties across the organization and its wider external environments. In short, such knowledge must become integrated, harmonized, and facilitative for its 352 RICHARD REEVES-ELLINGTON organizational members over time (Reeves-Ellington, 1995b; Reeves- Ellington & Anderson, 1997a, 1997b). In the pursuit of organizational performance, an organization’s capability to improve its core business activities becomes critically important. Business improvement works in tandem with improved social and human activities. A further requirement is a sharing of business and commerce knowledge both inter- and intra-organizationally (Oinas-Kukkonen, 2008). Business knowledge instructs the organization on achieving efﬁciency, and commerce knowledge informs the manner in which business outcomes are accomplished effectively (Sawy, Eriksson, Raven, & Carlsson, 2001). Hence the need for two knowledgescape phenotypes: business and commerce. The knowledgescape challenge is to support worker business knowledge in ways that provide stable learning environments, while expanding commerce knowledge results in worker social literacy (Gardner, 1999). The overall learning paradigm for understanding and creating both explicit and tacit knowledge is cooperative inquiry (Reason, 1994). For our purposes here, Reason’s cooperative research paradigm reﬂects the information collection, knowledge creation, and knowledge use functions. These functions are the fundamentals of a knowledge environment. The premise is that open communication will enable people to work together, view, select and structure information, and create wisdom for knowledge use for either the business or commerce climate (Argyris & Schoen, 1978, 1996). Adapting Reason’s (1994) four major methodological steps for doing cooperative research, four phases for knowledge creation and use are proposed:

Phase 1. Organizational actors agree on an area for inquiry and identify some initial data needs and research propositions. They try out particular skills or seek to change some aspect of their world. They also agree to some set of procedures by which they will observe and record their own and one another’s experiences. Phase 2. Lower-level organizational actors apply these ideas and procedures in their everyday life and work, both inside and outside the formal organization. Phase 3. The data collectors become fully immersed in this activity and experience.

Phase 4. After an appropriate period in Phases 2 and 3, data are forwarded to mid-level organizational actors who structure it into needed strategic knowledge, which is then forwarded upward to senior Enviroscapes 353

organizational actors, who use it to create a strategic context, which then serves as the basis of strategic intent. Fig. 2 outlines a process for organizational information creation and use that is valid for business and commerce phenotypes. Using this process, an organization has the capacity to develop organizational knowledge that is distributed both throughout the organization and to external agents with whom the organization interacts. The ultimate goal may be either business–economic aims as stipulated by the organization’s strategic intents or those aims’ execution, which requires

Fig. 2. Knowledge Creation Processes for Business and Commerce. 354 RICHARD REEVES-ELLINGTON detailed social and behavioral knowledge on the part of both internal and external organizational actors (Carroll, Rudolph, & Hatakenaka, 2005). Each of the stages requires differing leadership types that are located in three organizational levels. Each knowledge phenotype has speciﬁc needs that are met by different levels of organizational actors.

Knowledgescape Phenotype: Business

Business knowledge starts through learning. Organizational members collect data, which they then transform into knowledge. The resultant shared and agreed knowledge leads to specific business actions and outcomes. Conversely, commerce knowledge comes from individuals’ shared learning that either reinforces existing social realities or constructs new ones that are operational either within the organization or reaching out to its surrounding communities. Business knowledge use is the result of rational patterns of action designed to support organizational business strategic intent (Hamal & Prahalad, 1989) and executions of that intent (Hedberg & Wolff, 2005; Mintzberg, Raisinghani, & Theoret, 1976; Vera & Crossan, 2005). Fiol and Lyles (1985) note the acceptance of the relationship between organization learning and meaningful strategic performance. They see this relationship as consisting of patterns of actions that are goal directed and based on data collection and knowledge creation. The purpose of these actions is to inform organizational actors about necessary business directions of the organization (Stopford, 2001; Zack, 1999). The creation of organizational business knowledge and its subsequent use are bounded by extant strategies, but provide inputs designed for specific business outcomes that offer a competitive advantage to the organization (Nonaka, Toyama, & Byosiere, 2001). Alignment of strategic intent and resultant strategies assures that shared organizational meanings articulate a shared business agenda and defines issues that are important in meeting organizational business outcomes (Choo, 2002; Reeves-Ellington & Anderson, 1997a). Within the framework of constructed business meaning, expressed in strategic terms, organizations exploit their specializations or develop new capabilities so as to move toward their visions, missions, and desired outcomes. Movement that is blocked by gaps in knowledge requires the pursuit of additional data, which in turn lead to the creation of new knowledge that expands meaning. The resultant knowledge causes strategic revisions. Fig. 3 depicts the flow of business knowledge creation and the use of that knowledge to enhance the success of the organization. Enviroscapes 355

Fig. 3. Business Knowledge Process.

Most business data originate outside formal organizations, from customer, supplier, and retail sources. Necessary data are routinely collected by junior- level managers, who then organize the data into information formats commonly used by the organization. Middle-level managers then convert the information into meaningful knowledge necessary for the formation of strategic intents, which senior members of an organization develop and circulate. Formal statements of strategic intent encompass the resulting knowledge use that has been reviewed and agreed to at each lower managerial level. The global senior leaders control and dictate strategic intents. Middle and junior levels at the global level, as well as senior, middle, and junior levels at the regional and local levels, are responsible for strategic plans and actions based on senior leader intents.

Knowledgescape Phenotype: Commerce

Whereas business knowledge tends to be explicit if it is to be useful, commerce knowledge tends to be tacit and embedded in local contexts that are not shared as common knowledge throughout an organization 356 RICHARD REEVES-ELLINGTON

(Henderson & Cockburn, 1994). Whereas business knowledge uses deﬁned procedures to move data up the organization, commerce knowledge processes information through individual actions and reactions. The resultant exchanges of information create shared ‘‘truths’’ through the interacting actors’ actions – including truths that may be unacceptable to others. Commerce knowledge is rooted in human interactions and emotion, so it is necessarily heavily value laden and uses a process of bringing together tacit knowledge through shared experiences and joint activities (Nonaka et al., 2001). Senior organizational leaders cannot fully manage commerce knowledge, because it moves randomly throughout the organizational hierarchy and across organizational boundaries. Emphasizing shared values as well as expectations in commerce knowledge sharing and creation permits members to function socially both across and outside the organization (Ting-Toomey, 1999). Commerce knowledge is required for internal socialization and for outreach to external communities so as to create shared social realities that beneﬁt all parties.

Internal Commerce Knowledge Dorothy Lee describes a subjective experience of the importance of internal commerce knowledge and its use to gain social alignment: ‘‘The clarity of the structure within which I ﬁnd myself – that is, the ‘social constraint’ . . . makes it possible for me to act . . . when I live in dialogue with this structure ....Itmakes it possible for me to proceed in what would otherwise be a confusing jungle: It makes it possible for me to function’’ (1963, p. 101). Commerce knowledge creates understanding of one’s own constraints and those of others in ways that permit a new reality to emerge such that all parties can work together with mutual trust. Sackman calls social knowledge the employees’ ‘‘directory knowledge,’’ which she says is ‘‘knowledge about chains of events and about their cause and effect relationships . . . [and] delineates the ‘how’ of things and events’’ (1992, p. 82). Ultimately, commerce knowledge serves to integrate, orchestrate, and apply acquired social knowledge in business-enhancing activities throughout and beyond the organization’s formal boundaries. Krackhardt and Kilduff (1990) observe that much of social (commerce) knowledge is emergent from within organizational members’ own experience and networks. To the extent that commerce knowledge is outside the purview of business knowledge creation and strategic intents, it leads to discrepant attributions about the same people and events. Nevertheless, through common commerce knowledge, within an accepting organizational context, a shared truth emerges. Enviroscapes 357

The implication of these interpretations is interesting: An organization’s commerce climate boils down empirically to its commerce knowledge. This knowledge is affected both by the organization workforce’s processes – as enacted in commerce networks of people who interpret it similarly – and by the historical influence of the leaders’ values. Even so, the basis for these influences is not necessarily a conscious awareness of organizational actors. Internal commerce knowledge creates a strong ethnocentric organization (Bennett, 1993). The learning pattern involves a single loop and occurs in multiple dyadic relationships that are within internal networks (Goshal & Bartlett, 1990). Fig. 4 illustrates the desired outcome of a strong exchange of commerce shared knowledge that results in a powerful organizational commerce climate for each of the three levels of analysis (global, regional, and national). The complex dyadic or small-group interactions at each organizational level (depicted with solid arrows) assure the sharing of internal commerce knowledge, which results in a strong ethnocentric organization being shared by integrating all newcomers into the organization from the moment of their entry and regardless of level of entry. More informal communication lines (dotted arrows) provide constant reaffirmation of commerce climates throughout the organization.

Fig. 4. Internal Commerce Knowledge Flows. 358 RICHARD REEVES-ELLINGTON

A Procter and Gamble example illustrates the point. Procter and Gamble employees have traditionally worked to assure that commerce knowledge dictates social actions. This knowledge is tightly interwoven into the psyche of each new employee (Schiller, 1997). When successful, employees become ‘‘Proctoids’’ (Swasy, 1993), with the social rhythms being hidden through internalization while the business timelines are overt and consciously managed. Procter and Gamble’s socialization and business activity enculturation activities continue throughout an employee’s tenure at the company and reﬂect organizational use of both knowledgescapes (Schisgall, 1981). To assure systematic composition of the enculturation process, company principles clearly identify the underlying values and ethics of the company. These principles, which are initiated at all levels of the organization, are agreed to in ﬁnal form after what might be years of discussion. Ownership of these principles is almost universal, and deviance by any employee results in punishment, either formal or social. The principles are widely posted within and outside the Procter and Gamble organization.

External Commerce Knowledge External commerce knowledge is necessary when individuals from two or more diverse communities ﬁnd themselves in a mutually dependent, transactional relationship (Taylor & Osland, 2005). This form of knowledge sharing is highly interactive and important to understand and apply (Ting- Toomey, 1999). The knowledge learning shifts from an ethnocentric mode to a relational-centric one (Reeves-Ellington, 1995b; Reeves-Ellington & Anderson, 1997a, 1997b). Rather than its commerce climate and knowledge contexts being linked solely to external entities, new, shared realities are created by internal and external players. Sharing external commerce knowledge usually requires organizational actors to negotiate in frameworks where external actors are in a dominant position – a position with which bureaucratic organizations are unfamiliar. The dominant knowledge base is that of the external actors and, as such, remains outside the organization, requiring organizational actors to absorb the enviroscape, rather than the other way around. The desired outcome is a synthetic ‘‘imagined community’’ (Anderson, 1983; Calhoun, 1991). Fig. 5 shows the knowledge ﬂows necessary for external commerce knowledge sharing. The primary information sharing channels (depicted as heavy double arrows) occur mainly at junior organizational levels but also within middle organizational levels. Importantly, the internal actors must have native knowledge of the external community environments, if the communications are to be successful. These actors must also have the Enviroscapes 359

Fig. 5. External Commerce Knowledge Flows. capability to translate the external environments in ways that are under- standable to the non-nationals working within the organization. The translated information (lighter upward arrows) flows to the non-nationals, who must have the interest and patience to internalize the information received. Only with this absorption of the information do subsequent appropriate actions occur. The senior actors of the national organization (dashed downward arrows) then confirm the organizational understanding of the external environments and the approval of actions necessary to create important imagined communities. These actions remain in the national operations and are passed forward only as anecdotal information of interest to regional and global levels. If regional or global actors must be involved, the commerce knowledge must be translated into business knowledge. A Proctor and Gamble example of doing business in China illustrates this process. During the mid-1980s, the company had an agreement with the central government to manufacture and sell Pert shampoo throughout China; the two parties had agreements for distribution to all the major cities as well. However, the global business strategy for the brand was to sample drop at more than 50% of all households on a national basis. The company 360 RICHARD REEVES-ELLINGTON stood firm on this requirement, despite local prostrations about a lack of national advertising, the lack of a national sales force, and regionally controlled markets. Two Procter and Gamble employees, who had worked in China for many years, eventually came up with a solution that resolved the conflict. Using national employees, they worked out a plan for a national launch that met home office requirements. Every major apartment building and housing complex had a central government informant to report on citizen activities. Negotiations were conducted centrally, given that the deal was under the control of the central government, with regional training to follow. The basic premise was that these informants were underpaid and had substantial free time, so Procter and Gamble offered to pay them to sample drop in the areas of their control. A deal was made, and the result was a national Pert launch, on a single day throughout China. The brand immediately went to a leadership position. The knowledgescape phenotypes of business and commerce provide necessary organizational knowledge, strategic intent and organizational efficiencies, and wider social understandings that enhance organizational effectiveness. Underlying both climate and knowledge enviroscapes is an ethoscape with the phenotypes of trust and distrust.

Summary of Knowledgescape

Business knowledge strives to reduce organizational uncertainty, minimize risk, work toward strategic intent, and perform efficiently. The process focuses on standardizing product and process in stable contexts. Business phenotypic knowledge is explicit and formally processed from the bottom up, with strategic intent flowing from the top down. Business knowledge creation is the process in which an organization collects data, consolidates those data into information, and uses the subsequent knowledge to reach rational outcomes. The commerce knowledge phenotype requires a different knowledge paradigm. Commerce knowledge embraces uncertainty, works within contexts of complexity, thrives on variation, and recognizes voices that reflect community needs. Actors using commerce knowledge combine flexibility with the ability to troubleshoot, find problems, and solve those problems. This effort requires a frequent shift in organizational roles (dynamic context). The modern organization must, therefore, have a membership that can utilize more than one learning mode, is open to new data that create both standard and creative knowledge, and works across static and flexible Enviroscapes 361 knowledge categories (Langer, 1989; Langer & Piper, 1987). Extended organizational information systems require elements of both business and commerce phenotypes that provide knowledge for all enviroscapes at all organizational levels, in global, regional, and local contexts.

ENVIROSCAPE: ETHOSCAPE

Ethoscapes theory provides organizational researchers and practitioners with a theoretical framework that permits holistic integration of multiple levels of analysis (House, Rousseau, & Thomas-Hunt, 1995; Klein, Dansereau, & Hall, 1995; Klein & Kozlowski, 2000). An ethoscape reflects the distinguishing moral natures, or guiding beliefs of organizational actors and their interactions with organizational climates and accumulation of knowledge. Specifically, this is how a person or group expects to relate to its institution’s business and commerce environments. Ethoscapes operate in the usual two phenotypes of business and commerce, but each of these has an additional phenotype: business – distrust; commerce – trust. The business and commerce operational phenotypes, in conjunction with their attributes of distrust and trust, provide a general theoretical understanding how and where ethoscapes provide ethical underpinnings for the other enviroscapes. Conceptualizing trust and distrust, researchers should incorporate multi-level analysis to include dyadic and individual elements for understanding the ethos of commerce and business (see Table 3). Trust and distrust phenotypes are implicitly situated in existing research (Reeves-Ellington, 1995b, 1997), using variations of value orientation models, societal/organizational mutual value congruence, and identification (Doney, Cannon, & Mullen, 1998; Lane, 2000). Like their controlling phenotypes of commerce and business, trust and distrust are not conceptually two ends of a continuum, nor are they reciprocal; rather, they are mutually exclusive in their pure forms. Building on the work of Lewicki et al. (1998), trust and distrust can be seen as different constructs, and each construct relates primarily to either business or commerce if it has appropriate organizational application. Using the same logic as Herzberg (1968) to determine the independence of rewards and satisfaction, one can argue that they are independent paradigms primarily concerned with differing business and commerce constructs. Each can be discussed independently of the other, using differing vocabularies and driven by nonrelated outcomes. Trust and distrust reflect basic measures of how people value themselves in the climates of business and commerce. 362 RICHARD REEVES-ELLINGTON

Table 3. Multi-Level Trust/Distrust Constructs.

Level of Analysis Trust/Commerce Distrust/Business

Societal/organizational Value congruence Social intelligence Mutual Identiﬁcation Inter-organizational Closed social circuits Value congruence Dependency Generalized distrust Dyadic Risk taking Structured Vulnerability Manipulative Cooperation Opportunistic Personalized trust Controlling Social/economic exchange Personalizing distrust Information sharing Inﬂicting punishment Secretive Individual Self-worth Ego driven Trustworthy Untrustworthy

The two phenotypes are (1) asymmetrical, with each constructed from specific value needs; (2) genetically altruistic or selfish; and (3) dominant in specific contexts (Reeves-Ellington, 2004b).

Ethoscape Phenotype: Trust

Trust Definitions Trust is a psychological state (Rousseau, Sitkin, Burt, & Camerer, 1998) containing the intention to accept vulnerability based upon positive expectation of the intentions or behavior of another (Bhattacharya, Devinney, & Pillutla, 1996; Das & Teng, 1998; Hagen & Choe, 1998; Lewicki et al., 1998; Rousseau et al., 1998). Rousseau et al. (1998) also argue that trust is the willingness to be vulnerable under conditions of risk and interdependence. These authors add that trust is not a behavior, but rather a psychological condition that causes actions of risk taking. Working from a basis of trust as a relationship, Hardin (1998, 2001) develops what he calls the encapsulated-interest conception of trust. He defines trust as a three-part relation: A trusts B to do X (totally grounded). This dyadic, self-focused conception concentrates on specific actions. By defining trust as a relationship between actors or groups in which one party Enviroscapes 363 adopts the position, expressed either verbally or behaviorally, that the other party will pursue a course of action that is considered preferable to alternative courses of action, Braithwaite and Levi (1998) argue for shared values as a foundation for trust. Value congruence and mutual identification require shared commerce knowledge if they are to create institutional and societal trust. Trust, then, is a necessary condition for inter- and intra- organizational commerce to occur.

Trust Relationships To understand trust as expressed at the individual, dyadic, and societal levels, a vocabulary is required (Gambetta, 1988; Luhmann, 1979). The basic vocabulary of trust (see Table 3) relies on deeper concepts for each of the three levels of analysis: social knowledge for commerce relationships, information sharing for dyadic relationships, and self-esteem for the individual. The vocabulary of trust captures the notions of expected performance, competence, and commitment and influences the ways in which coordination occurs. Simon Blackburn (1998) has described the resultant actions engendered by the vocabulary of trust: (1) one person trusts another to do something (trusts to); (2) one person trusts institutions, governments, or ‘‘things,’’ and the institutions trust their members – individually and collectively (work in trust states); and (3) people and institutions in their interactive environments have a general trust (work from trust). Trust, then, covers a range of positive emotions and actions that generate a general sense of individual trustworthiness and a generalized sense of societal trust. Trust requires a mutuality of experience and intimacy (Lewicki et al., 1998). More specifically, Gambetta (1988) sees trust as enabling cooperative behavior, Miles and Snow (1992) promote the notion of adaptive organizational forms, and Meyerson, Weick, and Kramer (1996) would reduce harmful conflict, thereby decreasing transaction costs and creating ad hoc workgroups. Tyler (2001) views trust as an assessment that the motives of an authority are benevolent and caring. This general benevolence facilitates voluntary deference to rules and authority, which in turn creates a general sense of trustworthiness. Trustworthiness is under- pinned by four general elements:

Interdependence (Dasgupta, 1988) A way to cope with risk and uncertainty in interpersonal and exchange relationships (Lane, 2000) 364 RICHARD REEVES-ELLINGTON

Vulnerability resulting from the acceptance of risk (Lane, 2000) Expectation that vulnerability and risk will have positive outcomes (Bhattacharya et al., 1996; Das & Teng, 1998; Lewicki et al., 1998; Hagen & Choe, 1998; Rousseau et al., 1998).

In other words, actors working within powerful trust relationships are socially other-ends directed (Hardin, 2001; Mansbridge, 1999). In the commerce sphere, trust is driven by either empathy or sympathy.

Ethoscape Phenotype: Distrust

Distrust Definitions Barber (1983) works from an assumption that distrust appears in the absence of trust. He defines distrust as the lack of capable and responsible behavior from specific individuals. Govier (1997) states that distrust rests on the expectation that others will not act in one’s best interests, instead engaging in potentially injurious behavior to others. Lewicki et al. (1998) echo a similar thought by defining distrust in terms of confident negative expectations regarding another’s conduct. By ‘‘confident negative expectations,’’ these authors mean ‘‘a fear of, a propensity to attribute sinister intentions to, and a desire to buffer oneself from the effects of another’s conduct.’’ Giddens (1990) defines distrust as dread or angst in individuals caused by a strong randomness of events. Common to all of these definitions is a focus on the individual as a distruster of other people’s actions, with underlying negative expectations. Each of these definitions assumes that a continuum exists in which trust and distrust serve as the two extremes, rather than viewing them as distinct social realities. All of these definitions also omit the possibility of a general cultural or normative value on distrust. None of the definitions covers the wide range of distrusts discussed previously. Implicit in these distrust definitions is the definition of untrustworthiness as being an unpredictable and random response that acts in ways negative to others. Another assumption is the idea that both trust and distrust belong in commerce climate and knowledge, where trust is positive and distrust is negative. However, if one separates the two concepts by having trust drive commerce and by having distrust be the default for business climate and knowledge, these definitions are of little use. Unfortunately, little research exists in relating distrust to the business of organizations. Enviroscapes 365

Distrust Relationships Many researchers still assume that distrust occurs only when society lacks cohesion. Gabarro (1978) and Lindskold (1978) argue that distrust arises when an individual or group is perceived as not sharing key cultural values, which clarifies Lewicki et al.’s (1998) position that trust and distrust are alternative social realities. Implicitly, this position assumes that both trust and distrust are naturally occurring and separable (Giddens, 1990; Lewicki & Bunker, 1995; Lewicki et al., 1998; Reeves-Ellington, 1997). Certainly, within the discourse of commerce, distrust assumes negative attributes of isolation of experience, social and psychological distance, and a skeptical or negative attitude. In contrast, within a business climate, these very attributes are key positives for structuring information in ways that create strategic intents that work within a business climate. When considered in a climatescape framework, a depersonalized distrust that is girded by business orientations and bounded by them is highly valuable (Reeves-Ellington, 1995b, 1997, 1999b). The basic vocabulary of distrust (see Table 3) relies on a generalized skeptical climate that creates a generalized distrust, which is in turn kept in check by a rigid structure and a strong punitive ethic. Sober and Wilson (1999) argue that genetic selfishness is expressed as a focus on self-interest. Selfishness, however, is a key consideration if business strategies are to work efficiently (Reeves-Ellington, 2004b). The resulting opportunistic behavior resulting from selfishness is based on the assumption that all groups or people are out for themselves. This behavior is expressed through a high regard for cunning and work in environments that avoid risks and responsibilities. ‘‘Let the facts speak for themselves’’ is a key concept. Power asymmetries are the basis of knowledge asymmetries and hierarchical organizational distance. In short, totally distrusting contexts are pathological because of the malevolent intent of people within the system. Yamagishi (2001) and Hardin (1998) speculate that people in societies that are generally distrustful are less likely to enter into risky social interactions and more likely to interact in relatively closed circles, even at the risk of giving up new opportunities. Although people who have not made an investment in cognitive resources for improving social intelligence tend to be distrusters and expect others to be distrustful, within business climates and knowledge creation, distrustfulness of ‘‘facts,’’ motivations of others, and strategic plans is assumed to be a key to success. Distrust breeds distrust, and individuals in low social intelligence societies tend to refuse to 366 RICHARD REEVES-ELLINGTON make conscious efforts to develop the social intelligence required for detecting and taking risks when appropriate. Conversely, individuals with high business intelligence perform well in obtaining narrow results but fail in commerce interactions. Two consequences follow: A lack of social intelligence makes people more distrustful when dealing with others and more reliant on business distrust. Realizing their vulnerability, people avoid engaging in such interactions and try to survive on cunning. Within the business climate, closed informational circuits are environments in which untrustworthiness is the default expectation. This is a learned defense strategy for people who cannot protect themselves properly in risky situations. If distrust is assumed in organizations, business relationships dominate at the expense of human ones, asymmetrical power drives those business relationships forward, and strong distrust environments ensue (Bacharach & Gambetta, 2001).

Summary of Ethoscapes

The ethoscape phenotypes of business and commerce are expressed through the concepts of distrust and trust. Trust and distrust frameworks provide organizational value orientations necessary to operate both business and commerce aspects of organizational activities. Business distrust requires value congruence founded on dependency, closed knowledge circuits, and some social isolation that leads to a dispassionate stance toward business climate and knowledge. Further, personal interactions assume manipulative, secretive, and controlling behavior that is opportunistic and requires the presence of a rigid structure to avoid risk and penalties. Finally, within the business environment’s multi- level formulations of individuals, each person self-selects for similarly directed peopled who are selﬁsh, ego driven, and untrusting. In trust environments, each person has a predisposition to self-select for other similarly oriented people who have shared commerce knowledge and climate. Trust at the inter-organizational level assumes shared commerce knowledge and social knowledge based on value congruence, which in turn creates mutual identiﬁcation and shared commerce environments and commerce knowledge. Inter-organizational trust relies on dyadic, unstructured relationships that are driven by personalized trust (Reeves-Ellington, 2004b). Enviroscapes 367

ENVIROSCAPE: TIMESCAPE

The fourth enviroscape, time, provides a basis of speed and pace of business and commerce organizational actions (Reeves-Ellington, 2007b). Timescapes are inherent in all organizations (Kluckhohn & Strodtbeck, 1961; Reeves- Ellington, 1995b) and are acted out in the organizational climate by key stakeholders. In other words, timescapes are embedded in organizational systems in ways that affect both systems and people (Ekvall, 1987; Joyce & Slocum, 1984; Koyes & Decotiis, 1991). Organizational climate, knowledge, and ethos, as related to timescapes and related activities, constitute a part of the shared history, expectations, unwritten rules, and social mores that affect the behavior of everyone in an organization (Denison, 1990). This section discusses the timescape and its business and commerce phenotypes in an effort to provide clearer understanding of both. It then puts key attributes of both phenotypes into an organizational model that responds to organizational needs and uses of time. The model and time parameters are ﬁnally considered as they apply to business and commerce phenotypes of the organizational timescape. Systematic study of organizational timescapes requires a model. The proposed model shown in Table 4 (Reeves-Ellington, 2007b) reﬂects the business and social timescape phenotypes, with their six dimensions of type, movement, use, action, management, and work habits. Each of these dimensions has a bipolar set that informs the understanding of the time dimensions of business and social timescapes. Understanding the importance of timescapes requires a detailed examination of organizational time. Bluedorn and Denhardt (1988) postulate two timescapes (fungible and epochal time), set boundaries for

Table 4. Timescape Dimensions and Parameters.

Dimensions Parameters

Business Commerce timescapes timescapes

Type Quantitative Qualitative Movement Linear Cyclical Use Monochronic Polychronic Actions Counterfactual Climate Management Asymmetric power Symmetric power Work habits Amoral Moral 368 RICHARD REEVES-ELLINGTON organizational tensions, and provide utilitarian dimensions (Adam, 1995, 1998; Bluedorn, 2002). Fungible and epochal time are complex constructs that exist in all organizations; they cover the ways that organizations experience time (Flanherty, 1999; Kern, 1983). The interplay of the tensions determines how organizations approach their activities and relationships in creating and sustaining successful commerce. The two timescapes are inescapably intertwined, such that both are required for an organization to function. Based on field discussions with organizational practitioners these time constructs are labeled as the business timescape (fungible), which references the chartered outcomes of the organization (what they do), and the social timescape (epochal), which celebrates human activities and interchanges (how they do what they do) (Reeves-Ellington, 2007b). Using specific organizational questions, organizational timescapes are operationalized through the application of the six time dimensions (type, movement, use, actions, management, and work habits) to practitioner actions. These questions and dimensions provide a framework for mapping business and commerce phenotypes of timescapes (Reeves-Ellington, 2007b). Understanding the complexity of time and its business and social manifestations allows organizational actors to use time more efficiently and effectively across organizations and their boundaries. While the timescapes and their six dimensions (each of which has bipolar parameters) are holistic, organizational actors initially study time serially by relating one organizational issue to one dimension. Applying the wrong time dimension to an issue results in erroneous information. Specific organizational questions relate to specific dimensions, which in turn require dimensions that foster further definition of and bring greater clarity to the understanding of organizational time. Timescape theory provides a richness of concepts with which organizational time can be studied and understood. Although timescapes, their dimensions, and dimension parameters exist in complex organizations, they force researchers and practitioners to make choices about the appropriate application of each timescape, the preference for a specific timescape dimension, and the parameter weighting that is most apt. While theory and conceptual frameworks inform the researcher and practitioner, they must be contextualized using other research methods and analytical tools. The next section considers such contextualization. In time theory, dimensions are often expressed as one of three pairs that can be identified as polar opposites: linear–cyclical (Fine, 1998), qualitative– quantitative, and monochromic–polychronic (Bluedorn, 2002; Clark, 2000; Hall & Hall, 1987). Each of the bipolar indicators provides parameters for Enviroscapes 369 the type, movement, and use dimensions of the two timescapes, though parameters for actions, management, and work remain missing. To rectify these omissions, three additional pairings are proposed: counterfactual– climate, asymmetric power–symmetric power, and amoral–moral parameters. Each of the dimensions offers insight into organizational time understanding and is used at different levels of the organization and organizational enculturation (Reeves-Ellington, 2007b). As shown in Table 4, while each timescape shares the same dimensions, the parameter for each is different. Bluedorn and Denhardt (1988) warn that these pairs of parameters should not be treated as dual entities, but rather as parameters to timescape dimensions. Organizational actors must overcome a sense of dualism by acknowledging that the two timescapes, their dimensions, and their dimensional parameters are interwoven at every organizational level and arise from the same source – a way of perceiving organizations as envisioned by the tasks they do and the manner in which they do them. A discussion of each time dimension and its social and business parameters follows.

Timescape Phenotype: Business

Bluedorn and Denhardt (1988) expand on Clark’s (2000) notion of clock time, stating that ‘‘clock time’’ is amenable to exact measurement, moves forward linearly, and flows evenly. Business time, by contrast, is dubbed ‘‘private time.’’ Urgency and speed drive business time rhythms (Waller, Conte, Gibson, & Carpener, 2000). Certainly, business time is the driving force behind organizations’ quests to meet their chartered reasons for existing – that is, the business objectives for which they are formed. To operationalize business time, one can postulate that business time (1) is even, abstract, and moves linearly; (2) attracts monadic individuals who prefer to do tasks serially; and (3) uses the past counterfactually to change the future rapidly through the use of asymmetric power (Reeves-Ellington, 2007b). Business time is discrete, fixed, and universally measured in standard units. Quantitative time occurs in a business timescape and refers to clock time that is amenable to precise measurement via discrete division, is subject to a unitary interpretation, moves forward linearly, and flows evenly (Adam, 1998; Bluedorn & Denhardt, 1988). The notion of clock time dominates the Western sense of time (Flanherty, 1999; Gell, 1992; Kern, 1983; McKay, 1968). Efficiency, as a measure of the speed of activities, becomes the 370 RICHARD REEVES-ELLINGTON objective of clock time use, with the ultimate goal being to increase the value of the firm. Organizational business activities flow from the past into the future, as measured by periodic monetary measures. Newton’s view of linear time is a major feature of the Western cultural worldview. His mechanistic view of time portrays it as an absolute physical reality and infers that the passage of time occurs independently of consciousness. In the linear view, time flows like a conveyor belt that moves horizontally from past to present to future at the same unchangeable speed for all of us (Hall & Hall, 1987). Linear time is conceived of as an endless process, without beginning yet created by a succession of events in which earlier events imply later ones and later ones are consequences of earlier ones. Linear time has the ethos of unilateral determinism. Business timescape is grounded in linear time (Adam, 1995)– the past, present, or future – and each temporal location requires different temporal orientations and different abilities to project into the future through the use of the past and present. Kosselleck (1985) suggests that temporal categories of space of existence and horizon of expectation might be used to frame the past–present–future continuum. The space of existence allows for the integration of the past into the present, and the horizon of the future reveals the way of thinking about the future. The space of existence constructs itself on a selectively remembered past. Organizational business activities, which move from the past through the present and are forecasted into the future, occur serially and quasi- independently one from the other. Monochronic time focuses on a single task, is controlling, is imposed on activities, has a short-term orientation (Hall & Hall, 1987), and takes place in personal space. People who prefer monochronic activities tend to work on tight deadlines, are serially task oriented, work monadically, and avoid team activities. Expressions such as ‘‘First things first,’’ ‘‘I’ll decide on priorities,’’ and ‘‘One thing at a time’’ express the urge to work monochronically and monadically.

Timescape Phenotype: Commerce

Organizational commerce time (also known as sociotemporal time), as suggested by Fraser (1987), rests upon commerce climate, knowledge, and trust constructions. Organizations socialize their members by establishing reciprocal, relational networks of people within specific organizations (Major, 2000). Such socialization takes place in constructed events and through integrative processes by exposing newcomers and insiders to Enviroscapes 371 company traditions and climates. Such indoctrination conforms to Clark’s (2000) description of recurrent action patterns as sequences of actions that are repeated, distributed by communication and authority, spread among several actors, interlocked by role sets, operated on tacit and unarticulated knowledge, and created by emergent routines. Specifically, social time is event time that is concrete, moves cyclically, and attracts integrators who execute a variety of tasks simultaneously and who rely on climate and policy to maintain the present status quo through the use of diffuse, symmetric power. Social time is ‘‘lived’’ time (Minkowski, 1970), occurs in the present, and is public time in the organization (Zerubavel, 1981). That is to say, time is in the events (Bluedorn, 2002) rather than the events being in time. The operationalization of this concept is expressed in organizational rhythm through deliberate actions, paced by human interactions, and bounded by traditions (DiMaggio & Powell, 1983; Leblebici, Salancik, Copay, & King, 1991). Social time is event driven and subjectively applied by actors as they interact with one another. Qualitative (event) time, as a measure of social timescapes, can follow different indeterminate event trajectories, can be subject to multiple interpretations (plurality of meaning), can flow discontinuously, and can be neither measured nor manipulated easily (Bluedorn & Denhardt, 1988). Organizing occurs in qualitative time by the building of consensually validated grammars for process and action, thereby creating organizational effectiveness (Clark, 2000). Human social activities rely on celebratory, reoccurring, and seasonal events, all of which are driven by cyclical time. Cyclical time springs from earth and life cycles (Hall & Hall, 1987). It is identified as variable and diurnal (Adam, 1995), seen through a climate prism (Reeves-Ellington, 2007b), expressed through collective memory (Cohen & Bacdayan, 1994), and perceived as recurrent and periodic (Clark, 2000). Cyclical time is important only in social timescapes, as this ‘‘lived time’’ promotes stability and coherence though processes, such as social rites of passage, and discontinuous events, such as shifting homes or job changes. It has high emotional content. There are two ways of using cyclical time: nostalgia (Boym, 2001) and climate expressed as tradition (Reeves-Ellington, 1997, 1999b). Organizational social activities often require multitasking. Most people prefer monochronic time usage when dealing with business activities, but will accept polychronic time usage when interacting with people (Reeves- Ellington, 2007b). Polychronic time focuses on multitasking (undertaking two or more tasks simultaneously), outcomes, and the future. People who prefer polychronic work tend to change plans easily, develop more 372 RICHARD REEVES-ELLINGTON human relationships, have strong organizational commitment (Bluedorn, Kaufman, & Lane, 1992; Hall & Hall, 1987), and use public space as an extension of themselves. Such individuals prefer working integratively. Within organizations – and especially within managerial ranks – monochronic behavior is encouraged early in one’s career, but a plethora of needs quickly causes managers to work in polychronic ways (Bluedorn, 2002). Organizational commerce interactions are most successful when symmetric power is assumed by all. Machiavelli provides the two significant foundations for these power sets – asymmetric and symmetric power. Machiavelli’s concerns relate to power’s imprecise, contingent, strategic nature (Clegg, 1992). Effectual or distributed power has multiple-power centers with distributed agency (Reeves-Ellington, 2004b).

Timescape Summary

The reader should take away three points from this section. First, commerce timescapes and their attendant dimensions and parameters are learned through dyadic relationships and socially distributed in complex organizations. Through social distribution, learners of social time applications, who are often not even consciously aware of such applications, have more difﬁculty in studying socially driven organizational activities. Second, business timescapes and their attendant dimensions and parameters are individually learned and practiced. Business timescape applications demand monadic work practices, serial tasks, and counterfactual reasoning to change the present course of events in ways that cause future changes. They require asymmetric power applications to task assignments. Third, both business and social timescapes are found at entry, middle, and senior managerial levels, but each of these levels has primary responsibility or organizational leadership for only certain dimensions. When dimensional leadership is blurred or shifts occur among the three managerial levels, organizational confusion may ensue.

ENVIROSCAPE: LEADERSCAPE

Leaders are required at all organizational levels and locations. Leadership needs vary according with the precise location, however. In addition, as demonstrated for the other four enviroscapes, leaderscapes must work within both business and commerce phenotypes. This section provides Enviroscapes 373 models in which leadership constructs provide the basis of understanding both business and commerce leader requirements. Each construct includes a set of leadership descriptors, communication metaphors, and tasks. Under- standing organizational leadership requirements is necessary to understand when and how to make the other four enviroscapes work efﬁciently (business) and effectively (commerce). Successful organizations require a mixture of both leadership activities, which need to occur at the right organizational level and the right geographic location. The two types of leaders operating in each of the phenotypes provide the emotional intelligence to govern the other enviroscapes (Nye, 2008). Two organizational paradigms that differ greatly in their approach to social responsibility, yet are commonly used, are the producer/consumer model and the customer/supplier model (Reeves-Ellington, 1995b). Within the producer/consumer model (a ‘‘pure’’ business paradigm), ‘‘we’’ (the producer) provide goods for ‘‘you’’ (the consumer). ‘‘We’’ produce goods or services for ‘‘our’’ company to thrive. Perhaps not surprisingly, this paradigm requires business leaders. By contrast, within the customer/supplier model (commerce paradigm), a constant interchange of information leads to socially embedded business activities. This paradigm requires commerce leaders. See Table 5 for a detailed comparison of business and inspirational leadership. Each of the two phenotypes of business and commerce has separate communications and language metaphors: discourse for business, and dialogue for commerce. As Weedon states,

If language is the site where meaningful experience is constituted, language also determines how we perceive possibilities of change. Further if experience is based in cultural contexts of cultural logic, then ethics and cultural artifacts, language is at least partially shaped by these. (1987, pp. 86–87)

Table 5. Business and Commerce Leadership Descriptors.

Business Commerce

Senior Middle/junior level Senior/mid-level

Thought Dichotomous Dichotomous Holistic Role Teacher Boss Learner Relations Follower Subordinate Disciple Activities Transformer Consolidator Creator Methods Aligner Controller Liberator Focus Organization Task Community 374 RICHARD REEVES-ELLINGTON

Table 6. Communications Metaphors.

Business Discourse Commerce Dialogue

Reasoning Calculating Rapture Process Excluding Integrating Methods Controlling Liberating Results Villains Heroes

Complementing leadership descriptors are the metaphors of business and commerce language used by organizational leaders. The language metaphor discussion includes how and where leaders use discourse and dialogue. Table 6 provides the basis of understanding of the underlying communications metaphors.

Leaderscape Phenotype: Business

Within the business phenotypes, senior and mid-level leaders are the backbone of business activities, actively managing business activities. Webster’s New Universal Unabridged Dictionary (1983) states that ‘‘to manage’’ (in the original) is to ‘‘train a horse in his pace, make docile in order to control.’’ A manager is ‘‘one who conducts affairs with economy and frugality – a good economist.’’ Thus the major traditional skill set required for managers is controlling and directing others in matters of business. The definition of management adds another dimension to managers’ activities: ‘‘management is concerned with goals and objectives; getting things done’’ (Webster’s New Universal Unabridged Dictionary, 1983). To the skills of controlling and directing, a manager must add efficient task management as the underlying driver. Wesson agrees when he says, ‘‘Management denotes the process of formulating and executing business strategies that are aligned with senior management strategic intents. These strategies are the result of functional activities of planning, organizing, directing, coordinating, and controlling’’ (1964, p. 403). The authority resides in the formal authority of the office (Nye, 2008). The traditional business literature makes little distinction between leadership and management. When such a distinction is made, it is not maintained. For example, according to Kanungo and Mendonca, ‘‘Leaders are expected to provide direction, exercise control, and generally execute such functions that are necessary to achieve the organization’s objectives’’ (1996, p. 2). Right-sizing, flattening, hollowing out, and use of devices such Enviroscapes 375 as market-focused divisions (Fatsis, 1992; Xerox Corporation, 1993) reflect the readiness of American management and leadership to develop and implement visions that shed and change people and organizational culture as necessary to increase penetration and control (Robbins, 1983).

Business Leaderships’ Managerial Metaphors Global, regional, and national business leaders all use their vision as the basic ideation from which to create a sense of the organization’s structure and boundaries (Bass, 1985; House, Spangler, & Woycke, 1991). That is, these leaders focus on setting group goals and moving the group toward these goals by providing meaning to organizational ideologies and values. They strive to achieve change by establishing their view of order within the organization, such that the organization is protected from external influences (Thomas & Humphrey, 1994). The measurements of success for business leaders are universally accepted. These leaders believe that ‘‘the most commonly used measure of leader effectiveness is the extent to which the leader’s group or organization performs its task successfully and attains its goals . . . . Objective measures of performance or goal attainment are available such as profit growth, profit margin, sales increases, market share, sales relative to targeted sales, return on investment, productivity, cost per unit output, costs in relation to budgeted expenditures, and so on.’’ Sinha would measure a leader’s effectiveness as depending ‘‘on the extent to which the demands are met and the subordinates are motivated to realize shared goals’’ (Yukl, 2006, p. 6).

Business Leaderships’ Language Metaphors The language of business leaders is controlled by centralized discourse. Such discourse might be defined as the act of telling or communicating the results of a study or thought process in a logical and organized manner. Business language is the language of management. Managerial discourse sets parameters to maintain focus, direction, and outcomes of business activities. Managerial language is used to perpetuate static situations, not to create change. The architects of business language – business leaders – structure it for efficient discourse to remove ambiguity concerning profits, sales, and other key business measures. As one of my managers once said, ‘‘There are only three reasons for communicating in this company: to inform management, to recommend to management, or to ask management for help on an issue of interest to them. Communications should contain only business facts, and each communication should start with a position of why the communication will drive business forward.’’ Kanungo and Mendonca 376 RICHARD REEVES-ELLINGTON support this point of view: ‘‘When morality intrudes into the using organization, it has the potential of directing business leaders from the organization’s primary objectives, and as a result, causing it to be inefficient and to deprive stockholders of their due returns’’ (1996, p. 2). Business leaders’ visionary language focuses on aligning organizational members with the strategic directions of the organization. Sinha (1995) rightly identifies business visionary language as that which touches the very core of the followers’ thoughts, feelings, and aspirations. This discourse controlled by the leader, however, and is driven down the organization by managerial language and logic. Business leadership in today’s enacted business environments uses language to reinforce and rationalize information flows into the organization and to filter out that which is extraneous to business knowledge creation. The discourse of business language and business action frees managers and leaders from what they see as the drag of interdependency and responsibility.

Business Leaders People working in business phenotype environments tend to focus on efﬁciencies, given that they spend their time in amoral contexts independent of their traditional social moral ones (see Table 2). The resultant disassociation from commerce increases organizational actors’ fears and anxieties (Heinze, 1991). As Klimo has stated eloquently, ‘‘the fruits of dissociation are endless degrees of unconsciousness and ignorance, isolation and powerlessness, as kinds of limitation in the ability to access information and energy from beyond oneself’’ (1988, p. 104). Individuals become dissociated from the larger community in which they are embedded. As a consequence, holistic ways must be found to integrate the task of earning a living, which we view as profane, with our cultural heritage of spirituality, which we view as sacred (Heinze, 1991). The question then becomes: What kinds of leadership are required for each phenotype and organizational level? Business mid-level leaders and senior leaders, who tend to be located in global and regional organizational levels, think in managerial terms of ‘‘tit for tat.’’ Inspirational leaders, who are usually located in national organizations at the senior and middle levels of leadership, feel comfortable with Harmon’s (1995) ideas of integration of communal and individual responsibility. Global and regional mid-level and senior leaders are comfortable in social spaces that call for social distance. Their transactional bias is expressed in terms of boss/subordinate or teacher/follower, whereas inspirational leaders prefer small to nonexistent social distances when forming their relationships. These mid-level leaders work for senior leaders Enviroscapes 377 in terms of executing strategic intents in the form of operational strategies. When being task oriented, mid-level leaders consolidate, senior leaders redeﬁne, and transformational leaders create. When working with people, mid-level leaders control, senior leaders align, and inspirational leaders liberate. Within the task and human interaction contexts, mid-level and senior-level leaders require structure and constancy, whereas inspirational leaders prefer limited structure and ambiguity.

Purposes and Roles of Business Leaders Business leaders tend to use charisma to develop followers, define organizational values, and gain others’ alignment with desired outcomes. These leaders work on altering organizational structure and changing attitudes of people within the organization through the dissemination of their visions. Nevertheless, the relationships they foster always adhere to a leader–follower pattern (Reeves-Ellington, 1999a). Operating in this context, Sinha believes the business leader’s responsibility is to establish a vision that ‘‘touches the very core of the followers’ thoughts, feelings, and aspirations. These leaders should communicate it in a manner that followers perceive to be genuinely natural’’ (1995, p. 41). Pfeffer (1981) and Smith and Peterson (1988) state the problematic slightly differently: Leaders create shared meanings. Schein supports this position, noting that ‘‘a unique function of ‘leadership’ . . . is the creation and management of culture’’ (1987, p. 17). He sees organizational culture as existing apart from its external environment. The business-focused leader is responsible for creating an organizational vision and ensuring others’ alignment with that vision. Business phenotype leaders use the vision as the basic ideation with which to create a sense of the organization’s structure and boundaries (Bass, 1985; House et al., 1991). They focus on setting group goals and moving the group toward these goals by providing meaning to organizational business ideologies and values. They strive to achieve change by establishing their view of order within the organization in ways that ensure the organization is protected from external influences (Thomas & Humphrey, 1994). For the business leader, ‘‘the most commonly used measure of leader effectiveness is the extent to which the leader’s group or organization performs its task successfully and attains its goals ....Objective measures of performance or goal attainment are available such as profit growth, profit margin, sales increases, market share, sales relative to targeted sales, return on investment, productivity, cost per unit output, costs in relation to budgeted expenditures, and so on’’ (Yukl, 2006, p. 6). In contrast, Sinha 378 RICHARD REEVES-ELLINGTON would measure a leader’s effectiveness as depending ‘‘on the extent to which the demands are met and the subordinates are motivated to realize shared goals’’ (1995, p. 33).

Leaderscape Phenotype: Commerce

Inspirational Leaders Inspirational leaders (Humphrey, 1997) are the heart of commerce intercourse. Inspirational leaders are sense makers, whole makers, moral architects, value definers, stewards, servants, guides, optimists, and warriors (Campbell, 1973). They are spiritually rather than materially oriented. They offer direct inspiration in either a learner or a teacher relationship; that is, inspirational specialists either have followers who take their guiding doctrines, or they have disciples (pupils and adherents) (Webster’s New Universal Unabridged Dictionary, 1983). Inspiration is ‘‘concerned with vision, integrity, values, states of mind, energy, sense of community, acknowledgment’’ (Webster’s New Universal Unabridged Dictionary, 1983, pp. 1030, 1667) and deals with effectiveness rather than efficiency (Yukl, 2006, Sinha, 1995). Contrary to business leadership, commerce leadership lies in the authority of individuals. Inspirational leaders take responsibility for interacting with others in ways that integrate seemingly disparate rules of conduct (Kanungo & Mendonca, 1996) through self-empowerment. They cause organizations to interact with a variety of social milieus. These leaders are self-defining, having strong internalized values and ideals. They forgo personal pursuits for actions they believe to be correct. Such individuals have a strong sense of self-worth that allows them to make unpopular decisions. They exhibit a strong sense of inner purpose and direction (Bass & Avolio, 1994, p. 18). Inspirational Leaders: Shaman as Metaphor The shaman is the traditional culture model for inspirational leaders (Krippner, 1988) and provides a compelling metaphor for understanding how this role works in an organization (Hamayon, 1994). We might view such leaders’ function as being to help commercial organizations ‘‘recover their souls.’’ Such a recovery requires inspirational leaders who live out of their imaginations, indulge in humor, and have more profound philosophical interpretations than do other individuals. The shaman is responsible for integrating the sacred (values and ethics) with the profane (the mundane) (Winkelman, 1982). Enviroscapes 379

Shamanism is ‘‘humankind’s earliest and longest-lasting healing, psy- chotherapeutic, and spiritual tradition’’ (Walsh, 1994, p. 7). While shamans have a keen awareness of differences and peculiarities (Krippner, 1988), they alter the Cartesian paradigm of dichotomous thinking by drawing a sense of relatedness and continuums (Klimo, 1988). They satisfy the need for abolishing polarities (Elide, 1964). Shamanism is an appropriate inspirational specialist model for commerce-driven organizations, as humans often need healing to reconcile the methods by which they earn their living and the social environment they inhabit. Determination of the acceptability of ‘‘shaman as empowered and healer’’ requires examination of shamanistic definitions, mission, personal attributes, activities, roles, and outcomes. Walsh believes that shamanism is ‘‘a family of traditions whose practitioners focus on voluntarily entering altered states of consciousness in which they experience themselves or their spirit traveling to other realms at will and interacting with the entities to serve their community’’ (1994, p. 7). Eliade agrees: ‘‘Shamanism is the interaction with the spirit world on behalf of the community, particularly in healing and protection’’ (1964, p. 352). ‘‘Shamans can be defined as men and women who deliberately alter their consciousness to obtain power and knowledge that can be used to help and heal members of their community’’ (Krippner, 1988, p. 306). Shamans achieve their mission of integration through a process of catharsis – purification and ordering of the psyche, an increase in self- confidence and security of those they help (Krippner, 1988). ‘‘The main task is to mediate between the sacred and the profane – to provide others access to the divine’’ (Heinze, 1991, p. 216). As mediators, shamans work between the spiritual and human worlds. Normative rules of social order are suspended in this space, and the encounter with the omnipotent spiritual world is screened. On the one hand, shamans protect the spiritual world from being polluted by human weaknesses and ignorance; on the other hand, they channel spiritual energy into useful ways to avoid overtaxing the capacity of their contemporaries. Shamans successfully mediate by using vision quests, impersonal power, and moral status (Winkelman, 1982). They must actively empathize with persons of both genders and heal individuals, families, and societies through their actions as counselor, healer, prophet, and spiritual guide (Reeves-Ellington, 1999a). Their healing arts and power synthesize a comprehensive worldview. This is the root of human cultural processes that mediate spiritual, cosmic, and material forces. Shaman agency is responsible to communities for making the sacred available. In this way, shamans help communities discover shared belief and 380 RICHARD REEVES-ELLINGTON value systems. This role is crucial, because shared belief and value systems are a vital part of effective relationships (Frank, 1985; Walsh, 1994). Harmon provides specific context in his discussion of social responsibility and agency:

The paradoxes of agency are produced by the splitting off of . . . . the responsible self – that of maker, one who is responsible for the authorship of his or her actions . . . one who is accountable to other members of a moral community. Because agency for and accountability to are complementary, the denial or neglect of one inevitably leads to the atrophy of the other . . . when agency is construed chieﬂy as entitlement, the atrophy or individual agency, or under-responsibility, predictably follows. (1995, p. 156)

In their support of commercial activities, shamans attempt to identify and develop a unifying whole with the external environment. Their mission is that of world making and taking a wider concern with the structure of the several symbol systems of sciences, philosophy, arts, perception, and everyday discourse (Reeves-Ellington, 1999a). Within this context, shamans provide several constructs of their visions – both those that demonstrate unity and those that demonstrate differences. Success in world making requires commercial shamans to increase their clients’ ability to learn about the external world so that they can apply their new knowledge in the organizational world. Successful shamans learn theory, act as mythologists, and become world- wise within the contexts in which they work by learning the terrain of the multilayered, interconnected universe in which they will quest for knowledge and power for their community (Walsh, 1994). They also create an altered terrain for those disciples who want to learn. Shamans are not missionaries working with divine inspiration so as to save people and organizations that do not recognize their need of being saved. Rather, they are visionaries who transfix those capable of seeing and creating a holistic vision and are open to being healed (Pattee, 1989). The shaman as a creative human being experiences empathy with all sentience and, therefore, can enter into the spirit of other sentient beings. As Pattee says, ‘‘Creativity is power to manifest spiritual directives close to the source for the benefit of sentient beings and their greater purpose’’ (1989, p. 152). This creativity is also used for the transformation of emotions, reduction of fear and anger, and development of love, joy, and compassion – all of which play key roles in shamanistic healing practices (Walsh, 1990). Reflecting creative shamanistic acts, clients have the power to transmute elemental forces into spiritual dimensions directed toward self-healing and Enviroscapes 381 integration (Pattee, 1989). The purpose of shamanistic creativity might be said to follow Sufi advice: ‘‘If you hear hoof beats, expect to see horses, but watch out for zebras’’ (Le Vie, 1988, p. 16). Shamans create the ability in others to share in the making and remaking of prospective worlds in ways that are considered both socially acceptable and responsible on the part of the commercial organization and the society in which they are situated.

Commerce Language of Inspirational Leaders While discourse is the internal form of business language in commercially focused organizations, dialogue is the external and interactional language of commerce. Discourse provides assurance of business focus and direction; dialogue is the natural communication form preferred by society. Dialogue is used to effect healing of the organization and integration of the organization with its wider environments. Understanding when to use discourse or dialogue requires knowing the individuals involved, the subject at hand, and the context of the problematic issue. The shamanistic metaphor also applies to transformational leaders’ dialogue – that is, the dialogue between the internal commercial organization and its wider social environment. Shamanistic dialogue provides an ability to make social processes transparent – social processes that express relationships between the sacred and profane in ways that others can understand and that use the opportunities offered by these relationships. In this way, shamanistic dialogue provides the language of a helper and healer to those wanting to integrate with the larger world in responsible ways. The transformational leader who uses the shamanistic metaphor expresses empathy but does not take sides in any dialogue. Rather than act in a confrontational manner, the transformational leader provides healing powers that seek to resolve differences or disharmony through the transparency of language. The underlying assumption is that power of clariﬁcation and synthesis is the power of the shaman but the actual healing power is that of the client. Shamanistic language emphasizes wholeness, wellness, and integration. Those using it provide symbolic representations of the sacred in the material world through the words, body language, and processes with which they work. Within the commercial context, a dialogue is necessary for people to reclaim their sense of personal vision, divinity, and destiny. Transforma- tional dialogue is the dialogue of androgyny, which is accomplished by transcending pairs of opposites and attaining equilibrium between two polar forces in a relationship of dynamic tension, uniﬁcation of planes of beings, 382 RICHARD REEVES-ELLINGTON the intuitive and the rational, and the visionary and the ordinary (Halifax, 1978, p. 28).

Inspirational Leaders at the Periphery Contrary to current beliefs that change leading to strong commerce interactions almost always occurs at the organization’s center in contexts reﬂecting uncertainty, inspirational leaders tend to cluster at organizations’ margins. The diverse orientation of inspirational leaders requires them to inhabit different organizational realms and to operate within different rules and roles. Inspirational leaders think of hierarchy in terms of the total society and its various operating climates. They work from the margins where people tangentially touch one another. Inspirational leaders occupy themselves with options and opportunities to interact continuously and pragmatically with the wider environment. They necessarily mediate and interpret between systemic levels, but they often resist or reinterpret mandates and limitations imposed by outlying social systems. If imposed constraints cannot be avoided, successful inspirational leaders creatively open the organization to reinterpret the dissonance and ambiguities within the larger systems under which they operate. Once open to these dissonances, they provide processes by which other organizational members can become integrated into the wider environment, while still maintaining their drive toward business objectives. In his early days at Apple, Steve Jobs was an exemplar of an inspirational leader working from the margins. His success in this role sprang from his independence from traditional, internally oriented organizational roles. Jobs provided a trusting bond between Apple and the wider society Apple inhabited. He personiﬁed Apple’s creative urge to transform the computer industry from a business-oriented model to a commercially oriented one. Early in the history of Apple, Jobs – acting out a shamanistic metaphor – created a learning organization that was attempting to integrate itself into a particular cultural logic and prescriptive ethical set.

Leaderscape Summary

The purpose of business organizational leaders is to create an environment that is conducive to organizational gain. This gain may be distributed in uneven ways starting at the top. By comparison, the purpose of commerce inspirational leaders is to enter into discourse in ways that create new realities and imagined communities that beneﬁt all parties. Enviroscapes 383

Business leaders define success in terms of efficiency, as measured by task completion, working within the law, and internal recognition of leaders and managers. They do not concern themselves with social morality, unless it is defined and enforced by law. Instead, these leaders manage rewards, have the high contextual IQ required for understanding an evolving business environment, and adjust organizational business to better suit current needs. Inspirational, commerce-driven leaders consider socially responsible actions and measure success in terms of the levels at which they integrate with those with whom they work, both inside and outside the organization of employment. They have the ability to manage wide-ranging relationships, are persuasive, and balance ideals and capabilities.

ENVIROSCAPES SUMMARY

Two organizational phenotypes are common to all of the enviroscapes: business and commerce. Each has a paradigm that differs greatly in its processes and outcomes. The business phenotype paradigm is a producer/ consumer model, and the commerce phenotype paradigm is a customer/ supplier model (Reeves-Ellington, 1995a). Drucker supports the business phenotype paradigm: ‘‘Business performance comes ﬁrst – it is the aim of the enterprise and the reason for its existence’’ (1955, p. 16). Levitt states this point even more forcefully: ‘‘Altruism, self-denial, charity, and similar values are vital in certain walks of life. But for the most part, those virtues are alien to economics. Further, the governing rule in industry should be that something is good only if it pays. Otherwise, it is alien and not permitted’’ (1958, p. 48). Within the customer/supplier model (commerce paradigm), a constant interchange of information leads to socially embedded business activities. While the focus remains on the movement of goods and services at a proﬁt, actual exchanges occur in ways that are socially acceptable to all key constituents. When commerce organization interactions occur in practice, they involve commercial actors situated marginally to the centers of authority. Such marginal actors operate closer to other institutional centers of the immediate social and physical environments. In exchanges between organized systems of different size, complexity, and resources, the circumstances are asymmetric, and socially responsible business is always organizationally compromised (Reeves-Ellington, 1999a). 384 RICHARD REEVES-ELLINGTON

EMPIRICAL CASE STUDY

Indonesian Plant Start-up: Three Enviroscapes Illustrated

Using the enviroscapes phenotypes of business and commerce, this case study illustrates how the three enviroscapes of climate, knowledge, and leadership affected the start-up of an Indonesian plant. Each set of materials is analyzed using levels of analysis (LOA) and levels of management (LOM) as shown in Table 1. Without grounding enviroscape theory in concrete examples, the richness of this theory as well as the difﬁculties it presents in a variety of contexts are less clear and open to misunderstanding. The case study takes the reiﬁed theory and places it in environments in which organizational actors work. Given that empirical case materials and ethnographic analyses exist elsewhere for ethoscapes (Reeves-Ellington, 2004a, 2004b) and timescapes (Reeves-Ellington, 2007a, 2007b), they are ignored in these case materials here.

Building a Pharmaceutical Plant in Java

This case study recounts the circumstances and business decisions associated with buying land, building a pharmaceutical manufacturing plant, and starting up manufacturing operations in what was at the time a rural area of East Java. In this plant start-up process, a number of considerations operated together at various levels and from various perspectives. The staging of events began with the initial requirements for regional approval of a site by the Surabaya Department of Agriculture and then moved through speciﬁc landowner identiﬁcation and subsequent negotiations of land purchase, actual purchase and payment, and plant siting and construction. National project leaders in Indonesia worked with and within three national cultures – Javanese, Chinese, and corporate American (in this case, the US company P. T. Essex). There was a constant need to integrate the three cultures into a common and agreed-upon commerce strategy that met the ‘‘business’’ needs of each of the parties. Each business task brought up its own particular issues of commerce, yet the whole had to be seen as both relatively economically successful and socially coherent for its participants.

Business Phenotypes of Knowledge, Climate, and Leadership During this period, Indonesia was experiencing the same pattern that has become all too familiar in newly sovereign countries: international firms Enviroscapes 385 wanting to invest, international business and investment consultants appearing to support both the new investors and the local government, and a supportive banking system being developed. This consulting environment spanned the full social and ideological range. At the same time, many of the Indonesian government’s officials had become expert in identifying and overcoming internal barriers and, occasionally, in creating further ones that were most conveniently remediated through the right payments to the right people. International political and intergovernmental advisors were concerned with reducing what they then perceived as the Communist threat through the creation of a popular Rostowian concept widely known as economic takeoff. The national government maintained a putative one-stop central Investment Board, then under the direction of Barli Halim, a colleague of President Suharto. The realities for Western investors, however, were anything but one stop: Halim’s board would hand out the forms, but the investors had to find, cultivate, and collect a multitude of signatures from appropriate government agencies. A partial Jakarta-level list of such government agencies, for Essex’s intended project, included the following: for the site, the city planning board; for the equipment and design, the Ministry of Light Industry; for financial aspects, the Central Bank and the Ministry of Finance (whose financial officers reputedly did not get along); for duty relief, Customs; and for all employee matters, the Ministry of Labor. At each ministry, a minimum of 12 signatures was required from given departments and levels – this included 115 signatures for the central Jakarta portion alone (Reeves-Ellington, 1976). Only upon approval of the central government offices could an investor start work with the regional counterparts for each of the central government ministries. Money exchanges were the norm for gaining ministry signatures at all levels. In the late 1960s and through the 1970s, the investment policy of the Indonesian government emphasized import substitution in manufacturing, with pharmaceuticals being an important consideration. Until 1972, all pharmaceutical investment had been located in the Jakarta area. Initially, there were no minimum investment or product line requirements. After 1972, however, the government closed Jakarta to small industry. Any new investor would have to go either to the outer islands (where there was virtually no market, transportation, or other infrastructure) or to East Java. All foreign investors worked on a cost-plus basis until the late 1960s. After 1969, any new foreigners were required to set up joint ventures with local partners. Local Indonesian nationals had to have a minimum 5% investment, with a recommended entry level of 10%. Virtually no one had 386 RICHARD REEVES-ELLINGTON any money to invest, and partner selection was a difficult issue; most who wanted and could afford to invest were ethnic Chinese. In the mid-1960s, Schering Corporation, a major US-based multinational corporation, under the guidance of its chairman, W. H. Conzen, embarked on an Asian strategy of an expansion program, one leg of which was investing in pre-identified markets characterized by substantial populations, rising incomes, investment-friendly governments, and willingness to opt for important substitution. In the execution of this strategy, the company’s Hong Kong-based regional management ran all of the distribution arrangements throughout North and Southeast Asia. Based on this experience, the Hong Kong vice president, Roland Weiser, and the regional director, Fred Shear, recommended investing in Indonesia. Richard Reeves- Ellington, the future vice president of the Indonesian subsidiary, was tasked with the gathering of information for the investment and provided on-the- ground knowledge and execution of the investment strategy in Indonesia. In 1972, Essex received Indonesian Investment Board approval to invest in Indonesia, as a subsidiary of Schering Corporation. By the time of Essex’s arrival, 16 pharmaceutical companies were already active in the Djakarta area. As a result, the Investment Board required the company to locate in East Java. Essex’s plant would be making tablets, sugar-coated pills, ointments, liquids, and sterile injectables. By the time the project was ready to begin, the minimum investment would come to $1 million. The business steps involved working with the regional East Java government in negotiating the rights to purchase land in an approved industrial estate, making an offer for land purchase, making payment for the land, registering its new ownership, ensuring land protection, plant siting and dedication, hiring workers, and opening the plant.

Commerce Phenotypes of Climate, Knowledge, and Leadership Essex’s strategic intent of what to do from a business perspective was clear. In addition, anticorruption principles and US laws offered some guidance for how to do business in emerging markets. Nevertheless, national ‘‘ways to do business’’ conflicted with those traditionally espoused by US firms. Commerce could be difficult in an Indonesia environment, because local contexts would not go away nor bend very far to accept US commerce demands. Commerce discussions then revolved around the payment of ‘‘tea money’’ and the method in which it would be paid. The following discussion focuses on how tea money worked and how Western businesses accommodated this practice. Three formalized responses were then possible: exit, voice, and loyalty (Hirschman, 1970). Enviroscapes 387

As has been the case in some other national business settings, foreign businesspeople in Indonesia often found that providing a gratuity for services – cash, gifts, or business opportunities – seemed to make it easier to get administrative gatekeepers to expedite their shipments, applications, or requests. From this recognition arose the concept of ‘‘tea money’’ – that is, a little extra money to buy tea (Reeves-Ellington, 1980). Tea money had long been integral to the Indonesian circumstances of political and social emphasis on minute negotiations and mutual social obligation. As an artifact of subtle accommodation among the potentially competing interests of different social groups, the roots of tea money lay deep in the country’s often bitter and quietly ambivalent relations between rulers and ruled, from the times of the early Southeast Asian kingdoms to their colonial and modern military successors. Multiple layers of social hierarchy persisted, reflecting both continuities and sequences of Javanese contact with Asian northerners, the Dutch, and the Japanese military. The elaborate system of ritual exchanges and customary expectations of mutual help extended through all areas of life. They could clearly be seen in the social institutions of sharing resources during wet rice irrigation and in the hierarchical prerogatives accorded to the prijaji, the gentile, customary ruling class (Anderson, 1972), who became the civil service class in modern times, and whose familial-based aristocratic status extended historically to traditional Javanese court society. In the tea money system, help returned was viewed as commensurate with the help rendered. Status and ritual obligation, rather than the preference for impartial fairness of law or duty that lies at the basis of Western business ethics, formed part of the cultural logic for a social knowledge of business relations. The cultural ethic of rukun, which has been well described by Geertz (1961), prescribes an appropriate demeanor and etiquette – the emphasis is on dignity, pleasant composure, and avoidance of direct confrontation related to individuals’ differences. This understanding has supported a patient working out of minute and complex interdependencies in this hierarchical framework. In the cultural context of US multinational businesses at that time, the popular and politically correct view was that virtually any informal or personal compensation went beyond the pale of ethical business relations. Many Westerners interpreted tea money gratuities entirely from within their agenda-driven ethic: It was corrupt bribery, no matter in what scale or form – and, besides, it would add to costs. Americans in particular often responded to tea money requests with righteous anger, impatience, and arrogant outbursts regarding services they felt entitled to demand. For the 388 RICHARD REEVES-ELLINGTON most part, this response simply resulted in higher payments of tea money being demanded and less help and service being extended. Both sides, affronted, could see their business relationship spiral upward in dollars and downward in respect. This was problematic in that respect was another very high priority in the regional values orientations – indeed, it was a quality absolutely necessary for continuing successful business relations in Indonesia. An alternative response was to flout the rules of Western business ethics and buy cooperation with copious amounts of US dollars. This practice would also result in resentment and a heightened distrust between Western business partners and local people, with the added result of further inflation in payoff expectations. All Western companies that stayed to do viable business in Indonesia did from time to time pay some type of gratuity that Indonesians would have seen as tea money. The tea money concept spanned the entire range of statuses, functions, and amounts. Did all firms pay tea money in circumstances and amounts that would look, from the US business standpoint, like corrupt bribes? No, but it is safe to say that most did, on the report of multiple sources Reeves-Ellington considered reliable at that time. This was also highly plausible in light of the fact that most Western businesspeople – particularly Americans – believed they had little time to invest in the complicated task of learning and responding appropriately to local cultural interpretations. Economic drivers were the only factors seen as being important to most business investments at the time. Without a commensurate investment in cultural learning for the fitting response, US companies would not have begun to see and maximize mutually responsible opportunities in the relations of this highly compromised and culturally complex setting. If they had done so, the many pitfalls of payoffs might have been avoided. The option of righteous refusal was also generally unsuccessful. Some righteous scenes of American confrontational outbursts did occur; these were received by Indonesians, at least in Reeves-Ellington’s experiences, as unseemly and disrespectful. (According to a local saying, ‘‘The empty barrel makes the most noise.’’) The complainant would either end up out of business or pay more anyway, and receive poorer assistance in exchange. To learn a more culturally responsive ethic for this context took time, tenacity, and astuteness. It also took the cultivation of relationships with persons who could both tolerate the investors’ lack of social knowledge and guide them toward more successful interpretations for action. Essex’s foreign managers, after observing how irritated objections to requests for service gratuities simply made things worse, set about discovering Indonesian Enviroscapes 389 equivalents to business gifts and relationships that would be both respectful and culturally acceptable for both sides. This process necessarily involved mistakes, recoveries, approximations, small compromises, and continual and persistent information gathering and verification. Building good faith among area officials eventually obviated the need for bribes in most cases; greater and cultural astuteness provided ways to avoid them in others. In one example (which was typical of many other exchanges), the demand of a Jakarta health department official for a bribe to obtain his signature was overcome by social relationship building and culturally appropriate demonstrations of respect. The Essex management staff, which included Javanese individuals, learned of the birthday of the official’s wife’s. They arranged to take the right gift and flowers to the man’s house at the right time in the evening. While such touches came to typify the actions of many so-called customer-driven US firms in the 1990s, especially at the high-end customer and executive partnership levels, they were unusual in the 1970s and in a foreign country. This aptness of this response was particularly notable in that it occurred on Indonesian terms in an Indonesian context. In the case in question, without paying a bribe, Essex obtained the official’s signature within days. Many other instances like this one arose during the project. Hosting wayang, the area’s traditional and highly esteemed shadow-puppet play performances, and inviting local guests as well as company people to participate in ritual occasions worked against the foreigners’ inevitable social mistakes and helped to counteract somewhat the artifacts of excess and disrespectfulness in foreigners that would be likely to attract demands for bribes. This sort of thing also was not widely done in the early 1970s in Indonesia. Beginning in the later 1970s, however, US firms were forced to consider personal relationships and responses to alternative cultural contexts in the face of competitive examples of manufacturing and international trade success established by the Japanese.

Doctrinal Responses to Indonesian Commerce Building on the work of A. O. Hirschman (1970), three doctrinal options may be identiﬁed that match with either commerce phenotype: exit, voice, and loyalty. One of these options (loyalty) actually covers both phenotypes. These options reﬂect organizational actors’ preference to work predominately within an organizational context or be willing to developing a comfort zone in the commerce intersections in methods of doing business. 390 RICHARD REEVES-ELLINGTON

Exit. A strict exit doctrine with respect to working in a commerce climate requires a rejection of business if viewed as in conflict with the wider social mores of the home organizational environment. Exit doctrine is grounded in the desire for risk-free, well-structured environments that are managed efficiently. Exit or near-exit proponents have generally stopped looking for solutions from within business itself. Instead, they tend to look for social and ethical effectiveness of interventions that come from alternate business and allied business climates. If they assume a political economy perspective with conflict between collective interest groups, they postulate a need for social change at a larger scale. This typically includes some kind of business- based, political agenda. Bodley (1996), for example, addresses large-scale finance capital and its associated outcomes of inequality in business globalization as real and imminent threats to small-scale cultures. He concludes that ‘‘Political power must regulate the business and financialization processes’’ (p. 210). Yet, he goes on to say, ‘‘The primary obstacle to such change will [be] . . . the ability of elites to control the cultural symbols that motivate human beliefs and behavior for their own interests’’ (p. 214). Bodley recognizes the need for social interactions (commerce) but also recognizes the difficulty of directing cultural change by agenda, so he is left hoping that a global commerce will emerge. In other words, he has largely given up on social responsiveness and would have organizational actors exit from a business climate without offering alternative options. O’Connor (1994) declares that radical greens, deep ecologists, socialist biocentrics, green anarchists, and ecological feminists all agree that global capitalism is not sustainable and that withdrawal from all existing business climates is the preferred action. The exit school validates only the commerce climate and would have the business climate disappear.

Voice. Voice as a basic doctrine actively questions how well business organizations work in broad international contexts. Proponents of the voice doctrine have no intention of abandoning existing business climates and knowledge creation nor of fundamentally restructuring or replacing them. Voice advocates are often more pragmatic about the proposed method and actions at smaller levels of inclusion and impact. They will often appeal to outside authority and public opinion to try to effect changes and reforms within existing institutions. As Hirschman (1970) points out, while the voice doctrine does not include imminent plans to leave the entity that is perceived as not meeting expectations, it does propose to influence change and reform through a mobilization of the effective threat of exit. This threat is one often Enviroscapes 391 employed by organizations that view commerce as a menace to their business climate. Within an organization, threats to leave by influential members rarely succeed in influencing organizational climates and knowledge creation (Reeves-Ellington, 2007b), for obvious reasons relating to power and authority. In rare cases, the parties are prepared to follow up by vigorously engaging in strategic battles for their own positions within the organization (Reeves-Ellington, 2007b). But, from inside late modern business itself, voice is a strong doctrine to support existing business climate practices.

Loyalty. Loyalty addresses climate and knowledge through a sense of integration of business and commerce climate phenotypes primarily within the scope, initiative, and power of business and social institutions as they exist (Reeves-Ellington & Anderson, 1997a). American business ethics textbooks reﬂect this outlook when they implicitly stake a claim of ownership to, and the adequate authority and expertise to determine, their own occupationally based moral community. The loyalist canon of normal business accommodates a loyalty canon and interpretation of organizational activities. The underlying principles are that well-run business and commerce are inherently driven by social compromise, are morally charged, and require ongoing social integration. Fundamentally, the loyalist works with existing business and commerce phenotypes but in ways that result in the integration of the two without constantly have one assert privilege over the other. The climate phenotypes of business and commerce provide organizational structure that shapes knowledge creation and use. For clearer understanding, climate knowledge can be augmented by examining social and organizational learning theory (James et al., 2007; Reeves-Ellington, 1995b; Reeves-Ellington & Anderson, 1997b).

Global Business and Commerce The Essex home office leadership, under the guidance of W. H. Conzen, had a clear strategic intent of business expansion throughout the world and developed financial, manufacturing, legal, and business support to support that intent. The intent was clearly communicated throughout the organization, with the expectation that it would be translated into regional and national actions. Once field recommendations were accepted at the global business level and implementation strategies set, support staff worked with regional and national organizations to facilitate approved actions. 392 RICHARD REEVES-ELLINGTON

Conzen’s background also provided a global perspective on doing business locally in terms of marketing, sales, and distribution efforts. Of more importance, his life experiences (discussed later) provided Conzen and those who knew him with the inspiration to allow leadership teams at the national level to work in ways that would fit in with the local social conditions. He did limit these interactions to ones that would not be considered illegal in the United States or likely to cause embarrassment to Schering. Certainly Conzen allowed flexibility for managers to work with the US facilitating-payments guidelines, but he expected the local leaders to understand how to do that in ways that would be socially acceptable to the hosts. If tea money was to be paid, it would be paid in the same ways that indigenous organizations would facilitate payments. Conzen encouraged the doctrinal position of loyalty in commerce interactions throughout the organization. This position required iterative learning throughout Essex. He was willing to accept effectiveness in the commerce climate, even if it caused some inefficiencies in the business climate. His concept of loyalty might include loyalty to the organization, but that would in turn require loyalty to those outside the organization who were able to affect business performance.

Regional Business and Commerce Schering’s regional office team, under the leadership of Roland Weiser, had more than 80 years of the experience in Asia. The members of this team also brought with them personal experiences in Hong Kong, Korea, Vietnam, Japan, and Australia. In short, they knew how to do business in all of Asia. Following the strategic intent of the US office, they collected information from all the markets that fit the profile for a distribution arrangement, joint venture, or green-field start-up. They worked with their national teams in collecting information, analyzing it, and forwarding the resultant knowledge in the form of business recommendations. They took responsibility for recommending how Schering should interact with each market and what level and type of interaction was appropriate. They also provided administrative support for the national operations. A final key business activity was to coordinate the hiring of expatriate staff, where necessary, for national operations. Weiser supported Conzen’s loyalty approach to commerce activities. When he inquired about national activities, he was learning how to do business in the area or else wanted to assure himself that commerce activities were thought through and did not expose the company to excessive risk. Weiser and his team would also serve as a buffer to headquarters in ways that supported national operations. Enviroscapes 393

Local Business and Commerce The local Schering team was responsible for all business activities in Indonesia, all budget activities that were approved regionally, local hiring, and direct discussions and negotiations with the investment board – in other words, the traditional work of a subsidiary. The team members were also the primary information source regarding local competitor activities, business climate, and political matters. They were expected to take leadership positions in local business organizations. Within their commerce activities, these individuals had far more ﬂexibility. They took responsibility for learning alternative forms and circuits of leadership in national government, regional government, and local government (where the plant was to be located). Each of these external ‘‘partners’’ required the Essex team to expand its knowledge of national, regional, and local differences, thus requiring a diverse leadership group within Essex itself. To ignore any of these factors would be to endanger the project, the people, or both. Rapid multicultural encounters required the Essex team to come to terms with discrepancies between government policies and internal political and regional actors at various jurisdictional levels. In short, the Essex Indonesia commerce leaders had to understand what tea money was, what it meant to various players, and how to arrange it in socially acceptable ways.

Leaders (and Management)

The global strategic intents of Essex’s business, their regional execution, and national implementation of detailed strategies were well supported through all leadership levels. Likewise, the managerial implementation activities were both appropriate and necessary. The local efforts at understanding and dealing with commerce climates and knowledge were also supported throughout the leadership ranks. The backgrounds of key leaders from the global office, through the regional office, to the local office provided an unusual global understanding of the importance of both business and commerce – an understanding that lead to substantial and sustained organizational success. A review of the leaders at each leadership level will demonstrate this point.

Senior-Level Leadership Understanding Conzen’s business and commerce leadership requires a short biography of the man. He began his career in Germany with Schering A.G. of Berlin in 1931. He was transferred to Schering (Pty.) Ltd., the company’s 394 RICHARD REEVES-ELLINGTON subsidiary in Johannesburg, South Africa, in 1938 and was named general manager in 1941. During World War II, the US subsidiary of Schering A.G. was seized by the US government and became Schering Corporation USA. After Schering Corporation USA acquired a major interest in Schering A.G. in 1947, Conzen played a principal role in Schering’s expansion and acquisition program in Europe, Asia, and Australia. In 1952, Conzen was appointed manager of Schering’s International Division. Seven years later, he was elected a vice president of the company and, in 1962, became general manager of the Schering Laboratories division. In 1966, Conzen was elected president, chief executive ofﬁcer, and a director of Schering Corporation. Other senior leaders followed Conzen’s business strategic lead, which was calculated to expand Schering’s business outside the US base, by developing and executing implementation strategies in their areas of responsibility. The processes used for developing action plans were deductive; that is, they operated by excluding all geographic areas that did not meet the necessary criteria. Senior leadership had primary responsibility for controlling strategic execution through budgetary and human resources control. In short, senior leaders managed the chief executive’s discourse of strategic intentions. Conzen had the breadth of experience from working in a variety of geographies and dealing with a range of business problems that required astute political knowledge and interactions. He also understood the necessity of local autonomy through his South Africa and US experiences. Within the Asia region he hired in his likeness to assure commerce sensibilities. His regional vice president, Roland Weiser, mirrored Conzen’s experiences, albeit in differing geographic areas. Weiser was born of Jewish parents in Alexandria, Egypt; attended secondary school in Switzerland; and received his tertiary education in the United States. His work experience included stints in the United States, Australia, Hong Kong, and Japan. Conzen’s focus was to value not only the community but also those communities surrounding the company. He practiced this in Bloom County, New Jersey, the home of Schering’s global headquarters, and expected his team to create much the same community feeling in areas where they worked. His commerce strategic goal was to integrate the company into its communities through the liberation of the company from narrow business structures to those that affected the surrounding communities. Weiser saw his role as supporting this strategy and assuring it was understood and implemented throughout his region. The commerce aim of senior leadership was to learn what was necessary at the local level, including tea money payments, and to achieve a communal, holistic result. Enviroscapes 395

Mid-Level Leadership Mid-level management’s primary commerce activities were managerial in nature. Those mid-level visitors not only brought their business expertise to Essex but also provided social support for employees that helped bond them to Essex. Their conversations with national counterparts outside the organization supplied Schering leadership outside Essex with business information that supported the local staff. Further, their returning war stories provided commerce knowledge that created a climate of trust between Essex and other Schering organizations. When local employees returned on visits to the US headquarters, adequate social events were planned around specific individual interests. For example, the Essex plant manager and the general manager enjoyed opera; on each visit that they made to the United States, Metropolitan Opera tickets were made available to them. An Essex Indonesia employee wanted to visit Radio City Music Hall, and his US counterpart arranged for tickets for the Christmas pageant and accompanied the employee to see it, even though it was a two-hour trip into Manhattan. The local-level management was necessarily managerial in dealing with business activities, but in their commerce activities they had to provide inspiration. It was largely at this level that mid-level nationals played key roles in learning what was necessary for integration of the company into its wider social environment. They took responsibility for learning whether tea money should take the form of cash, flowers, alcohol, or another type of personal support to someone outside the company. They also took responsibility for providing the information and establishing the personal climate that would create strong bonds throughout the organizational community. To Essex outsiders, these leaders were viewed as empathetic learners who understood local social and business climates and could faithfully inform and influence more senior Essex employees. Junior-Level Leadership The junior leadership members at the local level had an important role in commerce management. Their role was very similar to that of the senior and mid-level group but dealt with junior levels in other organizations. For example, when needing signatures from port authorities, the junior staff could push through the paperwork if encouraged by Essex junior staff. When higher levels attempted to carry out this task, failure was often the result. For example, when the company was having difficulty in getting key inventory released at the Djakarta port, Reeves-Ellington went to the senior customs authority to apply some pressure in the form of rational argument. 396 RICHARD REEVES-ELLINGTON

Logic failed, and he got loud. As he reported later, he could see shades drop behind the eyes of the Javanese ofﬁcial, alerting Reeves-Ellington that he better leave quickly. In a few days, a clerk from the company spoke with the Javanese ofﬁcial’s typist, learned some important social information about family events, and provided a small gift. The paperwork was then promptly processed. The lesson learned: Have the right level deal with commerce activities.

Commerce Activities for Obtaining Plant Land

While the business aspects of gaining investment permissions, land-use permissions, purchase laws, and building permits were reasonably transparent, the commerce actions necessary to successfully complete each of the activities – negotiating land purchase, making an offer, agreeing on a land- use transition, payment, and registration – were not. This section describes the climates, knowledge creation, and leadership contexts that had to be negotiated and outcomes that had to be arranged so that they aligned with the home ofﬁce, national, regional, local Indonesian government entities, and citizens of the village in which the plant would be located. It also analyzes the information in multi-level leadership and geographic constructs.

Negotiating the Land Purchase The regional administrative center for Essex’s intended project was located in Surabaya, East Java. This center represented yet another site in which signatures were required and added some new features to the social and political contexts. Essex’s advance work on the institutional background had revealed only that there was a lack of clarity about the relationships between the national and the regional authorities. These were unclear not only to the expatriate business consultants, but also to the firm’s US government contacts. When Essex pursued signatures at this level, as necessary for approval of the company’s offer to purchase the government-designated land, the vagueness of the linkages was made clearer: While the target site of Essex’s interest had been officially designated by the Jakarta office as an industrial location, it was still listed as farmland under the jurisdiction at the regional Surabaya level and was found to be under the purview of the Ministry of Agriculture. Furthermore, it was left up to the investors who wanted to obtain the required government approvals for the intended purchase under the new purpose to arrange for the entire transfer of uses themselves. Enviroscapes 397

Thus the regional government operated under Jakarta in theory, but not according to the practices of the legal infrastructure and the behavior of the officials. Put simply, Essex discovered that the central government at Jakarta had no practical control over the disposition of the land it was offering, nor would the central government want to coerce the local inhabitants. The peasants’ background was also infused with an older, syncretic mix of animism (a tradition known as abangan), overlain by the region’s previous Hindu–Buddhist tradition and by the more recent influence of Islam. The local peasants were strongly land centered and were said to have been staunch supporters of the deposed Sukarno (the first president of Indonesia). Thus, in East Java of the early 1970s, even a national government that has been described as notoriously repressive faced certain practical limits in the exercise of its power – notwithstanding its reification by some area scholars, such as Anderson (1972), for example. Given this local history and access to several insiders’ knowledge of this area, it was Essex’s strong impression that the regional Surabaya bureau- crats would be as wary of causing local offense as they were fearful of Jakarta’s wrath. This was not only consistent with the area’s recent political experiences, but was also culturally consistent with the dukun (a traditional abangan religious leader of a shamanistic tradition) ethic of conflict avoidance and the social knowledge stressing implicit rather than explicit communication of differences. Furthermore, from either a political or an economic standpoint, why would the regionally based authorities go out of their way to break local rice bowls for foreigners? While these regional conditions left the entire purchase agreement responsibility up to each company that wanted to build in East Java, and while the conditions were certainly anything but ideal, Essex officials’ held the opinion that this was still the best location available to investors, in an area widely deemed to have great long-term economic growth potential. The gap in background institutions had left the local communities open to a wide range of possible ways of dealing with those entities wanting to buy land, and it was here that stories of corrupt foreigners had found their most fertile ground. When Essex realized that it would have to make and negotiate its own offer for land at the village level, micro-level issues and face-to-face commerce relationships brought social responsibility issues to the fore. The company’s Javanese employees were worried about their fellow country people, so the firm’s lawyer in this transaction, who was from a nearby village and known in the hamlet where the purchase offer would be made, 398 RICHARD REEVES-ELLINGTON became centrally involved in the negotiation process. Other involved people, beyond the families themselves, included the local dukun (the curer and spiritual advisor, who also headed the local mosque) and the hamlet’s traditionally elected lurah (village head). (Geertz [1961] provides a good description of these roles.) The terms and conditions of the offer would have to be acceptable to all parties for the deal to succeed. The village setting presented a complex array of procedural traditions, but adat (traditional indigenous) legal conventions prevailed regarding land tenure and inheritance. It would be of major importance that all transactions be done publicly, and that they have the participation of all parties, including the land-owning families involved in the deal. But the real surprise, at least from the Western multinational company’s perspective, would be the entry into this process of a more modern religious influence, one with social prestige implications – specifically, the valuation the peasants placed on the hadj, the Islamic pilgrimage to Mecca.

Making an Offer The land purchase and subsequent activities involved a large cast of players. The purchase also involved several social and legal systems: Indonesian law (based on Dutch law), US law, Muslim law and religious prescriptions, and Javanese adat, with its infusion of local animism and social customs. Essex’s research in the area had given the firm a fairly clear idea of the commercial value of the land. Company representatives initially checked their perceptions by reviewing the firm’s intended offer with the two local village officials. This consultation took place before the representatives made their initial approach to the peasant families involved. The company’s intended offering price was confirmed as being fair by both the lurah and the dukun. Given that the lurah was selected by the villagers themselves and the dukun was locally powerful and influential, the input of these persons would be important to the foreign investors’ achievement of sufficient initial cultural alignment for a respectful approach to the families. The actual land negotiations excluded any officials other than these two local ones. The company’s own locally knowledgeable staff (including the Javanese personnel officer and the Essex lawyer) verified that the lurah and the dukun were the persons the landowners could be expected to trust most. The initial approach consisted mainly of polite social introductions, in ways traditional to local Javanese, between the peasant owners and Essex’s Javanese representatives. Essex’s personnel officer and lawyer had spoken initially with the lurah and the dukun. The purpose was to ‘‘introduce’’ Reeves-Ellington as Essex’s US manager (he was not present at the initial Enviroscapes 399 meetings) and to bring up the matter of Essex’s interest in purchasing the land. After proper introductions, Reeves-Ellington followed, observing appropriate local conventions of indirect approach. For example, he brought flowers, drank tea with the local officials, and joined polite conversation about wayang. During a series of meetings that followed, the three Essex representatives – two Javanese, one American – heard the lurah out as he described the families and their hardships on the land. He told them that over the generations, the division of inheritance of land in some large families was ‘‘breaking up vast holdings.’’ On follow-up investigation, Essex learned that the actual size of the largest holdings approximated 1.5 hectares (2.25 acres), and some of the current plots were down to meters square in size. About half were too small for the use of a water buffalo. (The families did not own these animals, but they would have had access to them during plowing season.) The local officials eventually introduced the Essex representatives to the landowners themselves. These initial meetings went quite well from Essex’s viewpoint. After several meetings, however, no deal had been struck, and Essex officials began to be frustrated. The peasants had shown interest and were not asking for more money, yet they continued to indicate that they did not want to sell for the stated offering price. The peasants who owned this particular land had been long integrated into the cash economy; they purchased much of their food and paid cash for the schooling of their children, for clothes, for personal items such as glasses, and for farm implements. Similarly, above-subsistence economic relationships and dependencies had existed with the ethnic Chinese, who controlled the milling system and would keep some of the rice as payment. The Chinese would also purchase rice produced should farmers need to sell it. Often, the woman of the household, and in some cases both husband and wife, would supplement the family income by running a warung (small market stall), turning over extremely small amounts of money by Western industrialized standards. (Geertz’s [1961] study of family economic roles and Alexander’s [1987] study of rural Javanese market traders give valuable ethnographic details relevant to these local systems.) Seeing the apparent impasse that had developed, Essex consulted separately with the local village officials, who in turn approached the landowners themselves; the landowners still did not budge. Reeves-Ellington then met separately with the lurah. He asked him if he would be willing to inquire about the possibility of other, nonmonetary issues that might be blocking progress in the negotiations. Of particular concern to Essex was whether pressure for payoffs was being applied to these peasants from the 400 RICHARD REEVES-ELLINGTON level of the regional government offices. If the regional investment or agricultural offices were demanding substantial payment from the peasants, perhaps Essex could help in resolving some of the needs. The lurah assured him that there were no issues of any consequence in this regard. Then, within two weeks, the lurah returned, relaying two concerns from the peasant families. Before reviewing these issues, he confirmed to the Essex representative that the price itself was satisfactory; it was not more money that the landowners wanted. Rather, the peasants wanted to use their land- sale money to buy tickets to go on hadj, the pilgrimage of purification to Mecca. Throughout Muslim society, hadjis (those who have made the pilgrimage) gain significant status, both religious and social. In Java, they were titled and wore special clothes. If contemporary Muslim connotations of the hadji’s piety and honesty are any indication, the status had some potential for enhancing the person’s chances of business success. This would have occurred through a higher general appraisal of trust and social expectations of piety on the part of a hadji within the community. While such poor people rarely had an opportunity to go on hadj, and this possibility was potentially compelling as an attraction of selling, the lurah reported that the peasants had not been able to gain access to the means for its arrangement, even with their prospects of having a cash windfall from a land sale in the offing. Not only was the number of hadj tickets limited, but the Ministry of Religious Affairs was also not responding to their requests for tickets. The lurah said they saw no need to close the deal for this amount of money if they could not use it to go on hadj. Without the goal of investing in the pilgrimage, the mayor explained, from their perspective they ‘‘had no use for so much money.’’ The Essex staff speculated that the peasants were worried that an unprecedented amount of undedicated money could attract extortion or theft. If this representation against the need of ‘‘so much money’’ seems incredulous to an outside businessperson, in the local context it might also be interpreted as an indirect statement expected to cause further questions that could then be answered. This was how Essex staff decided to receive the message, and this is exactly what happened. Essex responded with further questions, and the local intermediaries uncovered further answers. Direct questions from outsiders would not have been seemly or an appropriately Javanese way to do business in this context. The lurah’s second issue blocking the sale, viewed in retrospect, was related to the first. At the time, Reeves-Ellington discussed it with Essex’s Javanese staff and other Javanese professional contacts for the area; however, at the time it was construed in Western terms as a psychological Enviroscapes 401 need. The peasants still felt reluctance, Essex staff speculated, notwithstanding their awareness of the new opportunities and of the practical diminishing returns of their agricultural land. They were bound to the land in a deeper, moral way. It had been in their families’ possession for many generations. Viewed in local animistic context, the land represented a spiritual tie. According to the lurah, while the peasants wanted to sell, they were struggling with this idea of breaking ties to the land.

Arranging Hadji Tickets Essex representatives took these issues back to the Essex home office. First, as to the issue of hadj tickets: What if Essex attempted to obtain the tickets on behalf of the peasants? The initial Essex home office response was, of course, negative; the idea was viewed as ‘‘crazy.’’ Rationales given for this view hinged on everything from US ideals of separation of church, state, and business, to comments on the folly of a secular, non-Muslim organization trying to obtain tickets for the high holy Muslim pilgrimage. For some time this view prevailed, not only in the US home office, but (except for one manager, who was Swiss) at Essex’s Hong Kong office as well. After persistent arguments on the part of the Java team, but particularly those made by Essex’s Javanese lawyer, who told them the idea was a stroke of genius and should be done, the home Essex managers finally relented. Essex’s senior management agreed to approach their government contacts in Jakarta for advice. Such guidance would not have been available to the peasants. Not unlike Essex’s long trek downward through the hierarchy of officials’ signatures, the peasants would have had a long climb upward through successive layers of regional and district religious offices before being able to approach the Jakarta offices. As it turned out, Essex’s efforts proved quite successful. Going through the National Investment Board, Essex managers began a dialogue between Investment Board officials and senior officials in the Ministry of Religion. This communication began indirectly. One of the Essex managers was on the local international school’s board of education and had made friends in the Indonesian Ministry of Education; these friends had friends in the Ministry of Health and Pharmacy and the Ministry of Investment, in which further links existed to the Indonesian partner of a Western multinational company headquartered in Hong Kong. At the end of this network was the Minister of Religion in Jakarta. Why all the parties helped is a matter of speculation. Essex staff knew for certain that the regional religious ministers were not involved. (Indeed, Reeves-Ellington was told by a reliable source that the officials at this level 402 RICHARD REEVES-ELLINGTON would have been terrified to say anything.) The outcome might have been influenced by the gradual buildup of goodwill that had emerged through the business relationships Essex had cultivated earlier in Jakarta: socializing, birthday flowers, letters of support for college-bound children, the odd pair of tennis tickets in Jakarta – things that are little different from corporate business networking in the United States. Whatever the influences, they did not involve large cash bribes. Perhaps the partners looked favorably on the Essex manager’s support for young puppet carvers and dalang, or perhaps these particular officials liked Americans. Commerce actors do not always keep score objectively; they make relationships intuitively. Thus, without ever approaching directly and, therefore, without broaching certain tensions of Muslim–Christian sentiment that might have negatively influenced the results, Essex managed to gain access to the Minister of Religion himself. The required number of tickets was obtained to enable the peasants to go on hadj. Essex had decided to pay for the tickets as a premium, in addition to the initial purchase price offered to the peasants. This was done for two reasons. First, the cost was minimal, less than 10% of the purchase price. Second, Essex members believed that the impact of putting actual tickets on the table would be a greater incentive toward closing the sale than would offering a letter or merely giving a verbal confirmation that tickets would be available. This reasoning proved accurate for additional reasons. The tickets were concrete; they had the name of a ship, they were usable, and they could not be mislaid by an official or their existence denied, as was often the case with other kinds of official documents. That the tickets came through the Minister of Education directly from the Minister of Religion was not only made clear to the principals negotiating the deal, but became common knowledge at the regional government level, and no one would have dared interfere with the minister’s tickets. Putting the tickets on the table upfront was, indeed, a gamble – but it was also a concrete gesture that made Essex’s good faith clear, and no doubt signified the strength of the company’s relational power. This gesture proved auspicious in other ways. Essex later learned the company had achieved another unexpected bonus with the peasant sellers by providing the tickets upfront. In getting them directly, Essex learned they had avoided the obligation to pay a large sum of tea money to the Ministry of Religion and the Ministry of Transportation. This sum would certainly have been required of the peasants. If only a letter had been presented, it would have been passed to many hands along the way before being received by the peasants. Even if Essex had delivered the letter, it Enviroscapes 403 would have had to pass to still more hands on the way back to the tickets. All of these hands would have been extended for tea money. This amount, Essex learned, could have taken as much as 30% of the total payment for the land. The peasants, understandably, appreciated being able to keep this money for themselves.

Arranging Land Transition The second issue, relating to the peasants’ animistic bonds to the land, was perhaps even thornier than the bureaucratic arrangements for purchasing hadj tickets. The Essex side initially suggested to the lurah that, as the firm did not require all of the land in the initial stages of its commercial activity, it could allow the peasants to continue to farm the sections the company did not immediately need. Others better versed in adat legal conventions, including the lurah, said no: This would make it impossible to claim the land for use later. Essex and local lawyers, in consultation with other area lawyers and ministry officials, confirmed that allowing the same people to continue to use land they had formerly owned would give them de facto squatters’ rights. Essex might well lose any access to the land when the company eventually wanted to use it. After still further consultation with both the lurah and the dukun, another solution was reached. Essex agreed to make some of the land available for public use, with a gradual reduction of land availability over several years. The arrangements for the use of this land would be managed by the lurah.It was agreed that land availability would rotate among all the peasant families that were selling and would start with those who did not own the particular blocks that were being made available for public use. This would allow the peasants to wean themselves away from the land. According to the local lawyers, it would also avoid the establishment of squatters’ claims. The Essex field managers accordingly brought this second proposal to their Java headquarters. They argued that, while peasants were at least nominally Muslims, they were also syncretic animists. From the Essex team’s perspective at the time, the underlying commerce logic involved the argument that the spiritual bonds between people who had been the acknowledged owners and their former land were not easily or automatically broken; this separation would require some time. These individuals’ held beliefs in bonds to the land must receive some respect, and the bonds must be allowed to be relaxed in a more gradual manner. Conversely, allowing the original owners to remain on their former plot would only work in the direction of maintaining those bonds. 404 RICHARD REEVES-ELLINGTON

Finally, a plan was formulated that allowed use of available land that would otherwise have been abruptly put out of production. Because these peasant families were selling their total holdings, the mayor and Essex manager agreed that they would be given priority consideration for employment opportunities in the new Essex plant. This offer was prompted by the concern that the sellers might have no immediate means of support after selling their land. Headquarters agreed, and the purchase offer, including the land-use proposal, was put to the peasant families at the same time that the hadj tickets were put on the table. Within a week, a purchase deal was struck. Thus a Western transaction evolved that incorporated significant social elements of non-Western origin. While not easy to achieve, the land purchase agreement offered some advantages for all sides. The peasant families stood to gain status by being able to go on hadj, and they maintained some access to their traditional lands over a number of years. Their greater psychological and spiritual comfort with this transition was enhanced by the endorsement and subsequent ritual participation of their dukun. The involved parties also stood to gain community status through the multinational company’s efforts to make the land available for public use. Essex obtained the land it needed to carry out the factory project. In addition, the company gained strong local support in its later efforts at local employment. The increased prestige of the dukun and the lurah were also of immense value to Essex throughout the plant’s construction and in the hiring of employees over the next three years. During this time, these officials were instrumental in assuring a protected work site, a good workforce that came to work on time, and the services of some of the best-educated people in the village. The immediate outcome reinforced Essex’s intuitive impressions of the rightness of the way it had conducted business. No complaints emerged as the project moved forward, and the US manager received attention as a Western practitioner who was knowledgeable about Javanese concerns both locally and in the international business press. The lurah was admired for his wisdom in mediating the matter, and, significantly, Essex subsequently had no landowner or strike problems. The local officials not only gained prestige through helping to finalize the sale, but also continued to play supportive roles as the plant project progressed. The dukun officiated at the plant-siting and plant-opening ceremonies and was paid for these services.

Payment and Land-Use Registration Even with the deal struck, making payment for the land in the Essex project was not easy. Essex ofﬁces were in Jakarta, and the peasants were in East Enviroscapes 405

Java, a distance of five hours’ travel. The drive took two days owing to the bad roads. The peasants were accustomed to operating in cash and had never used banks. For them, payment had to be made in cash. The total paid for the roughly 10 hectares was $55,000 (in 1970s US dollars). At the time, the exchange rate was about 275 rupiah to $1, and Javanese per capita income was roughly $100 annually. In rupiah, the largest Indonesian bill was equivalent to about $0.50, necessitating the transport of approximately 135,000 paper bills. This transport of cash was a physical problem for Essex, as – at best – it would require four reasonably large suitcases of currency. The peasants wanted assurances that there would be no recriminations among themselves after payment for the land. For such assurance, they wanted their extended families and the lurah to be present at the counting. This event would involve some 50 or more people. In short, the activity would be reasonably public and well known. Further, the counting would take at least eight hours, so a meal had to be organized for the participants. This made the payment even more of a public event. A site needed to be organized and a caterer found to provide food. (Essex was expected to pay, but the peasants would be the food vendors.) Because of the heat, some cover had to be provided as well. The Essex home office had assumed that the payments would be made by bank transfer. They automatically approved payment by check should a bank transfer be difficult. Readers can imagine the consternation created when the request was made for transporting payment in the equivalent of paper 50-cent pieces. If that were not enough, the highly public nature of the event put the transfer at risk of robbery. Finally, the money paid to each of the landowners would be counted individually – an event that would take all day. The home office said, ‘‘No.’’ The managers then relented and said the payment would be approved if there was cash-in-transit insurance and the payment occurred under the control of the company’s bank. There was no insurance, however, and the bank declined on the basis that the entire event was too uncontrolled and, therefore, too risky. After at least two trips by Reeves-Ellington to Singapore just to correspond by phone and a further trip by a senior Essex finance officer to Indonesia, the payment method was reluctantly approved. The money would not be transferred by automobile, but rather by air, with the last leg of transport taking place by car. The bags would not be checked but would be hand carried, which required buying a seat for each of the bags. Three local managers, including Reeves-Ellington, would be responsible for the transfer. The company did indemnify them in case of robbery (although they were not bonded if any of them had decided to abscond with the money). 406 RICHARD REEVES-ELLINGTON

Purchase of airline tickets was no problem, but explaining to the boarding agent that each bag had a ticket, and therefore a seat, created a problem. Standby passengers wanted the seats, and they did not agree that bags had precedence over people in the seats. The Western logic that Essex managers had purchased the tickets and could put in them what they wanted did not prevail. At the time and place, people prevailed over property rights. The suggestion was made that the bags be checked so people could have the seats. Three two-suitor suitcases plus briefcases were hand carried by the Essex managers; each bag was stuffed with money. The managers did not want to broadcast that the bags were stuffed with money, which was why they could not be checked. Finally, the bags were taken and checked. The team was told that if this solution was not satisfactory, they would be taken off the flight. With the closing at stake, the bags were checked, and the team boarded the aircraft. All arrived safely at the Surabaya airport. The remainder of the trip and the money counting event itself went uneventfully. All the families stayed until all the money was paid. Eight family farms were settled in the same closing. It was unclear (and pointedly not the foreigners’ business to ask, according to the local officials and staff) how many people were involved from each family. Essex staff suspected that, in some cases, one of the selling families actually represented more than one owner; that is, that the eldest sibling represented others. This may have been one reason why so many people were at the closing. The entire economy was coming out of a modified state-owned system and was cash poor. Everyone carried their cash. Essex and the local negotiating team worried about the insecurity of the payment after delivery to the sellers. This led to queries to the local officials, who said, ‘‘Don’t worry – it is not your problem. Besides, we are friends here.’’ Certainly, if the Surabaya authorities had been involved, its officials would have been more than happy to assist; they would promptly have relieved the peasants of some – and perhaps a good portion – of the money. Reeves-Ellington remembers the sight of seven of the family parties walking barefoot down a dirt road, each carrying a rice sack full of money. (The eighth family, apparently, left by a different direction.) The Westerners wondered what the families would do with the money, but had been warned not to ask. Within the week, as it happened, all the sellers did host a party in the village with a well-known dalang. Someone told Reeves-Ellington that the surplus probably was buried somewhere for safekeeping. With the money’s ownership public, officially sanctioned, and religiously recognized, Essex hoped there would be some assurance of considerable social or perhaps even religious repercussions for anyone who dared steal someone’s hadj stake. Enviroscapes 407

Just before boarding a plane at the Surabaya airport for the return trip to Jakarta, one of the managers was approached by a peasant who had sold land. He said that he did not have all of his money. He was short one bundle. The manager replied in typical Western business manner, ‘‘I have yours and the lurah’s signatures on paper that you received all your money.’’ ‘‘Yes [you do],’’ said the peasant, ‘‘but I don’t [have it].’’ The manager looked in his case. There was one bundle of money – the amount the peasant was missing. The money was handed over, the matter was settled, and the peasant was very happy that he had the money.

Multi-Level Leadership and Geographic Location Analysis

While there were business phenotype issues involved in this set of vignettes, the most difﬁcult challenges occurred in the commerce area and in attempts to align commerce issues to business ones. Therefore, this analysis focuses on commerce issues and their alignment to traditional business structures and requirements. In keeping with the theoretical discussion that commerce tends to be local, the analysis presented here starts at the local areas. Collectively, these vignettes demonstrate three points: Organizations are open systems that work in multiple climates; organizational interdependence requires open knowledge building; and organizational environments are highly contextual.

Levels of (Geography) Analysis Local Commerce. Achieving desired business outcomes required the creation of a larger imagined community that incorporated local government and religious leaders into such a community. At the same time, this community had to have regional and national tacit acceptance of the ‘‘community’’ but without those authorities’ inclusion into it. Accomplishing this task required that Essex understand social, religious, and political contexts from the viewpoints of each level of Indonesian actors. It also required that Essex be understood by the local ofﬁcials and population where the plant was to be built. Success required Essex to build extended knowledge networks predicated on information sharing that resulted in shared understandings between the village, regional, and national actors. In addition, Essex had to present argumentation to its own regional ofﬁce that convinced the staff there that acting on the commerce knowledge was a valid course of action and that use of such knowledge would result in attaining desired business results. 408 RICHARD REEVES-ELLINGTON

Global Commerce. The global perspective was that the Indonesian effort not embarrass the company through actions that might cause unfavorable press reports or cause the company to become involved in bribes that might be illegal in the United States, yet take such actions that would result in meeting the company’s business objectives. As indicated, the purchase of hadj tickets worried the home ofﬁce, the transfer of cash in insecure ways caused doubts about doing business in Indonesia, and permitting villagers to use company land seemed risky. However, contacts between global headquarters and Indonesian counsel and dialogue with the US Embassy assuaged concerns. The modus operandi was to ‘‘trust, but verify’’ in a very business phenotype way.

Regional Commerce. The regional ofﬁce, too, relied on outside sources of information – namely, Indonesian contacts in Hong Kong and Chinese merchants who had extensive Indonesian dealings through their clan members who were resident in Indonesia. They would conﬁrm and verify the Essex reports. Information would then be tied to business activities in ways that supported local staff commerce activities.

Levels of Leadership (and Management) The land purchase activities involved a wide variety of players. Those primarily involved in these activities were the Essex employees located in Java (who included Javanese, Chinese, Americans, and Danes), the Essex attorney (Javanese), the sellers and their families (Javanese peasants), Essex’s partner (Chinese), the local lurah (Javanese), and the dukun (Javanese). A secondary cast of players was extensive and multinational: the Minister of Religion (a Javanese aristocrat), the Minister of Education (a Javanese aristocrat), the Director of Pharmaceuticals (Miningkabau), the Investment Board director (West Javanese), Essex’s Hong Kong business associate (Chinese), and Essex’s regional ofﬁce personnel (Egyptian, Swiss, and Swiss-Indochinese). The need was then to match leadership needs along the spectrum of activities.

Senior Levels. Within the local levels, the senior team of country manager, production manager, and human resources manager were aligned to specific extra organizational players. For example, the country manager was responsible for investment leadership, which required interactions with the Ministries of Investment, Religion, and Education. The external attorney assumed leadership for developing strategies for working with the village authorities (i.e., the lurah and the dukun). However, leadership of Essex’s Enviroscapes 409 interactions with these authorities fell to the human resources manager. Essex’s partner assumed leadership for gaining knowledge of the regional government activities and dealing with them. The precise leadership assignments were based on specific knowledge, responsibilities, and ethnicity and were aligned to non-organizational expectations. Finally, the strategic intent for each leader was to create shared social space that would support business activities – in other words, to provide inspirational leadership. After these relationships were solidified, the lurah and dukun played leadership roles in working with the peasants and providing key intelligence to the Essex team. The senior managers in Hong Kong took responsibility for required Chinese contacts and managing these to the advantage of Essex Indonesia.

Middle and Junior Levels. Local mid-level and junior-level leadership assumed responsibility for the execution of executed commerce strategies. For example, key junior staff provided support to the village leaders in ways that provided public acknowledgment of their senior roles in interchanges with Essex.

CONCLUSIONS

Organizations are situated in specific enviroscapes that influence the ability of individuals, their relationships, and their broader societies to remain stable or to affect paced changes. Organizational understanding requires scanning and understanding of all of these environments, if they are to be completely understood. These environments, which are both internal and externally related, operate within two phenotypes of each enviroscape: business and commerce. Each of these phenotypes is present in each enviroscape, and understanding the predominance of one over the other, or their interactions, provides a macro view of the organization. Multi-level research into geography and leadership provides theoretical frameworks for organizing enviroscape data. A next step would be to create several multi- level hypotheses that would demonstrate the power of each enviroscape and explore how these enviroscapes interact. In this regard, three lines of research become desirable. First, the business phenotype assumes a skeptical approach to activities and people, while commerce has an underlying assumption of openness and sharing. Successful organizations require an interaction of the two phenotypes in each of their enviroscapes. Specific larger contexts determined 410 RICHARD REEVES-ELLINGTON by organizational reach (global, regional, and local) will determine which enviroscape and phenotype is dominant for achieving success. Second, the two phenotypes require differing types of power: business asymmetrical and commerce more symmetrical. Power usage, then, requires mid-level, entry-level, and senior leaders to understand the differences and develop an enviroscape IQ for each environment and phenotype. Third, geographic locals have primary responsibility for leading and managing all enviroscapes. When leadership types and responsibilities are blurred or there are shifts among the three geographic levels, organizational confusion occurs. These lines of research could be pursued in future work via a paradigm for understanding organizational leadership realities, as presented in this chapter, through multi-level understanding of the organizational environments of climate, knowledge, ethnos, and time.

REFERENCES

Adam, B. (1995). Timewatch: The social analysis of time. Cambridge, UK: Polity Press. Adam, B. (1998). Timescapes of modernity. Oxford, UK: Routledge. Alexander, J. (1987). Trade, traders, and trading in rural Java. New York: Oxford Press. Anderson, B. (1972). The idea of power in Javanese culture. In: C. Holt (Ed.), Culture and politics in Indonesia (pp. 1–69). Ithaca, NY: Cornell University Press. Anderson, B. R. (1983). Imagined communities: Reﬂections on the origin and spread of nationalism. London: New Left Books. Antonakis, J., Cianciolo, A. T., & Sternberg, R. J. (2004). The nature of leadership. Thousand Oaks, CA: Sage. Antonakis, J., Schriesheim, C. A., Donovan, J. A., Gopalakrishna-Pillai, Pellegrini, E. K., & Rossomme, J. L. (2004). Methods for studying leadership. In: J. Antonakis, A. T. Cianciolo & R. J. Sternberg (Eds), The nature of leadership (pp. 48–70). Thousand Oaks, CA: Sage. Argyris, C. (1958). Some problems in conceptualizing organizational climate: A case study of a bank. Administrative Science Quarterly, 2, 501–520. Argyris, C., & Schoen, D. (1978). Organizational learning: A theory of action perspective. Reading, MA: Addison-Wesley. Argyris, C., & Schoen, D. (1996). Organizational learning II: Theory, method, and practice. Reading, MA: Addison-Wesley. Arvey, R. D., Zhang, Z., Avolio, B. J., & Kreuger, R. F. (2006). The determinants of leadership role occupancy: Genetic and personality factors. Leadership Quarterly, 17, 1–20. Ashkanasy, N. M., Wilderrom, C., & Peterson, M. F. (Eds). (2000). Handbook of organizational culture and climate. Thousand Oaks, CA: Sage. Avolio, B. J., Waldman, D. W., & Yammarino, F. J. (1991). Leading in the 1990s: Towards understanding the four I’s of transformational leadership. Journal of European Industrial Training, 15(4), 9–16. Enviroscapes 411

Bacharach, M., & Gambetta, D. (2001). Trust in signs. In: K. S. Cook (Ed.), Trust in society (pp. 148–184). New York: Russell Sage Foundation. Barber, B. (1983). The logic and limits of trust. New Brunswick, NJ: Rutgers University Press. Bass, B. M. (1985). Leadership beyond expectations. New York: Academic Press. Bass, B. M. (1990). Bass and Stogdill’s handbook of leadership: Theory, research, and managerial applications (3rd ed.). New York: Free Press. Bass, B. M. (1998). Transformational leadership: Industrial, military, and educational impact. Mahwah, NJ: Lawrence Erlbaum. Bass, B. M., & Avolio, B. J. (Eds). (1994). Improving organizational effectiveness through transformational leadership. Thousand Oaks, CA: Sage. Bennett, M. (1993). Developing intercultural sensitivity: An integrative approach to global and domestic diversity. In: R. M. Paige (Ed.), Education for the intercultural experience (pp. 21–71). Yarmouth, ME: Intercultural Press. Bennis, W. (2007). The challenges of leadership in the modern world. American Psychologist, 62, 2–5. Berger, P. I., & Luckmann, T. (1991). The social construction of reality: A treatise in the sociology of knowledge. Harnondsworth: Penguin. Bhattacharya, R., Devinney, T. M., & Pillutla, M. M. (1996). A formal model of trust based on outcomes. Academy of Management Review, 25, 459–472. Blackburn, S. (1998). Trust, cooperation, and human psychology. In: V. Braithwaite & M. Levi (Eds), Trust and governance (pp. 28–45). New York: Russell Sage Foundation. Bligh, W. (2001). Mutiny on the bounty. New York: Penguin Classics Series. Bluedorn, A. C. (2002). The human organization of time: Temporal realities and experience of time. Stanford, CA: Stanford Press. Bluedorn, A. C., & Denhardt, A. (1988). Time and organizations. Journal of Management, 14, 299–320. Bluedorn, A. C., Kaufman, C. F., & Lane, P. M. (1992). How many things do you like to do at once? An introduction to monochronic and polychronic time. Academy of Management Executive, 6(4), 17–26. Bodley, J. (1996). Anthropology and contemporary human problems. Mountain View, CA: Mayfield. Bontis, N., & Choo, C. W. (1998). The knowing organization: How organizations use information to construct meaning, create knowledge, and make decisions. Oxford, UK: Oxford University Press. Boym, S. (2001). The future of nostalgia. New York: Basic Books. Braithwaite, V., & Levi, M. (Eds). (1998). Trust and governance. New York: Russell Sage Foundation. Brown, D. J., & Lord, R. G. (2001). Leadership and perceiver cognition: Moving beyond first order constructs. In: M. London (Ed.), How people evaluate others in organizations (pp. 181–202). Mahwah, NJ: Lawrence Erlbaum. Burns, J. M. (1978). Leadership. New York: Harper Row. Burton, R. M., Lauridsen, J., & Obel, B. (2004). The impact of organizational climate and strategic fit on firm performance. Human Resource Management, 43(1), 67–82. Calhoun, C. (1991). Indirect relationships and imagined communities: Large scale social integration and transformation of everyday life. In: P. Bordieu & J. S. Coleman (Eds), Social theory for a changing society (pp. 95–120). Boulder, CO: Westwood Press. Campbell, J. (1973). The hero with a thousand faces. Princeton, NJ: Princeton University Press. 412 RICHARD REEVES-ELLINGTON

Carlile, P. R., & Christensen, C. M. (2005). Practice and malpractice in management research. Unpublished manuscript, Boston, MA. . Carroll, J. S., Rudolph, J. W., & Hatakenaka, S. (2005). Learning from organizational experience. In: M. Easterby-Smith & M. A. Lyles (Eds), The Blackwell handbook of organizational learning and knowledge management (pp. 575–600). Cambridge, UK: Blackwell. Chemers, M. M. (2000). Leadership research and theory: A functional integration. Group Dynamics, 4, 37–43. Chesher, J. E. (1992). Business: Myth and morality. In: R. W. McGee (Ed.), Business ethics and common sense (pp. 45–107). Westport, CT: Quorum. Choo, C. W. (2002). Sensemaking, knowledge creation, and decision making: Organizational knowing as emergent strategy. In: C. W. Choo (Ed.), Strategic management of intellectual capital and organizational knowledge (pp. 79–85). Oxford, UK: Oxford University Press. Clark, P. A. (2000). Organizations in action: Competition between contexts. London: Routledge. Clegg, S. R. (1992). Frameworks of power. Thousand Oaks, CA: Sage. Cohen, M., & Bacdayan, P. (1994). Organizational routines are stored as procedural memory: Evidence from a laboratory study. Organizational Science, 5(4), 554–568. Conger, J. A., & Kanungo, R. N. (1987). Toward a behavioral theory of charismatic leadership in organizational settings. Academy of Management Review, 12, 637–647. Dansereau, F., Graen, G. B., & Haga, W. J. (1975). A vertical dyad linkage approach to leadership within formal organizations: A longitudinal investigation of the role making process. Organizational Behavior and Human Decision Processes, 13, 46–78. Das, T. K., & Teng, B. S. (1998). Between trust and control: Developing confidence in partner cooperation in alliances. Academy of Management Review, 23, 491–512. Dasgupta, P. (1988). Trust as a commodity. In: D. Gambetta (Ed.), Trust making and breaking cooperative relations (pp. 129–148). Oxford, UK: Blackwell. Day, D. V., & Lord, R. G. (1988). Executive leadership and organizational performance. Journal of Management, 14, 453–464. Denison, D. R. (1990). What is the difference between organizational culture and organizational climate? A native’s point of view on a decade of paradigm wars. Academy of Management Journal, 21, 619–654. Dierkes, M., Anatal, A., Child, J., & Nonaka, I. (Eds). (2003). Handbook of organizational learning and knowledge. Oxford, UK: Oxford Press. DiMaggio, P. J., & Powell, W. W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in institutional fields. American Sociological Review, 48, 147–160. Doney, P. M., Cannon, J. P., & Mullen, M. R. (1998). Understanding the influence of national culture on the development of trust. Academy of Management Review, 23(3), 601–620. Drexler, J. A. (1977). Organizational climate: Its homogeneity within organizations. Journal of Applied Psychology, 62, 38–42. Drucker, P. F. (1955). The practice of management. London: William Heinemann. Easterby-Smith, M., & Lyles, M. A. (Eds). (2005). The Blackwell handbook of organizational learning and knowledge management. Oxford, UK: Blackwell Press. Eden, D., & Leviatan, U. (1975). Implicit leadership theory as a determinant of the structure underlying supervisory behavior scales. Journal of Applied Psychology, 60, 736–741. Ekvall, G. (1987). The climate metaphor in organizational theory. In: B. Bass & P. Drenth (Eds), Advances in organizational psychology (pp. 177–190). Beverly Hills, CA: Sage. Enviroscapes 413

Elide, M. (1964). Shamanism: Archaic techniques of ecstasy. Princeton, NJ: University of Princeton Press. Elkjaer, B. (2005). Social learning theory: Learning in participation in social processes. In: M. Easterby-Smith & M. A. Lyles (Eds), The Blackwell handbook of organizational learning and knowledge management (pp. 38–53). Cambridge, UK: Blackwell. Etzioni, A. (1964). Modern organizations. Englewood Cliffs, NJ: Prentice Hall. Fatsis, S. (1992, December 28). An original at corporate Xerox. The Daily Messenger, Canandaigua, NY, p. 7A. Fiedler, F. E. (1964). A theory of leadership effectiveness. New York: McGraw Hill. Fiedler, F. E. (1971). Leadership. Morristown, NJ: General Learning Press. Fine, C. H. (1998). Clockspeed: Winning industry control in the age of temporary advantage. New York: Perseus. Fiol, C. M., & Lyles, M. A. (1985). Organizational learning. Academy of Management Review, 10, 803–813. Flanherty, M. (1999). A watched pot: How we experience time. New York: New York University Press. Flieshman, E. A. (1953). Leadership climate, human relations training, and supervisory behaviour. Personnel Psychology, 6, 205–255. Frank, J. (1985). Therapeutic components shared by all psychotherapies. In: M. Mahoney & A. Freeman (Eds), Cognition and psychotherapy (pp. 49–79). New York: Plenum. Fraser, J. T. (1987). Time, the familiar stranger. Redmond, WA: Tempus. French, J. R. P., & Raven, B. (1968). The bases of social power. In: D. Cartwright & A. Zander (Eds), Group dynamics: Research and theory (3rd ed., pp. 259–269). New York: Harper Row. Gabarro, J. J. (1978). The development of trust, inﬂuence, and expectations. In: A. Athos & K. Gabarro (Eds), Interpersonal behavior: Communications and understanding in relationships (pp. 128–150). Englewood Cliffs, NJ: Prentice-Hall. Gambetta, D. (1988). Trust making and breaking cooperative relations. New York: Basil Blackwell. Gardner, H. (1999). The disciplined mind. New York: Simon & Schuster. Garsten, C. (1994). Apple world: Core and periphery in a transnational organizational culture. Stockholm: Stockholm Studies in Social Anthropology. Geertz, H. (1961). The Javanese family: A study of kinship and socialization. Prospect Heights, IL: Waveland. Gell, A. (1992). The anthropology of time. Oxford, UK: Berg. Giddens, J. J. (1990). The consequences of modernity. Stanford, CA: Stanford University Press. Gioia, D., & Pitre, E. (1990). Multiparadigm perspectives on theory building. Academy of Management Review, 15, 584–603. Goshal, S., & Bartlett, C. A. (1990). The multinational corporation as an interorganizational network. Academy of Management Review, 15(4), 603–623. Govier, T. (1997). Social trust and human communities. Montreal: McGill-Queens University Press. Guion, R. M. (1973). A note on organizational climate. Organizational Behavior and Human Performance, 9, 120–125. Hagen, J. M., & Choe, S. (1998). Trust in Japanese interﬁrm relations: Institutional sanctions matter. Academy of Management Review, 23, 589–600. Halifax, J. (1978). Shamanic voices. New York: Anchor Press. Hall, E. T., & Hall, R. H. (1987). Hidden differences: Doing business with the Japanese. New York: Anchor. 414 RICHARD REEVES-ELLINGTON

Halliday, D. (1983). Steve Paul Jobs. Current Biography, 204–207. Hamal, G., & Prahalad, C. K. (1989). Strategic intent. Harvard Business Review, 67(3), 63–76. Hamayon, R. N. (1994). Shamanism in Siberia: From partnership in supernature to counter- power in society. In: N. Thomas & C. Humphrey (Eds), Shamanism, history, and the state (pp. 76–89). Ann Arbor, MI: University of Michigan Press. Hanna, D. P. (1988). Designing organizations for high performance. New York: Addison Wesley. Hardin, R. (1998). Trust in government. In: V. Braithwaite & M. Levi (Eds), Trust and governance (pp. 9–26). New York: Russell Sage Foundation. Hardin, R. (2001). Conceptions and explanations of trust. In: K. S. Cook (Ed.), Trust in society (pp. 3–39). New York: Russell Sage Foundation. Harmon, M. M. (1995). Responsibility as paradox. Thousand Oaks, CA: Sage. Hedberg, B., & Wolff, R. (2005). Organizing, learning, and strategizing: From construction to discover. In: M. Dirkes, A. B. Antal, J. Child & I. Nonake (Eds), Handbook of organizational learning and knowledge (pp. 535–556). New York: Oxford Press. Heinze, R. I. (1991). Shamans of the 20th century. New York: Irvington. Henderson, R., & Cockburn, I. (1994). Measuring competence: Exploring ﬁrm-effects in pharmaceutical research. Strategic Management Journal, 15(Winter Special Issue), 63–84. Herzberg, F. (1968, January–February). One more time: How do you motivate employees? Harvard Business Review, pp. 53–62. Hirschman, A. O. (1970). Exit, voice and loyalty. Cambridge, MA: Harvard University Press. Hogan, R., & Kaiser, R. (2005). What we know about leadership. Review of General Psychology, 9, 169–180. House, J. R., Spangler, W. D., & Woycke, J. (1991). Personality and charisma in the US presidency: A psychological theory of leader effectiveness. Administrative Science Quarterly, 36, 364–396. House, R., Rousseau, D. M., & Thomas-Hunt, M. (1995). The meso paradigm: A framework for integration of micro and macro organizational. In: L. L. Cummings & B. Straw (Eds), Research in organizational behavior (Vol. 17, pp. 71–114). Greenwich, CT: JAI Press. Humphrey, C. (1997). Shamanistic practices and the state in Northern Asia: Views from the center and periphery. In: N. Thomas & C. Humphrey (Eds), Shamanism, history, and the state. Ann Arbor, MI: University of Michigan Press. Hunt, J. G. (1991). Leadership: A new synthesis. Newbury Park, CA: Sage. Hunt, J. G. (2004). What is leadership. In: J. Antonakis, A. T. Cianciolo & R. J. Sternberg (Eds), The nature of leadership (pp. 17–47). Thousand Oaks, CA: Sage. Hunt, J. G., & Larson, L. L. (Eds). (1973). Leadership frontiers. Kent, OH: Kent State University Press. James, L. R., Choi, C. C., Ko, C. H. E., et al. (2007). Organizational and psychological climate: A review of theory and research. European Journal of Work and Organizational Psychology, 3, 230–252. James, L. R., & Jones, A. P. (1974). Organizational climate: A review of theory and research. Psychological Bulletin, 81(12), 1096–1112. Joyce, W. F., & Slocum, J. (1984). Collective climate: Agreement as a basis for deﬁning aggregate climate in organizations. Academy of Management Journal, 27, 721–742. Kanungo, R. N., & Mendonca, M. (1996). Ethical dimensions of leadership. Thousand Oaks, CA: Sage. Enviroscapes 415

Katz, D., Maccoby, N., Gurin, G., & Floor, L. G. (1951). Productivity, supervision and morale among railroad workers. Ann Arbor, MI: Survey Research Center, Institute for Social Research, University of Michigan. Kern, S. (1983). The culture of time and space. Cambridge, MA: Harvard University Press. Klein, K. J., Dansereau, F., & Hall, R. J. (1995). Levels issues in theory development, data collection, and analysis. Academy of Management Review, 19, 125–229. Klein, K. J., & Kozlowski, S. W. (Eds). (2000). Multilevel theory, research, and methods in organizations. San Francisco, CA: Jossey-Bass. Klimo, J. (1988). The emergence of a new paradigm. In: R. I. Heinze (Ed.), Proceedings of the 5th international conference on the study of shamanism and alternate modes of healing (pp. 394–402). Berkeley, CA: Independent Scholars of Asia. Kluckhohn, F. R., & Strodtbeck, F. L. (1961). Variations in value orientations. Westport, CT: Greenwood. Kosselleck, R. (1985). Futures past. (K. Trith, Trans.). Cambridge, MA: MIT Press. Koyes, D., & Decotiis, T. (1991). Inductive measures of psychological climate. Human Relations, 44, 265–285. Krackhardt, D., & Kilduff, M. (1990). Friendship patterns and culture: The control of organizational diversity. American Anthropologist, 92, 142–154. Krippner, S. (1988). The use of dreams by tribal shamans. In: R. I. Heinze (Ed.), Proceedings of the 5th international conference on the study of shamanism and alternate modes of healing (pp. 22–64). Berkeley, CA: Independent Scholars of Asia. Kunda, K. (2007). Development of the conceptual framework of organizational climate. Vidyasagar University Journal of Commerce, 12, 99–108. Lane, C. (2000). Introduction: Theories and issues in the study of trust. In: C. Lane & R. Bachmann (Eds), Trust within and between organizations: Conceptual issues and empirical applications (pp. 1–30). Oxford, UK: Oxford University Press. Langer, E. J. (1989). Mindfulness. Reading, MA: Merloyd Lawrence. Langer, E. J., & Piper, A. I. (1987). The prevention of mindlessness. Journal of Personality and Social Psychology, 53(2), 275–287. Leblebici, H., Salancik, G. R., Copay, A., & King, T. (1991). Institutional change and the transformation of interorganizational fields: An organizational history of the US radio broadcasting industry. Administrative Science Quarterly, 36, 333–363. Lee, D. (1963). Freedom and social constraint. In: D. Bidney (Ed.), The concept of freedom in anthropology (pp. 1–19). The Hague: Mouton. Le Vie, D., Jr. (1988). The ‘‘significant experience’’ as a mnemonic device in shamanic symbolism. In: R. I. Heinze (Ed.), Proceedings of the 5th international conference on the study of shamanism and alternate modes of healing (pp. 161–167). Berkeley, CA: Independent Scholars of Asia. Levitt, T. (1958). The dangers of social responsibility. Harvard Business Review, 36(5), 41–50. Lewicki, J. R., & Bunker, B. B. (1995). Trust in relationships: A model of development and decline. In:B.B.Bunker&J.Z.Rubin(Eds),Conflict, cooperation, and justice (pp. 113–173). San Francisco, CA: Jossey-Bass. Lewicki, J. R., McAllister, D. J., & Bies, R. J. (1998). Trust and distrust: New relationships and realities. Academy of Management Review, 23, 438–453. Lewin, K., Lippit, R., & White, R. K. (1939). Patterns of aggressive behavior in experimentally created ‘‘social climate’’. Journal of Social Psychology, 10, 271–290. 416 RICHARD REEVES-ELLINGTON

Lindahl, R. (2006). The role of organizational climate and culture in the school improvement process: A review of the knowledge base [electronic version]. National Council of Professors of Educational Administration. Retrieved January 30, 2008, from http:// cnx.org/content/m13465/latest/ Lindskold, S. (1978). Trust development, the FRIT proposal and the effects of conciliatory acts on conﬂict and cooperation. Psychological Bulletin, 85, 772–793. Litwin, G. H., & Stringer, R. A. (1968). Motivation and organizational climate. Cambridge, MA: Harvard Business School, Division of Research. Lord, R. G., Foti, R. J., & Philips, J. S. (1982). A theory of leadership categorization: Internal structure, information processing and leadership perceptions. Organizational Behavior and Human Performance, 34, 343–378. Lowe, K. B., & Gardner, W. L. (2000). Ten years of The Leadership Quarterly: Contributions and challenges for the future. Leadership Quarterly, 11, 459–514. Luhmann, N. (1979). Trust and power. Chichester, UK: Wiley. Machiavelli, N. (1990). The prince. Oxford, UK: Oxford University Press. Mackenzie, K. D. (2004). The process approach to multi-level organizational behavior. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and processes (Vol. 3, pp. 347–415). Boston, MA: Elsevier. Major, D. A. (2000). Effective newcomer socialization. In: N. M. Ashkansay, C. P. M. Wilderrom & M. F. Peterson (Eds), Handbook of organizational culture and climate (pp. 355–368). Thousand Oaks, CA: Sage. Mann, R. D. (1959). A review of the relationship between personality and performance in small groups. Psychological Bulletin, 56, 241–270. Mansbridge, J. (1999). Altruistic trust. In: M. Warren (Ed.), Democracy and trust (pp. 101–140). Cambridge, UK: Cambridge University Press. Marshall, L. (1997, March). The world according to Eco. Wired, pp. 144–196. McGregor, D. M. (1960). The human side of enterprise. New York: McGraw-Hill. McKay, J. T. (1968). The management of time. Englewood Cliffs, NJ: Prentice-Hall. Meyerson, D., Weick, K. E., & Kramer, R. M. (1996). Swift trust and temporary groups. In: R. M. Kramer & T. R. Tyler (Eds), Trust in organizations: Frontiers of theory and research (pp. 166–195). Thousand Oaks, CA: Sage. Miles, R. E., & Snow, C. C. (1992, Summer). Causes of failure in network organizations. California Management Review, 72–93. Minkowski, E. (1970). Lived time: Phenomenological and psychopathological studies. (N. Metzel, Trans.). Evanston, IL: Northwestern University Press. Mintzberg, H. (1979). The structuring of organizations. Englewood Cliffs, NJ: Prentice-Hall. Mintzberg, H. (1983). Power in and around organizations. Englewood Cliffs, NJ: Prentice-Hall. Mintzberg, H., Raisinghani, D., & Theoret, A. (1976). The structure of ‘‘unstructured’’ decision processes. Administrative Science Quarterly, 21, 246–261. Moran, E. T. (1992). The cultural approach to the formation of organizational climate. Human Relations, 45(1), 19–47. Nonaka, I., Toyama, R., & Byosiere, P. (2001). A theory of organizational knowledge creation: Understanding the dynamic process of creating knowledge. In: M. Dierkes, A. Anatal, J. Child & I. Nonaka (Eds), Handbook of organizational learning and knowledge (pp. 490–517). Oxford, UK: Oxford Press. Nye, J. S. J. (2008). The powers to lead. New York: Oxford. Enviroscapes 417

O’Connor, M. (1994). Introduction: Liberate, accumulate – and bust? In: M. O’Connor (Ed.), Is capitalism sustainable? (pp. 1–21). New York: Guilford. Ohmann, O. A. (1989). Skyhooks. In: K. R. Andrews (Ed.), Ethics in practice: Managing the moral corporations (pp. 58–69). Boston, MA: Harvard Business School Press. Oinas-Kukkonen, H. (2008). The 7C model for organizational knowledge creation and management [electronic version]. Retrieved February 26, 2008, from http://oasis.oulu.ﬁ/ publications/OKLC-04-hok.pdf Owens, R. G. (2004). Organizational behavior in education: Adaptive leadership and school reform (8th ed.). Boston, MA: Allyn & Bacon. Pattee, R. (1989). Neo-shamanism: A source of creativity for our time. In: R. I. Heinze (Ed.), Proceedings of the 5th international conference on the study of shamanism and alternate modes of healing (pp. 150–166). Berkeley, CA: Independent Scholars of Asia. Pawlowsky, P. (2001). The treatment of organizational learning in management science. In: M. Dierkes, A. Anatal, J. Child & I. Nonaka (Eds), Handbook of organizational learning and knowledge (pp. 61–88). Oxford, UK: Oxford Press. Payne, R. L. (2001). Climate and culture: How close can they get? In: N. M. Ashkanasy, C. Wilderrom & M. F. Peterson (Eds), Handbook of organizational culture and climate (pp. 163–176). Thousand Oaks, CA: Sage. Pfeffer, J. (1981). Management as symbolic action: The creation and maintenance of organizational paradigms. In: L. L. Cummings & B. N. Staw (Eds), Research in organizational behavior (Vol. 3, pp. 1–52). Greenwich, CT: JAI Press. Powell, G. N., & Butterﬁeld, D. A. (1978). The case for subsystem climates in organizations. Academy of Management Review, 3, 151–156. Reason, P. (1994). Three approaches to participative inquiry. In: N. K. Denzin & Y. A. Lincoln (Eds), Handbook of qualitative research (pp. 324–339). Thousand Oaks, CA: Sage. Reeves-Ellington, R. (1976). Field notes, P. T. Essex. Reeves-Ellington, R. (1980, January/February). Pitfalls of payoffs in Indonesia. Asia, pp. 6–9. Reeves-Ellington, R. (1995a). Anthropology and total quality management: Improving sales force performance in overseas markets. In: J. F. J. Sherry (Ed.), Contemporary marketing and consumer behavior (pp. 169–208). Thousand Oaks, CA: Sage. Reeves-Ellington, R. (1995b). Organizing for global effectiveness: Ethnicity and organizations. Human Organization, 54(3), 249–262. Reeves-Ellington, R. H. (1997). Cross cultural organizational stress: The American University in Bulgaria. Human Organization, 57(1), 94–107. Reeves-Ellington, R. (1999a). Altered states of leadership responsibility. Leadership and Organizational Development Journal, 19(2), 97–105. Reeves-Ellington, R. H. (1999b). Local value orientations and global customer chains: A Bulgarian case study. South East Europe Review for Labour and Social Affairs, 2(4), 105–122. Reeves-Ellington, R. (2004a). Cultural bases of research: Integrating research cultures and complex problems. In: F. J. Yammarino & F. Dansereau (Eds), Research in multi-level issues (Vol. 3, pp. 167–177). Boston, MA: Elsevier. Reeves-Ellington, R. (2004b). Trustscapes and distrustscapes: A multi-level approach for understanding stability and change. In: F. J. Yammarino & F. Dansereau (Eds), Multi- level issues in organizational behavior and processes (Vol. 3, pp. 91–148). Boston, MA: Elsevier. 418 RICHARD REEVES-ELLINGTON

Reeves-Ellington, R. (2007a). Let us theory build rather than theorize. In: F. Dansereau & F. J. Yammarino (Eds), Research in multi-level issues (Vol. 6, pp. 339–349). Boston, MA: Elsevier. Reeves-Ellington, R. (2007b). Timescapes: A multi-level approach for understanding time use in complex organizations. In: F. Dansereau & F. J. Yammarino (Eds), Multi-level issues in organizations and time (Vol. 6, pp. 259–317). Boston, MA: Elsevier. Reeves-Ellington, R., & Anderson, A. (1997a). Business, commerce, and social responsibility. Lewiston, NY: Mellon Press. Reeves-Ellington, R., & Anderson, A. (1997b). The ethnology of information systems. Accounting, Management, and Information Technologies, 7(3), 168–180. Rescher, N. (1996). Process metaphysics: An introduction to process philosophy. New York: University of New York Press. Robbins, S. P. (1983). Organizational theory: The structure and design of organizations. Englewood Cliffs, NJ: Prentice-Hall. Rousseau, D. M., Sitkin, S. B., Burt, R. S., & Camerer, C. (1998). Introduction to special issue on trust in organizations. Academy of Management Review, 23, 387–392. Sackman, S. (1992). Culture and subcultures: An analysis of organizational knowledge. Administrative Science Quarterly, 37, 140–161. Sandys, C. (1965). We shall not fail: The inspiring leadership of Winston Churchill. London: Oxford. Sawy, O. A. E., Eriksson, I., Raven, A., & Carlsson, S. (2001). Understanding shared knowledge creation spaces around business processes: Precursors to process innovation implementation. International Journal of Technology Management, 22(2), 149–173. Schein, E. H. (1987). Organizational culture and leadership. San Francisco, CA: Jossey-Bass. Schein, E. H. (2001). Sense and nonsense about culture and climate. In: N. M. Ashkanasy, C. Wilderrom & M. F. Peterson (Eds), Handbook of organizational culture and climate (pp. xxiii–xxx). Thousand Oaks, CA: Sage. Schiller, L. J. (1997, August). An old coach with new discipline. Business Week, p. 51. Schisgall, O. (1981). Eyes on tomorrow. New York: Doubleday. Schneider, B. (1985). Organizational behavior. Annual Review of Psychology, 36, 447–611. Schneider, B., & Bartlett, J. (1968). Individual differences and organizational climate I: The research plan and questionnaire development. Personnel Psychology, 21, 323–333. Schneider, B., & Reichers, A. (1983). On the etiology of climates. Personal Psychology, 36, 19–39. Scott, J. (1991). Networks of corporate power: A comparative assessment. Annual Review of Sociology, 17, 181–203. Senge, P. M. (1990). The ﬁfth discipline: The art and practice of the learning organization. New York: Doubleday Currency. Sherry, J. F. J., McGrath, M. A., & Levy, S. J. (1995). Monadic giving: Anatomy of gifts given to the self. In: J. F. J. Sherry (Ed.), Contemporary marketing and consumer behavior: An anthropological sourcebook (pp. 399–432). Thousand Oaks, CA: Sage. Sinha, J. B. P. (1995). The cultural context of leadership and power. New Delhi: Sage. Smith, A. (2007). The wealth of nations. Retrieved February 4, 2008, from http:// www.bibliomania.com/2/1/65/112/frameset.html Smith, P. B., & Peterson, M. F. (1988). Leadership, organizations, and culture. London: Sage. Sober, E., & Wilson, D. S. (1999). Unto others: The evolution and psychology of unselﬁsh behavior. Cambridge, MA: Harvard University Press. Enviroscapes 419

Stogdill, R. M. (1948). Personal factors associated with leadership: A survey of the literature. Journal of Psychology, 25, 35–71. Stopford, J. (2001). Organizational learnings as guided responses to market signals. In: M. Dierkes, A. Anatal, J. Child & I. Nonaka (Eds), Handbook of organizational learning and knowledge (pp. 264–281). Oxford, UK: Oxford Press. Swasy, A. (1993). Soap opera. New York: Random House. Tagiuri, R., & Litwin, G. H. (1968). Organizational climate: Explorations of a concept. Cambridge, MA: Harvard School of Business Administration, Division of Research. Taylor, S., & Osland, J. S. (2005). The impact of intercultural communication of global organizational learning. In: M. Easterby-Smith & M. A. Lyles (Eds), The Blackwell handbook of organizational learning and knowledge management (pp. 212–232). Cambridge, UK: Blackwell. Thomas, N., & Humphrey, C. (1994). Shamanism, history, and the state. Ann Arbor, MI: University of Michigan Press. Ting-Toomey, S. (1999). Communicating across cultures. New York: Guilford. Tyler, T. R. (2001). Why do people rely on others? Social identity and the social aspects of trust. In: K. S. Cook (Ed.), Trust in society (pp. 285–306). New York: Russell Sage Foundation. Uhl-Bien, M., Graen, G. B., & Scandura, T. (2000). Implications of leader–member change (LXM) for strategic human resource management systems: Relationships as social capital for competitive advantage. In: G. Ferris (Ed.), Research in personnel and human resource management (Vol. 128, pp. 137–185). Greenwich, CT: JAI Press. Van Vugt, M. (2006). Evolutionary origins of leadership and followership. Personality and Social Psychology Review, 10, 354–371. Van Vugt, M., Hogan, R., & Kaiser, R. (2008). Leadership, followership, and evolution. Unpublished manuscript. Vera, D., & Crossan, M. (2005). Organizational learning and knowledge management: Toward an integrative framework. In: M. Easterby-Smith & M. A. Lyles (Eds), The Blackwell handbook of organizational learning and knowledge (pp. 123–141). Cambridge, UK: Blackwell. Waller, M. J., Conte, J. M., Gibson, C. B., & Carpener, M. A. (2000). The effect of individual perceptions of deadlines on team performance. Academy of Management Review, 26(4), 586–600. Walsh, R. (1990). The spirit of shamanism. Boston, MA: Shambhala. Walsh, R. (1994). The making of a shaman calling, training, and culmination. Journal of Humanistic Psychology, 34(3), 7–30. Webster’s New Universal Unabridged Dictionary (2nd ed.). (1983). New York: Simon and Schuster. Weedon, C. (1987). Feminist practice and poststructuralist theory. Oxford: Blackwell. Weick, K. E. (1995). Sensemaking in organizations. Thousand Oaks, CA: Sage. Wiley, J. W. (1998). Using survey results to predict dealership satisfaction and operational performance. Paper presented at the 13th Annual Meeting of the Society for Industrial and Organizational Psychology, August 13–15, Seattle, WA. Winkelman, M. J. (1982). Magic: A theoretical reassessment. Current Anthropology, 23(37), 59–66. Woodman, R. W., & King, D. C. (1978). Organizational climate: Science or folklore? Academy of Management Review, 3, 816–826. 420 RICHARD REEVES-ELLINGTON

Xerox Corporation. (1993, May). The Xerox quest for quality. Staff presentations at the annual conference of the New York State Association of Rehabilitation Facilities, Albany, NY. Yamagishi, T. (2001). Trust as a form of social intelligence. In: K. S. Cook (Ed.), Trust in society (pp. 121–147). New York: Russell Sage Foundation. Yukl, G. A. (Ed.) (2006). Leadership in organizations (6th ed.). Englewood, NJ: Prentice Hall. Zaccaro, S. J., Foti, R. J., & Kenny, D. A. (1991). Self-monitoring and trait-based variance in leadership: An investigation of leader ﬂexibility across multiple group situations. Journal of Applied Psychology, 76, 308–315. Zack, M. (1999). Developing a knowledge strategy. California Management Review, 41, 125–145. Zerubavel, E. (1981). Hidden rhythms: Schedules and calendars in social life. Chicago, IL: University of Chicago Press.

TARGETING THE CULTURAL PROCESSES OF PARTNERING FOR ANALYSIS

Elizabeth K. Briody

ABSTRACT

Describing and explaining the interface between organizational culture and community culture necessitate an exploration into assumptions, expectations, beliefs, symbolism, and behaviors. This commentary examines the successful integration of an expanding US pharmaceutical firm into Indonesia’s multicultural environment, a context marked by the interweaving of market exchange and reciprocity exchange. It directs attention to the interactions occurring among key leaders within the firm, and between those in the firm and those in the peasant and governmental communities. By focusing on the cultural processes of partnering, the contribution of cooperative, healthy relationships in achieving the firm’s business goals is revealed.

INTRODUCTION

The chapter in this volume linking organizational leadership to ‘‘enviroscapes’’ contains a very interesting and insightful discussion of cross-cultural

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 421–429 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008016 421 422 ELIZABETH K. BRIODY interactions from the perspective of Essex, a subsidiary of the Schering Corporation. The author, Richard Reeves-Ellington, has been a colleague and friend of mine for many years; we are both members of the National Association for the Practice of Anthropology, a section of the American Anthropological Association. Richard’s extraordinary experiences in Indonesia have always intrigued me, and I am delighted to have been given this opportunity to explore some of his work in detail. This commentary is divided into three sections. First, it identiﬁes some of the methodological techniques and approaches that Reeves-Ellington used in Indonesia that help explain Essex’s success in Indonesia. Next, it examines the contrast he proposes between the concepts of business and commerce. Finally, it offers a perspective on the ethnographic data that focuses attention on the cultural processes at work in the evolution of this success story.

EXPLAINING THE ESSEX SUCCESS STORY

It is appropriate that this chapter emphasizing the role of leadership (with an occasional example about particular leaders) was written by a leader – and a leader in several respects. Richard Reeves-Ellington was very obviously a leader within Essex, rising to a critical position within the firm. He played an important interface role with the various local cultures (e.g., Javanese, Chinese). He served as a cultural broker and mediator back to the Essex home office from his positions in Indonesia. Moreover, his career opportunities in Asia have positioned him well to act as a translator for academic and business communities as he describes the multicultural Indonesian context from a leadership standpoint. The ethnographic material presented as case scenarios is fascinating. An understanding of, and learning associated with, the various Asian cultures paves the way for Essex’s overall success in Indonesia. As evident in Reeves- Ellington’s chapter, cultural sensitivity matters and affects business outcomes (Reeves-Ellington, 2009). The author’s discussion of the communication chain linking Essex with the Minister of Religion in Jakarta illustrates the importance of knowing to use and then activating social networks to further business prospects. Listening to the lurah (village head), who represented the peasant families, enabled Essex to choose a highly desirable and appropriate exchange for the sale of their land (i.e., Hajdi tickets). Problem solving with the lurah and dukun (religious leader) to deal with the impending loss of spiritual bonds between the local population and their land symbolized Essex’s willingness to ‘‘go the extra mile’’ so that the Targeting the Cultural Processes of Partnering for Analysis 423 sellers were content. Essex’s clear intent was to satisfy the sellers on the sellers’ terms; collaborating with the local community was reinforced in other examples as well (e.g., the land-sale process). All of these strategies reflected Essex’s high degree of cultural competency and leadership. Another important aspect of the case material is the emphasis on the leadership role and perspective that is derived from the author’s overall methodology. Reeves-Ellington describes Essex business professionals using a variety of techniques at the global, regional, and local levels of leadership. His account is based on his own direct experience as an Essex executive, his discussions with other Essex leaders, and his interactions with Indonesian locals and others. His methodology informs his knowledge base, his interactions, and his decisions – all of which contributed to Essex’s overall success. Moreover, Reeves-Ellington’s account is longitudinal with insights and experiences spanning a number of years. Using a longitudinal approach in combination with ethnography – including on-site experience as a participant – represents a powerful set of mechanisms for data gathering. But there is another critical element at work here: A longitudinal approach positions an organization for success in ways that a short-term approach does not. When organizational leaders make a long-term commitment to a goal, a region, and/or a community, they are in it to make things work, to cultivate relationships, and to learn. The methodological approach taken by Reeves-Ellington and aligned with Essex’s own orientation to expansion was internally consistent and complementary. As a result, Reeves-Ellington presents a clear, concise description of Essex’s leaders and leadership associated with those three levels, their interactions with other leaders and constituencies, and the ways in which they were perceived by others – including members of the local Indonesian communities. An orientation toward a culturally sensitive globalization emerges at all Essex leadership levels in Indonesia, as Essex seeks to engage in work practices that will be acceptable both to the home office and to the local communities. Reeves-Ellington’s actions indicate his strong interest in arriving at a set of agreements that each party considers successful.

EXAMINING THE PHENOTYPES OF BUSINESS AND COMMERCE

Reeves-Ellington has proposed a contrast between the notions of ‘‘business’’ and ‘‘commerce.’’ His intent seems to be to identify the differences in cultural expectations, assumptions, beliefs, processes, practices, and 424 ELIZABETH K. BRIODY behaviors linked to these notions in the literature and supported by the case material. He describes business and commerce as ‘‘phenotypes’’ for several different organizational environments (i.e., enviroscapes), setting each up in opposition to the other without the potential for overlap. I prefer the phrase ‘‘ideal types’’ in Weber’s usage (i.e., one type is understood in relation to another), as Reeves-Ellington seems to be suggesting a generalized conceptualization linked with comparative analysis (Gerth & Mills, 1946, pp. 59–61). Weber’s concept emphasizes particular dimensions of the type rather than connotations that are ‘‘exemplary’’ or indicative of ‘‘an ideal way of life’’ (Gerth & Mills, 1946, p. 59). Three aspects of the business and commerce discussion are especially surprising. First, I wondered about the rationale for selecting these two concepts to compare and contrast rather than any others. I do not think this contrast was based on ‘‘emic’’ (i.e., local, native) categories; the description of the high-context Indonesian world suggests a greater degree of integration or holism with respect to these two concepts. Second, I wondered why the core elements of the concepts were not referenced explicitly in the labeling; instead, differentiation is left to the attributes associated with the concepts. For example, an alternative labeling of the two types such as ‘‘business–organization culture’’ and ‘‘peasant– community culture’’ might have offered a more robust and accurate depiction of the distinction Reeves-Ellington was trying to draw. Perhaps most importantly, I wondered about the particular attributes Reeves-Ellington chose to link with business as well as the attributes he identified with commerce. The phenotype labeling suggests that business and commerce are mutually exclusive when, in fact, the attributes can and do cross-cut both business and commerce. Tables 2–4 in the Reeves-Ellington chapter are particularly good examples of the fluidity between the two types. (Because these tables are included in the literature framing of the chapter, I assume that they apply to the ethnographic material.) To suggest or even imply that business is characterized as ‘‘manipulative’’ and ‘‘opportunistic’’ and that commerce is not (and, in fact, is characterized as vulnerable and cooperative) is buying into a stereotype that Reeves-Ellington’s portrayal of the ethnographic data does not support. Each type shares characteristics with the other in varying degrees, at particular times, and under certain conditions. While there were attempts to ‘‘exhibit efficiency in achieving stated outcomes’’ by the firm (e.g., bank transfer of funds, payment by check), it was Essex’s view of efficiency that then clashed with the peasant community’s view of efficiency (e.g., having the 135,000 paper bills counted in the presence of kin). Efficiency mattered in both cultural contexts. Targeting the Cultural Processes of Partnering for Analysis 425

As to ‘‘relationship effectiveness’’ (the presumed contrast with efficiency), the ethnographic material indicates that relationship effectiveness flourished within Essex, between Essex and other Schering organizations, and between Essex and the local community. Within Essex, Conzen, the head of the home office, certainly understood the importance of relationships given his background in sales, marketing, and distribution, and his expectation that local Essex leaders implement a payment system ‘‘that would be socially acceptable to the hosts [within the Indonesian environment].’’ He reportedly ‘‘expected his team to create much the same community feeling in areas where they worked.’’ Conzen also ‘‘hired in his likeness,’’ including Weiser, the regional vice president for Essex. The ‘‘returning war stories’’ offered by mid- level Essex management ‘‘created a climate of trust between Essex and other Schering organizations.’’ Moreover, social events were arranged for Essex managers visiting the United States (e.g., the US counterpart of an Indonesian manager hosted that manager at a ‘‘Christmas pageant’’). Finally, relationships were cultivated between Essex’s junior-level leaders and members of the local community. One example cited was the processing of paperwork from the port authorities. Apparently, an Essex clerk ‘‘learned some important social information’’ about the family of a particular Javanese official, ‘‘provided a small gift,’’ and soon found that the paperwork was ‘‘promptly processed.’’ The chapter is replete with examples of building and maintaining relationships as the foundation for Essex’s organizational culture. ‘‘Effective human relationships’’ is present not only in what is described in the notion of commerce, but also in the notion of business. Relationship effectiveness necessitates ‘‘trust’’ (Sengir et al., 2004). The ethnographic data suggest that ‘‘trust’’ is clearly linked with both the notion of commerce and the notion of business. Interestingly, although ‘‘distrust’’ was highlighted as a feature of business, I found little to no evidence of it in the data. I have no doubt, however, that examples of ‘‘distrust’’ exist within Essex, between Essex and other Schering organizations, and between Essex and the local community. One way to characterize organizations and communities, and their interaction, is along a continuum from trust to distrust, although any assessment is continually changing as social relations evolve.

OFFERING AN ALTERNATIVE FOCUS FOR ANALYSIS

As I read the chapter, the ethnographic material seems less pertinent to categorizing and defining the two types and more about understanding the 426 ELIZABETH K. BRIODY cultural processes that link them together. Reeves-Ellington describes one prominent cultural process that he labels ‘‘negotiation.’’ The word ‘‘negotiate’’ and its various forms (e.g., negotiations) appeared 15 times. As an example, Reeves-Ellington describes the relationship between Essex and members of the local community as a series of ‘‘negotiations of land purchase, actual purchase and payment, and plant siting and construction,’’ with the lurah acting as an intermediary in these activities. His use of the term ‘‘negotiations’’ directs attention to key players representing particular interests, with much of the action revolving around the lurah and Reeves- Ellington as a pair. Yet, both serve in negotiating and intermediary roles – the lurah on behalf of the Java peasants (as the chapter indicates) and Reeves-Ellington on behalf of Essex (a point not made very explicit in the chapter). Both had been charged by their constituencies with managing and settling the land-sale matter, and both sought to accrue benefits for their constituencies. Let’s put a spotlight for a minute on this cultural process described as negotiation. The ethnographic material reveals that the discussions between the lurah and Reeves-Ellington are predicated on a strong interpersonal relationship, and not just on the task of negotiating a single agreement. The relationship between the two began at the invitation of Essex’s Javanese personnel officer and lawyer, who spoke with both the lurah and the dukun about Essex’s desire to purchase the land. Relationship building began in earnest following Reeves-Ellington’s ‘‘introductions,’’ after which he ‘‘bought flowers, drank tea with the local officials, and joined polite conversations about wayang (shadow-puppet play performances).’’ Then, a ‘‘series of meetings’’ occurred involving Reeves-Ellington and two other Essex representatives with the lurah in which they ‘‘heard the lurah out’’ as he spoke of the impact of inheritance on the parceling of land. Ultimately, officials from the local community introduced the Essex representatives to the landowners. In each of these interactions, significant effort was expended in understanding the other group’s point of view, sharing information and insights, and developing cooperation. As indicated, the two sides followed the culturally appropriate protocol of indirect communication, using those in key local roles to gather further information or activate other social relationships that might be useful in helping them achieve their goals. What emerges is a complex network of interconnected relationships as Essex, with Reeves-Ellington as its key leader, begins its integration within the local culture. Essex demonstrates its intent through its collaborative orientation and remarkable flexibility. While obvious challenges had to be Targeting the Cultural Processes of Partnering for Analysis 427 overcome, Essex had a strong interest in problem solving and seemed to view the local communities and government officials with whom it interacted as potential allies. Essex was persistent in finding its way through a maze of unknown cultural rules and in its investment in building local relationships and sustaining them over the long term. Relationships were at the core of all of its interactions – whether those interactions were classified as commerce or as business. It was a case of organizational outreach and learning done, as well as Essex was able, Javanese style. What is so interesting is that the cultural process described as negotiation was not limited to Essex’s relationship with the local community: It also involved Essex’s own internal organizational interactions. Essex leaders in Indonesia were engaged in discussions about the land sale with other Essex leaders (e.g., those at the home office). They had to seek support for their ‘‘crazy’’ ideas. They had to reconcile Essex’s policies with local circumstances that did not always align with those organizational rules and procedures. They had to develop and sustain their own working relationships with those individuals in key organizational leadership positions. In considering the broker/intermediary roles filled by Reeves-Ellington and his Essex peers in Indonesia – both to those within the firm and to those in the local peasant and governmental communities – the range of activity suggests a broad and far-reaching network structure. A partnership model (Briody & Trotter, 2008, p. 11) is an alternative framework for documenting the cultural processes of network activity and functioning. In collaborative arrangements, such as the ones highlighted in the chapter, the parties are motivated to work together to overcome the constraints and challenges they face. (Social network diagrams can provide a visualization of the various connections between and within Essex and the local communities, starting with the land-sale negotiations and going through plant construction and beyond.) The primary partnership is the one between Reeves-Ellington and the lurah, although many other collaborative discussions were activated as a result of the discussions between these two parties. Many of the ‘‘partnering’’ interactions reported in the chapter – including those between Essex and the lurah on behalf of the local peasant community – occur within a mixed-exchange system. Market exchange is involved because Essex hoped to purchase land from the Javanese peasants. However, reciprocity exchange is also active because relationships are at the core of the high-context Javanese culture. A baseline of relationship strength is a necessary condition for any collaborative work (e.g., agreements) to proceed. As such, the interactions between Essex and the lurah are not simply ‘‘negotiations’’ (although specific points may be raised and considered), but 428 ELIZABETH K. BRIODY rather relationship-based cooperative exchanges directed toward particular (market and relationship) goals. This partnership pattern of interwoven market and reciprocity exchange (Trotter & Briody, 2006, pp. 4–6)is replicated throughout the network through various mechanisms including ‘‘tea money,’’ ‘‘the odd pair of tennis tickets in Jakarta,’’ and the creation of a communication chain through ‘‘friends’’ to friends of friends. To gain the maximum benefit from the use of a partnering framework as a heuristic device would require including additional data on local community assumptions, perspectives, and actions. As the chapter stands, there is rich detail on the key Essex roles associated with this business opportunity, leadership behavior, and the evolution of Essex views and policies over time. For example, the chapter describes numerous incidents when Essex made some accommodation with the local communities in an attempt to reach its business goals. If similar data were available documenting local peasant and governmental behavior with respect to Essex’s activities, as well as their perspectives of Essex’s business approach, the result would be a more complete and accurate understanding of community responses and compromises. With the current ethnographic data, it is difficult to know which accommodations were made by the local communities in their interactions with Essex. In focusing on the cultural processes of partnering, a variety of interesting questions for future investigations emerge.

Under what circumstances do relationships among kin, friends, work colleagues, and other associates incorporate attributes of market exchange in their interactions? To what extent does market exchange affect the quality of these relationships – such as the degree of trust, cooperation, and/or conﬂict? Which strategies do organizational and community leaders use to bridge the gap between their partners’ perspectives and behaviors and those of their own constituencies? To what extent do social networks remain open to future exchanges, particularly, if one of the partners knowingly or inadvertently violates a critical cultural rule during the exchange? How can the relationships that may have sustained any harm recover? How does the integration of market exchange and reciprocity exchange affect status and power in business organizations and local communities?

The answers to these and other questions would provide insight into the most effective and efﬁcient ways for these partners to collaborate in the Targeting the Cultural Processes of Partnering for Analysis 429 future. From an applications perspective, such information would be very valuable as a starting point for those interested in exploring potential partnership arrangements between organizations and communities. From a theoretical standpoint, the answers to these questions would help explain the precise linkages between market exchange and reciprocity exchange, their relative impact on each other, and the ways in which the two forms of exchange are combined in global economies today.

REFERENCES

Briody, E. K., & Trotter, R. T., Jr. (2008). Framing the partnership experience. In: E. K. Briody & R. T. Trotter, Jr. (Eds), Partnering for organizational performance: Collaboration and culture in the global workplace (pp. 3–14). Lanham, MD: Rowman and Littleﬁeld. Gerth, H. H., & Mills, C. W. (1946). From Max Weber: Essays in sociology. New York: Oxford University Press. Reeves-Ellington, R. (2009). Enviroscapes: A multi-level contextual approach to organizational leadership. In: F. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. Sengir, G. H., Trotter, R. T., Jr., Briody, E. K., Kulkarni, D. M., Catlin, L. B., & Meerwarth, T. L. (2004). Modeling relationship dynamics in GM’s research-institution partnerships. Journal of Manufacturing Technology Management, 15(7), 541–559. Trotter, R. T., Jr., & Briody, E. K. (2006). It’s all about relationships, not just buying and selling ideas: Improving partnering success through reciprocity (10540). Warren, MI: GM R&D Publication.

This page intentionally left blank

ENVIROSCAPES: THE CHALLENGES OF CULTURAL PARTNERING CONCEPTS

Richard Reeves-Ellington

ABSTRACT

Business and anthropological research create different constructions that frame the enviroscapes in which complex inter- and intra-organizational actors interface. This essay engages Briody’s (2009) three major areas of comment: methodology, theoretical concepts, and leadership processes used in the Essex scenarios. Briody’s is an anthropological construct, while Reeves-Ellington’s is that of a business practitioner/researcher. Expanding his original thoughts for conceptual organization, choice of research methods, and models for interaction allows the author to address Briody’s observations in ways that further an ongoing dialogue between academic research, practitioner, anthropology, and business, with the ultimate goal that others will join the conversation.

INTRODUCTION

I wish to thank Elizabeth Briody (2009) for a thoughtful and thought- provoking essay. Responding to her work has required me to be more

Multi-Level Issues in Organizational Behavior and Leadership Research in Multi-Level Issues, Volume 8, 431–443 Copyright r 2009 by Emerald Group Publishing Limited All rights of reproduction in any form reserved ISSN: 1475-9144/doi:10.1108/S1475-9144(2009)0000008017 431 432 RICHARD REEVES-ELLINGTON explicit in understanding the theoretical foundations I have used in my research for the last 20 years and to explain why I choose the research methodologies used and their benefits and limitations. With her discussions of partnerships, Elizabeth has encouraged me to revisit the anthropological literature and business literatures where I have been situated. She sums up three surprises within my discussions: the reasons for choosing the concepts that I did, the reasons for not using traditional anthropological core elements when discussing enviroscapes and their phenotypes, and the reasons for using the attributes for business and commerce. One might argue that these differences stem from the different research streams used by business and anthropology. The use of traditional business concepts, research methodologies, and multicultural interactions prompt conclusions that differ from those arrived at through an anthropological frame of reference. Briody’s conceptual starting point is that of a cultural anthropologist, who works within cultural frameworks that inform theory, methodologies, and complex social interactions. Within such an analytical framework, her concerns about my having two primary categories of ‘‘business–organization culture and peasant–community culture’’ seem explainable. My conceptual starting point is to identify, understand, and perhaps create interrelations that describe imagined communities (Anderson, 1983). I then inform my theory using the role of practitioner–ethnologist, who avoids cultural contexts by relying on enviroscapes. The practitioner– ethnologist demands methodologies that permit self-definitions of participants of specific activities (Reeves-Ellington, 2009). A researcher who starts with the viewpoints of a variety of actors involved often captures weltanshauungen of participant actors, rather than worldviews that are attributed to them by social science constructs. Briody and Trotter (2008a) would frame their methodology in terms of modeling the external environments that affect specifically defined cultural groups (Briody & Trotter, 2008b). In contrast, my work employs a series of theory-building activities that begin with a preliminary description based on observation (phenomena), progress through classification (frameworks and typologies of phenomena), and end in defining relationships (models that are statements of association) (Reeves-Ellington, 2007b). Briody suggests that a discussion of organizational relationships is best framed within partnership theory and the associated roles of leadership necessary to affect partnerships. I strongly support her ideas of the value of partnership and wish that I had seen them prior to writing my enviroscapes chapter. Nevertheless, the traditional leadership concepts she suggests as The Challenges of Cultural Partnering Concepts 433 valid are probably inadequate to build organizational partnerships that are based on multiple levels of individuals rather than on one or two key negotiators. Specifically, shamanistic leadership is required to build intra- organizational (legal and social) lasting relationships. Other transactional and transformational leadership acts come from other actors. The remainder of this chapter expands on these issues, raises possible areas of useful research, and explores possible ways to integrate ethnographic and cultural theory.

CONCEPTUAL CONSTRUCTIONS

Briody (2009) frames her essay in terms of a cross-cultural discussion. In contrast, I have avoided the concepts of culture and cross-culture, as I find that the first tends to be defined by static categories and the second causes a focus on differences. Briody and Trotter (2008b, p. 16) define culture as ‘‘the assumptions, expectations, beliefs, social structures, and values guiding behavior.’’ Such a definition leads to categorization that would indicate a need for a ‘‘business–organization culture’’ and ‘‘peasant–community culture’’ as suggested by Briody (2009). During my early work in Indonesia (Reeves-Ellington, 1994), my efforts at framing reporting in terms of culture demanded such categories, but the results were unsatisfactory because they remained in the what and largely ignored the how and why of interactive events. Within the business literature, the concept of culture as a major research topic typically focused on a question posed by Linda Smircich (1985): Is the concept of culture a paradigm for understanding organizations and ourselves? By 2000, the answer to this question was largely negative (Ashkanasy, Wilderrom, & Peterson, 2000; Sackman, 1992; Schein, 2001). To my mind, the problem revolved around static topologies that made action-oriented analysis difficult. I revised my basis (Reeves-Ellington, 1995a, 1995b) of understanding by using topologies created by experts to shared meanings within self- discovered contexts (Geertz, 1961, 1966; Kluckhohn & Strodtbeck, 1961). In short, I shifted from the study of academically defined cultural categories to the use of ethnographical, participant-defined categories. Within this frame of reference, researchers must respect the importance of shared meanings that are created by actors’ performances within webs of interactions. The learnings of participants are then expressed through self-revelation and with some feedback mechanisms assumed to be legitimate. After such revelation, the original actors can inform outsiders of what has been 434 RICHARD REEVES-ELLINGTON learned. Understanding such webs requires a ‘‘thick description’’ (Geertz, 1961) of intertwined layers of common meaning that underlie what people say and do. Apropos to this chapter, the strands of business, commerce, trust, and distrust phenotypes that are interacting through their enviroscape requires thick description. Finally, thick description requires the tracing of interactive webs as meaning evolves, thus requiring longitudinal study. The question of how to organize the resultant learning into meaningful structures that permit analysis and action then lies outside traditional cultural categories and within concepts that might be universally recognized. Literature reviews, my field notes review, and discussions with individuals involved in the Indonesian project suggested that all interactions between actors have three elements:

1. Human interactions occur in open systems that are interdependent with their environments. 2. Interdependence is the core of human interactions. 3. Environments are highly contextual.

I hypothesize that these elements are best studied through examination of the environment in which they occur (Reeves-Ellington, 2004, 2007b). The next question concerns which environments are both necessary and sufficient for understanding human interactions within structured situations. I hypothesize that five environments play key roles in this respect: climate, knowledge, values, time, and leadership (Reeves-Ellington, 2009). Within any structure or set of structures, whether social or organizational, primacy will vary according to context. As Briody contends, and I agree, each of these aspects has been the subject of an extensive literature that largely ignores other environmental issues, because these authors tend to start with the single environment rather than complex contexts. However, this issue must be addressed within my conceptual frameworks rather than avoided through other constructs. I further attribute two phenotypes, business and commerce (Reeves-Ellington, 2004, 2007b; Reeves-Ellington & Anderson, 1997a), to each of the environments, other than ethnos and leadership, to better structure and understand to what ends actors use specific generic environments. Briody (2009) interpreted my intent on the two phenotypes

. . . to be to identify the differences in cultural expectations, assumptions, beliefs, processes, practices, and behaviors linked to these notions in the literature and The Challenges of Cultural Partnering Concepts 435

supported by the case material. He describes business and commerce as ‘‘phenotypes’’ for several different organizational environments (i.e., ‘‘enviroscapes’’), setting each up in opposition to the other without the potential for overlap. I prefer the phrase ‘‘ideal types’’ in Weber’s usage (i.e., one type is understood in relation to another).

I do not believe in an either/or position for the two phenotypes but would rather agree with Briody that the two not only have the potential for overlap but, in fact, do overlap (Reeves-Ellington, 2004, 2007b). Particular social structures will use one or the other as a default position, but in specific contexts that default position might be overridden by the other phenotype (Reeves-Ellington, 2007b). For example, within the case material (Reeves- Ellington, 2009), Schering had a default position of a business phenotype. However, in dealing with the intricacies of the Indonesian contexts in which Schering found itself, key leaders were willing to shift to a commerce interaction. For this reason, I would avoid the ‘‘ideal types’’ concept and opt for ‘‘preferred phenotype’’ for specific contexts. Briody further questions whether the contrast between business and commerce was based on ‘‘emic’’ (i.e., local, native) categories that would suggest a greater degree of Indonesian integration or holism with respect to these two concepts. She is right that high-context Javanese did not have clear-cut categories; in contrast, Chinese Indonesians did, as did low-context Indonesians who worked primarily within a Dutch-educated and -trained hierarchy. All of these types, through interactions, worked within both concepts. The default positions, however, often led to an understood overt conflict that was caused by embedded tensions between the phenotypes. In trying to address the interrelationships and actions between various social structures and organizations, I relied on environments and their associated phenotypes as the differentiators. My work focuses on the margins of complex social interactions. These interactions are bounded by a variety of social structures that form and re-form to create changing patterns of interaction. As such, key cultural patterns of each social group are not useful for understanding and using the Indonesian case materials.

RESEARCH METHODS

In her review of the case study materials, Briody (2009) focused on the cultural processes that would examine the most effective and efﬁcient ways to examine interrelationships among differing social structures – in this case, 436 RICHARD REEVES-ELLINGTON the Schering organizations and the community where Schering constructed its plant. Some of the questions Briody raises are of scholarly interest:

Under what circumstances do relationships among kin, friends, work colleagues, and other associates incorporate attributes of market exchange into their interactions? To what extent does market exchange affect the quality of these relationships – such as the degree of trust, cooperation, and/or conﬂict? Which strategies do organizational and community leaders use to bridge the gap between their partners’ perspectives and behaviors, and those of their own constituencies? To what extent do social networks remain open to future exchanges – particularly if one of the partners knowingly or inadvertently violates a critical cultural rule during the exchange? How can the relationships that may have sustained harm recover? How does the integration of market exchange and reciprocity exchange affect status and power in business organizations and local communities?

Briody believes that from an applications perspective, such information would be very valuable as a starting point for those interested in exploring potential partnership arrangements between organizations and communities. I agree. The answers to these and other questions would provide insight into the most effective and efficient ways for these partners to collaborate in the future. The case study reflects activities that were undertaken to meet the ultimate goals of Schering. Schering managers had a primary business purpose. To achieve this goal, they had to learn about and focus on commerce activities of various Indonesian communities: the Department of Health, the Foreign Investment Board, the Ministry of Region (Islam) in Djakarta; the Regional Land Office in Surabaya (East Java); village government authorities (from a variety of regions in Java), and local religious leaders (not Muslim, but rather traditional folk religion). All of these groups were acting as negotiating surrogates for others, which precluded Schering from gaining direct access to key decision makers in Djakarta, the Regional Governors office in Surabaya, and the villagers in East Java. The task at hand led me to adopt action research and participant action research methodologies as part of my study (Reason, 1994; Reeves-Ellington, 1995b; Reeves-Ellington & Anderson, 1997a, 1997b). The overarching methodological paradigm of the case material is cooperative inquiry. The Challenges of Cultural Partnering Concepts 437

Four fundamental methodological steps required for cooperative inquiry were used:

Phase 1. The coresearchers agree on an area for inquiry and identify initial propositions to be examined.

Phase 2. The coresearchers apply these ideas and procedures in their everyday life and work.

Phase 3. The coresearchers become fully immersed in this activity and experience.

Phase 4. After an appropriate period engaged in Phases 2 and 3, the coresearchers return to consider their original propositions in light of experience, modifying, reformulating, and rejecting them as necessary.

The research methods used had to provide information that would be sufficient to meet all parties’ needs to resolve agreed issues. By conducting systematic, firsthand participant observation (PO) using action research, an experiential and interactive mode of learning emerged (David, 1985). Thus, the learning environment was at the fore during much of the period. All the parties were co-opted as participant observers partaking in action research. Within this framework, all participants are researchers. In the case study (Reeves-Ellington, 2009), their number included all the Schering leaders and managers involved with the Indonesian project, the government officials who worked with Schering to make the investment successful, and the religious leaders who helped the peasants with their ethical issues. The effectiveness of the information system depended on communicative transparency by all actors involved. At the same time, the cooperative research paradigm prevents the development of a holistic understanding of all actors’ motivations and rationale for actions. Within the Indonesian material, by using the cooperative paradigm, we developed ethnography of living and changing contents of specific enviroscapes. None of the participants pursued a gestalt understanding of all variables and their interactions. Had we attempted a total understanding, as Briody judges to be important, we might have jeopardized our creation of an imagined community at the points of intersection important to all of us. Further, had we pursued such knowledge – knowledge that was not necessary to improve our cooperation – we could have created a distrust that would have separated us in our efforts to succeed. Finally, the 438 RICHARD REEVES-ELLINGTON pursuit of extraneous knowledge would have impaired the building of trust in the ethnos environment. Such a loss would have hindered long-term collaboration and partner building. A fruitful area of future research might be to examine ways that practitioner–participant methodologies might be used to support more traditional cultural methods of study. A first step would be to agree on a general framework for theory building (Carlile & Christensen, 2005)and then proceed from that point.

ANALYSIS

A partnership model (Briody & Trotter, 2008a, p. 11) is a framework for documenting the leadership interactions that I describe consistently as negotiations. Briody (2009) correctly points out that in collaborative arrangements, such as the ones highlighted in the case study (Reeves-Ellington, 2009), the parties were motivated to work together to overcome the specific constraints and challenges they faced. She states that the case materials support a partnership pattern of interwoven market and reciprocity exchange (Briody & Trotter, 2008b). Briody would prefer a baseline of relationship strength as a necessary condition for any collaborative work to proceed. I agree, and would postulate that the relationship strengths reflect the adoption of a specific leadership phenotype that is applied within another specific enviroscape and phenotype – for example, ethos and trust. In the case study, the need for a variety of partnerships required Schering personnel and their Indonesian institutional counterparts to practice different types of leadership, depending on operational levels within the company and the geographic location of the leader. Coupled with the leadership style variances, each set of activities involved working in a variety of enviroscapes, with an appropriate phenotype, to address and resolve shared issues and goals. Let’s return to the case material for examples. Given the complexity of interactions between a myriad of people in the case materials, I will limit my discussion examples to Essex and village partnership frameworks within Schering, and between Reeves-Ellington and the lurah. Within the Schering organization, the company president, E. Conzen, practiced transformational leadership by having the company temporarily suspend its primary business orientation to a commerce preference for the Indonesian operation. He assured alignment across organizational levels and geographic boundaries that guaranteed (1) business would be done but The Challenges of Cultural Partnering Concepts 439

(2) the commerce side would control business activities. The primary enviroscape affected by Conzen’s transformational leadership was that of climate – specifically, a climate characterized by relationship-based cooperative exchanges directed toward particular business goals. Within this climate, the regional leaders could reach out to Hong Kong Chinese who had influence in Indonesia and engage in transactions that would further Schering’s Indonesian operations. Both the external Chinese actors and the Schering participants worked primarily within the leadership environment to achieve shared objectives. Within the leadership environment, their activities were largely transactional business phenotype, market exchange activities. The regional vice president, R. Weiser, held the exact nature of these exchanges private within Schering’s Hong Kong office. At the national level (Indonesia), Reeves-Ellington’s leadership model when dealing with his Indonesian counterparts was that of a shaman who would integrate the profane (business) and the sacred (commerce). Actions concerning the sacred required working in the enviroscape of ethos for the development of the phenotype of trust. The Schering managers’ primary goal was to build institutional relationships (both formal and informal) between Schering and its Indonesian counterparts. Those primary counterparts included the Director of the Foreign Investment Board, the Director of the Pharmaceutical Division of the Department of Health, the lurah of the village where the plant was to be constructed, and the lurah’s dukun in the village to assure that all was done in keeping with traditional religious belief. Reeves-Ellington’s profane side of the equation was not to be overly active in transactional specific activities. Rather, his leadership role was to assure that such activities were carried out in ways that created external institutional trust. At the intracompany level, Reeves-Ellington’s leadership role was primarily transactional within the time enviroscape, using the business phenotype. This leadership role might be considered the profane side of shamanistic leadership. The primary business goal was to complete investment approval, land purchase, plant construction, and start-up within a two-year period. These goals were the business measures of success for the entire Essex team. Business reports did not discuss peasant issues but rather land purchase transactions. Land authorization was not reported in terms of official corruption holdups but rather in terms of progress impediments at the regional level. Within the context of plant activities that involved Schering, the lurah had a shamanistic leadership role in aligning commerce (the profane) and commerce (the sacred) of his village communities: Islam, folk religion, and 440 RICHARD REEVES-ELLINGTON communal values on the sacred side; land sales, jobs, and communal land use on the profane side. As with Reeves-Ellington, the need to build trust between the villagers and Essex required the lurah to work within the enviroscape of ethos. The changes in the village that increased Islamic leadership, the need to find jobs for the peasants who sold their land, and the need to find work for women of the village that was plant related all dictated a transformational leadership role for the lurah. Many of the profane activities pushed the enviroscape of time to the fore owing to the existence of time-driven goals, although these goals were not universally recognized by all parties. The peasants selling the land wanted to go on hadj within the calendar year. The lurah wanted the land use decided prior to rice planting time. Some women of the village wanted to start food services from the beginning of plant construction. These goals were not discussed in terms of cost of the hadj trip, lost crop sales, or lost income, but rather in terms of the village good: more hadi to lead the village, maintenance of family through rice production, helping the landless women with village jobs. All transactions were set within the commerce phenotype of time. In summary, the baseline of relationship strength necessarily involved the interactive use of three enviroscapes and both phenotypes for specific activities. These interactions were a necessary condition for collaborative work to proceed. As Briody (2009) states, to gain the maximum benefit from the use of a partnering framework as a heuristic device would require including additional data on local community assumptions, perspectives, and actions. Sadly, this possibility was not an alternative at the time of data collection and actor interactions. Such a broad-scope activity would have required several full-time, multidisciplinary researchers. Social network diagrams would have provided a visualization of the various connections between and within Essex and the local communities (Briody, 2009), but they were not done. Even if they had been considered at the time of data collection and actor participation, they would have been incomplete and perhaps misleading. Consider the Essex Indonesian attorney, Nugroho, and his extensive networking for Essex. His primary enviroscape relationships with the home office were those of information and time, both within the business phenotype. His leadership phenotype was transactional. To accomplish the shared Essex goals, Nugroho had business and social relationships with Essex’s home office attorneys (with attendant confidentiality), Filipino consultant, and accounting firm; the Department of Foreign investment; the Ministry of Health; East Java Regional Government authorities; and village The Challenges of Cultural Partnering Concepts 441 authorities, both civil and religious. In addition, Nugroho acted as an informal dukun for Reeves-Ellington. In this role, he used his home village contacts to inform him of local events and attitudes. This role was particularly important for Reeves-Ellington’s success. While his relationships with Schering could be networked by enviroscape and appropriate phenotype, the information needed input from the Indonesian networks for complete understanding. Nugroho kept the commerce side of his actions backstage and hidden from view. With the current ethnographic data and lack of network diagramming, Reeves-Ellington could not know which accommodations were made by the local communities in their interactions with Essex. However, 30 years after the events described in the case study, the Essex plant had never had labor problems or village conflicts that resulted in serious confrontation (Swaak, 2002).

CONCLUSIONS

There are two underlying differences in the way Briody and I approach fieldwork and ethnography studies. First, Briody seeks a holistic approach that would be all inclusive for understanding all the cultures of all the actors, whereas I am satisfied with the levels of understanding that provide the basis for the creation of shared communities among the actors involved. Second, Briody approaches research materials as an academic anthropologist, whereas my approach is much more attuned to practitioner business anthropology. Such a practitioner works toward decision making that results in satisfaction by all involved actors. There is common ground between the traditional cultural anthropologist and the practitioner–ethnologist approaches. The first step would be to gain agreement for a common research paradigm that would allow past work to be evaluated and then integrated into new constructs. The next step might be to build a theory that moves scholarship forward in ways that satisfy all the research community. I have suggested such a model in my recent work (Reeves-Ellington, 2007a). As Briody suggests (Briody & Trotter, 2008a), any such integration must take into consideration evolutionary theory, structural theory, cognitive theory, and symbolic theory. Business literature of climate, knowledge, time, ethnos, and leadership must also be thrown into the mix. Such an undertaking by a single researcher is unlikely, but would rather require a team of scholars to bring the information together in ways that would provide an integrated conceptual framework. 442 RICHARD REEVES-ELLINGTON

REFERENCES

Anderson, B. R. (1983). Imagined communities: Reflections on the origin and spread of nationalism. London: New Left Books. Ashkanasy, N. M., Wilderrom, C., & Peterson, M. F. (Eds). (2000). Handbook of organizational culture and climate. Thousand Oaks, CA: Sage. Briody,E.K.(2009).Targeting the cultural processes of partnering for analysis. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues.Bingley,UK:Emerald. Briody, E. K., & Trotter, R. T. (2008a). Framing the partnership experience. In: E. K. Briody & R. T. Trotter (Eds), Partnering for organizational performance (pp. 3–14). New York: Rowman & Littlefield. Briody, E. K., & Trotter, R. T. (Eds). (2008b). Partnering for organizational performance. New York: Rowman & Littlefield. Carlile, P. R., & Christensen, C. M. (2005). Practice and malpractice in management research. Unpublished manuscript. Boston, MA. David, K. (1985). Participant observation in pharmaceutical field selling. Norwich, NY: Norwich Eaton Pharmaceutical. Geertz, H. (1961). The Javanese family: A study of kinship and socialization. Prospect Heights, IL: Waveland. Geertz, H. (1966). Person, time and conduct in Bali: An essay in cultural analysis. New Haven, CT: Yale Southeast Asia Program Cultural Report Series. Kluckhohn, F. R., & Strodtbeck, F. L. (1961). Variations in value orientations. Westport, CT: Greenwood Press. Reason, P. (1994). Three approaches to participative inquiry. In: N. K. Denzin & Y. S. Lincoln (Eds), Handbook of qualitative research (pp. 324–339). Thousand Oaks, CA: Sage. Reeves-Ellington, R. (1994). Corporation anthropologists at work in the Third World. In: V. H. Sutlive & T. Hamada (Eds), What can multinationals do for peasants (pp. 213–240). Williamsburg, VA: College of William and Mary. Reeves-Ellington, R. (1995a). Anthropology and total quality management: Improving sales force performance in overseas markets. In: J. F. J. Sherry (Ed.), Contemporary marketing and consumer behavior (pp. 169–208). Thousand Oaks, CA: Sage. Reeves-Ellington, R. (1995b). Organizing for global effectiveness: Ethnicity and organizations. Human Organization, 54(3), 249–262. Reeves-Ellington, R. (2004). Trustscapes and distrustscapes: A multi-level approach for understanding stability and change. In: F. J. Yammarino & F. Dansereau (Eds), Multi- level issues in organizational behavior and processes (Vol. 3, pp. 91–148). Boston, MA: Elsevier. Reeves-Ellington, R. (2007a). Let us theory build rather than theorize. In: F. Dansereau & F. J. Yammarino (Eds), Research in multi-level issues (Vol. 6, pp. 339–349). Boston, MA: Elsevier. Reeves-Ellington, R. (2007b). Timescapes: A multi-level approach for understanding time use in complex organizations. In: F. Dansereau & F. J. Yammarino (Eds), Multi-level issues in organizations and time (Vol. 6, pp. 259–317). Boston, MA: Elsevier. Reeves-Ellington, R. (2009). Enviroscapes: A multi-level contextual approach to organizational leadership. In: F. J. Yammarino & F. Dansereau (Eds), Multi-level issues in organizational behavior and leadership. Vol. 8 of Research in Multi-Level Issues. Bingley, UK: Emerald. The Challenges of Cultural Partnering Concepts 443

Reeves-Ellington, R., & Anderson, A. (1997a). Business, commerce, and social responsibility. Lewiston, NY: Mellon Press. Reeves-Ellington, R., & Anderson, A. (1997b). The ethnology of information systems. Accounting, Management, and Information Technologies, 7(3), 168–180. Sackman, S. (1992). Culture and subcultures: An analysis of organizational knowledge. Administrative Science Quarterly, 37, 140–161. Schein, E. H. (2001). Sense and nonsense about culture and climate. In: N. M. Ashkanasy, C. Wilderrom & M. F. Peterson (Eds), Handbook of organizational culture and climate (pp. xxiii–xxx). Thousand Oaks, CA: Sage. Smircich, L. (1985). Is the concept of culture a paradigm for understanding organizations and ourselves. In: P. J. Frost (Ed.), Organizational culture (pp. 55–72). Beverly Hills, CA: Sage. Swaak, R. (personal communication, September 2, 2002). No strikes in Indonesia. In: R. Reeves-Ellington (Ed.). Binghamton, NY.

This page intentionally left blank

PART VI ABOUT THE AUTHORS

This page intentionally left blank

ABOUT THE AUTHORS

Neal M. Ashkanasy is a Professor of Management at the University of Queensland, Australia. His research interests lie in organizational and ethical behavior, leadership, culture, and emotions. He is Editor-in-Chief of the Journal of Organizational Behavior and the book series Research on Emotion in Organizations. Janice A. Black (Ph.D., Texas Tech University) an Associate Professor at California State University, Bakersﬁeld, has published articles in several leading journals (The Leadership Quarterly, Strategic Management Journal, Entrepreneurship: Theory and Practice) and in the Advances in Applied Business Strategy: Competence-Based Management series. Dr. Black and colleagues use agent-based modeling to examine leaders and followers. Elizabeth K. Briody, Ph.D., is a Technical Fellow and cultural anthropologist at General Motors R&D, in Warren, Michigan. She conducts research on various aspects of GM’s culture. Her most recent projects have focused on views of integrated health and wellness, partnership effectiveness, and perceptions of an ideal plant culture. She is an adjunct professor at Northern Arizona University, Michigan State University, and Wayne State University, and has served as President of the National Association for the Practice of Anthropology, a section of the American Anthropological Association. Jay J. Caughron is a doctoral student in industrial and organizational psychology at The University of Oklahoma. His research interests include leadership and planning. Fred Dansereau, Ph.D., is Professor of Organization and Human Resources and Associate Dean for Research in the School of Management at the State University of New York at Buffalo. He received his Ph.D. from the Labor and Industrial Relations Institute at the University of Illinois with a specialization in Organizational Behavior. Dr. Dansereau has extensive research experience in the areas of leadership and managing at the individual, dyad, group, and collective levels of analysis. Along with others,

447 448 ABOUT THE AUTHORS he has developed a theoretical and empirical approach to theorizing and testing at multiple levels of analysis. He has served on the editorial review boards of the Academy of Management Review, Group and Organization Management,andLeadership Quarterly. Dr. Dansereau is a Fellow of the American Psychological Association and the American Psychological Society. He has authored 12 books and over 80 articles and is a consultant to numerous organizations, including the Bank of Chicago, Occidental, St. Joe Corp., Sears, TRW, the United States Army and Navy, Worthington Industries, and various educational institutions. John N. Davis completed his Ph.D. at Texas Tech University. He is assistant professor of management at the Kelley College of Business at Hardin- Simmons University in Abilene, Texas. His research examines how leaders change over time, with emphasis on application to practice. Shelley D. Dionne received her Ph.D. and MBA from Binghamton University, State University of New York, School of Management in 1998 and 1994, respectively. Her research interests include leadership, creativity, teambuilding, and training and development; and her publications have appeared in Journal of Applied Psychology, Leadership Quarterly, Research in Multi-Level Issues, Journal of Organizational Behavior, and Human Relations. She teaches leadership and organizational behavior in the School of Management’s undergraduate, graduate, and executive programs. She is the Associate Director of the Center for Leadership Studies at Binghamton University. Peter J. Dionne received his Ph.D. and MS in mechanical engineering at Rensselaer Polytechnic Institute in 2006 and 1990, respectively. His research interests include discrete modeling of polymer nanocomposite systems and multiscale modeling of materials; and his publications have appeared in Macromolecules and Journal of Chemical Physics. He is Senior Mechanical Engineer at Sensis Corporation in East Syracuse, NY. Katey E. Foster is a student in the doctoral program in Industrial/ Organizational Psychology at the University of Akron. Her research interests include topics in leadership and feedback seeking. Tamara L. Friedrich is a doctoral student in industrial and organizational psychology at The University of Oklahoma. Her research interests include leadership and creativity. Rosalie J. Hall is an Associate Professor of Psychology at the University of Akron. She received her Ph.D. from the University of Maryland in 1988. About the Authors 449

Her research focuses on interpersonal perceptions in organizations and research methodology. James G. (Jerry) Huntw (Ph.D., University of Illinois, 1966) was Paul Whitﬁeld Horn Emeritus Professor of Management and former Director of the Institute for Leadership Research at Texas Tech University. He was a Fellow of the Academy of Management and Southern Management Association and authored, coauthored, edited or coedited more than 20 books and monographs and nearly 200 articles, book chapters, and related materials. Dr. Hunt also edited both the Journal of Management and The Leadership Quarterly. He served on the national governing board of the Academy of Management, as the chair of the Academy’s organizational behavior division, and as president of the Southern Management Associa- tion and chair of the Midwest Academy of Management. Dr Hunt also won a career-long service award at the national level and at one of the regional levels and won an outstanding university-wide research award at Texas Tech. At the time of his death (in 2008), his key interests were dynamic and macro approaches to OB and leadership and to the sociology of the science of management. Samuel T. Hunter is an Assistant Professor in the Industrial and Organizational Psychology program at Pennsylvania State University. He received his Ph.D. in Industrial and Organizational Psychology from the University of Oklahoma. He is a member of the Society of Industrial and Organizational Psychology and the Academy of Management. Dr. Hunter has worked for over four years with the Department of Defense and published multiple articles and book chapters in the areas of leadership, creativity, innovation management, and multi-level innovation. David A. Kenny is a Distinguished Alumni and Board of Trustees Professor at the University of Connecticut. He is the author of six books, the most recent being Dyadic Data Analysis. He has published in the areas of interpersonal perception, mediational analysis, and analysis of data from dyads and groups. Stefano Livi is a researcher in social psychology at the Department of Developmental and Socialization Processes, Sapienza University of Rome, Italy. He is currently interested in small-group processes, including norms transmission and socialization, communication, and group composition. Robert G. Lord is a Distinguished Professor of Psychology at the University of Akron. He received his Ph.D. from Carnegie-Mellon University in 1975. 450 ABOUT THE AUTHORS

His research focuses on motivation and self-regulation, emotions, leadership, and information processing. He has coauthored the books Leadership and Information Processing: Linking Perceptions and Performance with Karen Maher, and Leadership Processes and Follower Self-Identity with Douglas Brown. Michael D. Mumford is the George Lynn Cross distinguished professor of Industrial and Organizational Psychology at the University of Oklahoma where he directs the Center for Applied Social Research. Dr. Mumford received his Ph.D. from the University of Georgia in 1983. He has held faculty positions at the Georgia Institute of Technology and George Mason University. Dr. Mumford has written more than 150 articles focusing on leadership, creativity, planning, and integrity. He is currently the senior editor of The Leadership Quarterly and serves on the editorial board of the Creativity Research Journal, The Journal of Creative Behavior, and IEEE Transactions on Engineering Management. He is a fellow of the American Psychological Association (Divisions 3, 5, and 14), the American Psychol- ogical Society, and the Society for Industrial and Organizational Psy- chology. He is a recipient of the M. Scott Myers award for applied research in the work place. Richard L. Oliver (Ph.D., New Mexico State University), an Associate Professor in the College of Business at New Mexico State University, has published in journals including: The Leadership Quarterly, Strategic Management Journal, Accounting Education: An International Journal, Journal of Engineering Education (JSE), Journal of Education for Business, and in the Advances in Applied Business Strategy: Competence-Based Management series. Dr. Oliver pursues research in simulation and modeling, performance analysis, and distance learning paradigms. Lori D. Paris (Ph.D., New Mexico State University), an assistant professor at California State University, Bakersﬁeld, is currently working on a number of papers addressing topics such as women in leadership, motivational aspects of knowledge sharing, and cross-cultural leadership. Dr. Paris recently published in the Leadership and Organization Development Journal. She received the Best Dissertation Award from the Academy of Management. Richard Reeves-Ellington is currently Professor Emeritus in the School of Management at Binghamton University, State University of New York, where he has served as Associate Dean and Acting Dean. He is also on the faculty and a consultant for Excelsior College. He taught at the American University in Bulgaria and Soﬁa University in Bulgaria as a Fulbright About the Authors 451

Senior Scholar. His fields of interest revolve around cross-cultural aspects of global organization, marketing, and business strategy. He has also served on the Fulbright Selection Committee for Southeast Europe (acting as chair for one year), the Muskie Foundation for students from the Confederation of Independent States, and the Fulbright Senior Scholars Program. His initial 33-year career in the pharmaceutical industry included 19 years in Asia, Europe, and Latin America. Eduardo Salas, Ph.D., is Trustee Chair and Pegasus Professor of Psychology at the University of Central Florida. He has coauthored more than 300 journal articles and book chapters, has edited 18 books, has served or currently serves on 15 editorial boards, is past Editor of Human Factors journal, and is current Associate Editor of Journal of Applied Psychology. He is a Fellow of Division 14 (twice recipient of the Division’s applied research award), 19 (recipient of the Division’s Gersoni award for scientific contributions to the field), and 21 (recipient of the Division’s Taylor award for scientific contributions to the field) of the American Psychological Association. Dean Keith Simonton received his Ph.D. in social psychology from Harvard University. He is currently Distinguished Professor of Psychology at the University of California, Davis. His more than 350 publications – including 10 books – treat various aspects of genius, creativity, leadership, talent, and aesthetics. For information, see http://psychology.ucdavis.edu/Simonton. Jessica L. Wildman is a doctoral student in the Industrial and Organiza- tional Psychology program at the University of Central Florida, where she has been a graduate research assistant at the Institute for Simulation and Training since 2007. Her current research interests include team performance, multicultural performance, teams in complex settings, simulation in team research, and virtual teams. Francis J. Yammarino, Ph.D., is SUNY Distinguished Professor of Management and Director and Fellow of the Center for Leadership Studies at the State University of New York at Binghamton. He received his Ph.D. in Organizational Behavior (Management) from the State University of New York at Buffalo. Dr. Yammarino has extensive research experience in the areas of superior–subordinate relationships, leadership, self-other agreement processes, and multiple levels of analysis issues. He has served on the editorial review boards of eight scholarly journals, including the Academy of Management Journal, Journal of Applied Psychology, Journal of Organizational Behavior, Leadership Quarterly, Organizational Research 452 ABOUT THE AUTHORS

Methods, and Personnel Psychology. Dr. Yammarino is a Fellow of the American Psychological Society and the Society for Industrial and Organizational Psychology. He is the author of 13 books and has published over 100 articles. Dr. Yammarino has served as a consultant to numerous organizations, including IBM, Textron, TRW, Lockheed Martin, Med- tronic, United Way, Skills Net, and the US Army, Navy, Air Force, and Department of Education.