Action Science

Concepts, Methods, Skillsand for Research and Intervention Chris Argyris Robert Putnam Diana McLain Smith Action Science

Jossey-Bass Publishers San Francisco ACTION SCIENCE Concepts, Methods, Skills Researchand for Interventionand Chriy b s Argyris, Robert Putnam Diand an , a McLain Smith

Copyrigh : 198Jossey-Bas© t by 5 s Inc., Publishers 350 Sansome Street San Francisco, California 94104

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording otherwiser o , , withou prioe th t r written permissio publishere th f no .

Library of Congress Cataloging-in-Publication Data Argyris, Chris (date) Action science. (The Jossey-Bass social and behavioral science series) (The Jossey-Bass management series) Bibliography; p. 451 Includes index. . Socia1 l sciences—Research . Socia2 . l sciences—Methodology . Actio3 . n research. I. Putnam, Robert (date). II. Smith, Diana McLain (date). III. Title. . SeriesIV . SeriesV . : Jossey-Bass management series. H62.A663 1985 300'.72 85-18054 ISBN 0-87589-665-0

Manufactured in the United States of America

JACKET DESIG WILLY NB I BAUM

FIRST EDITION HB Printing 1098765

Code 8528 A joint publication in The Jossey-Bass Socia Behaviorad an l l Science Series and The Jossey-Bass Management Series

Consulting Editor Methodology of Social and Behavioral Research

Donald W Fiske University of Chicago students,Toour from whom learnwe muchso Preface

Creating usable knowledg s becomini e n increasinglga y important e sociatopith n ci l sciences. Lindbloo Cohed man n (1979), for example, have written about producing knowledge than ca t be used to formulate policies. Our focus is on knowledge that usee b produco dt n ca e action, same whilth t ee a tim e contrib- utina theor o t g f action o y e concepTh . f usablo t e knowledge has produce uneasn da y mixtur f enthusiaso e skepticismd man . It has generated enthusiasm because we need more usable knowledge to help manage interpersonal, community, and organizational affairs. Moreover, technological spinoffs from the physical sciences suggest thae sociath t l sciences might generate similar benefit r sociafo s l practice t therBu .s widespreai e d skepticis s wellma . Policie r dealinfo s g with poverty, discrimina- tion, and unemployment bog down in the complexities of im- plementation retrospectn i d an , , some observers argue that these policies have made the problems worse. Programs for transform- ing organizations succeed each other with the seasons, leaving in their wake the weary wisdom that nothing really changes, Re- sponsible social scientist respony sma theso dt e disappointments

IX x Preface

by turning inward to research that seems increasingly esoteric to practitioners. In proposing an action science, we hope to articulate the feature a scienc f o s e than generatca t e knowledge tha s usei t - ful, valid, descriptiv e worldth d informativf e o ew an , w ho f o e might change it. This emphasis on advancing basic knowledge while also solving practical problems has had a long and distinguished caree sciencen i re naturath n I . l science illustrates i t si d by the work of Louis Pasteur, who discovered much about the role of germs in illness while trying to solve problems of fer- mentatio r Frencfo n h vintners s alsi ot I . illustrate y earlb d y wor operationn ki s research: scholar t asidpu s e their interesn i t basic researc helo ht p England solve critical practical problems during World War II. In the course of this work, they discovered exciting intellectual problems whose solution contributed to basic knowledge. sociae th n lI sciences this emphasi combininn so g science d practican s usualli e y entitled action research woule W . e b d content to use the term action research if it were not for two factors. First, ovee yearth r s action researc s ofteha h n been separated from theory buildin d testinggan . Leading social scientists distinguish action research from basic research by assert- ing that the intention of action research is to solve an important problem for a client and not necessarily to test features of a theory (Coleman, 1972). We believe there is value in combining e studth f practicao y l problems with research that contributes theoro t y buildin testingd gan . Second, many action researchers understandably conduct their empirical work by following the current ideas about stan- dard scientific research. The dilemma is that some of the currently accepted ideas of rigorous research may be self-limiting. o attaiT ncertaia n leve f rigoro l methodologe th , becomy yma e so disconnected from the reality it is designed to understand that it is no longer useful. For example, the research that fol- lowed and built on the early studies of Lewin, Lippitt, and White (1939 n leadershio ) p style d grouan s p climate- in s wa s deed more rigorous, yet far less usable by human beings in real- life conditions (Argyris, 1980) tha originae nth l studies. i x Preface

Two of the best known exemplars of action science researchers as we understand them were Kurt Lewin and John Dewey. Both designe d executean d d actio r demonstratioo n n experiments whose consequences they studied systematically. Both were intereste addinn di fundamentao gt l knowledge while solving practical problems such as educating youngsters, influ- encing eating habits during World War II, or reeducating individuals about their prejudices. Dewe d Lewian y n were committe o notiondt f betteo s r societies or to what has recently been described as liberating alternatives n theiI . r worlds, citizens woul e heldb d responsible for becoming inquiry oriented in order to produce a society that was learning oriented and experimentally minded. This f organidescriptivo x mi c d normativan e e interests also charac- terized the great early social scientists such as Weber (Asplund, 1972). r vie f Ou actiowo n science e ideabuildth f thes o sn o s e early practitioners. We maintain that social science should have an important role in generating liberating alternatives. This objective canno e accomplisheb t d without challengin e statugth s quo. socian I l lifestatue th existo , squ s becaus norme eth d san rules learned through socialization have been internalized dan are continually reinforced. Human beings learn which skills work within the status quo and which do not work. The more the skills work, the more they influence individuals' sense of competence. Individuals dra sucn wo h skill justifd san y theie us r y identifyinb e valueth g s embedde n the i dd adherin man o t g these values. The interdependence among norms, rules, skills, and values creates a pattern called the status quo that becomes so omnipresent as to be taken for granted and to go unchal- lenged. Precisely because these patterns are taken for granted, precisely because these skills are automatic, precisely because value internalizede ar s individualsd an e statu th o , qu s ' personal responsibility for maintaining it cannot be studied without confronting it. orden I conduco t r t research that include e optioth s f no changing the status quo, one must have models of the status xii Preface differena f o d an t universo qu usee eb thacreato dn t ca t diaea - lectic .e intereste Thusar e w , n researci d h that generated an s tests propositions concernin e variableth ) g(1 s embeddee th n di statu thao qu st statue keeth t p i variablese quoth ) (2 ; s involved in changing the status quo and moving toward liberating alternatives; (3) the variables in a science of intervention that will be require e previouth f di s proposition e tested b evee ar so d t r an ; finally (4) the research methodology that will make change possible and simultaneously produce knowledge that meets rigorous tests of disconfirmability. In our discussions with social science colleagues on how o product e valid usablan d e knowledge e encountew , r several objections to research that attempts to alter the status quo. These objections raise valid concerns t thesbu , e concerne ar s often dealt wit wayn hi s thae counterproductivar t scienco t e e an practiceo dt . The first objection begins with a premise of normal science: the primary objective of science is to describe reality as accurately as possible. Hence, mainstream scientists focus on describin n changino e t Th e worlt no existi . th g git d s a dan s paradox is that this approach cannot describe many important features of the world as it exists. Among these features are the defensive routines that protec e statu th o against qu s t change. We will probably never get a valid description of the resiliency of defensive routine jusy b s t watchin d waitinggan . Som- de e fenses do not even surface until the first layer of defenses has been engaged (Argyris, 1985). A corollar e premisth o yt e thae purposth t f scienceo s ei o describt e realit s thai y t generating knowledge about change is a second step, one that must wait until basic descriptive knowledg s beeha e n accumulated n actioI . n scienc e agrew e e importans thai t i t understano t t chango t e ar worle e dth w f di e alsw o t believeBu . it s Kura , t Lewin said, tha opposite th t s ei bese th tf wayo true e understano t son : o t y tr e worl dth o t s di chang choosinn I . exploro t eit t gno e way f changino s stae gth - s quotu , researchers choos o perpetuatt e a worle whicn i d h there is little knowledge about the defensive routines that maintain the status quo. Preface xiii

A second objection is that defensive routines may be functional and hence should not be challenged. Defenses do serv o protect e t individual organizationd an s importann i s - re t spects. But what if such defenses are functional and dysfunc- e samtionath t ea l time r datOu ?a suggest that some defenses n significantlca y limi individual'n a t n organization'a d an s - ca s pacity to learn and adapt and hence to survive and flourish. To point to the positive aspects of defensive routines as reasons t studyin foo changno t r w eho g their negative - aspectit y ma s defensiva sel e b f e routine. A third and related objection is that attempts to change might get out of hand and unintentionally harm participants. This is an important concern, one that researchers must con- stantly respond to. But what leads researchers to believe that clients will allow them to create dangerous conditions? Our experience is that social scientists are successfully denied access by subjects who do not trust the researchers or do not agree with the research. We should add that our experience is based a mode a collaborativn o f o l e relationship between researcher d subjectan r clientsn whico i s e hon , client - n makin ca sn a e formed choice about proceeding with the research. To the degree that the researcher has unilateral control, subjects may be less able to protect themselves. The notion that clients have ways of protecting themselves leads to a fourth objection: the researcher could be kicked out. Confronting organizational defensive routinea n i s group could be dangerous. The group could unite and turn agains e researcheth ts tryin i o o discusgt wh r s issues thae th t group prefers to leave undiscussed. We agree that this is a dan- believe w t gerebu , tharesponse th t o withdrat e b nee t wdno from such studies. Some scientists should consider conducting research to illuminate under what conditions these dangers can be overcome. majoe th On f ro e contributions that action science makes to researchers is to help them develop the knowledge and skills needed to reduce the likelihood that they might unintentionally harm people or that participants might turn against them and to increase the clients' or subjects' commitment to research. The v xi Preface

knowledge require s relatedi o additionadt l mode f inquiryo s , methodw ne f research o se interpersona th d an , l skill o cont s - duct this research successfully. This represent sprimara y thrust bookr oou f . Several features of normal science, including intersuhjec- tively verifiable data, explicit inferences, disconfirmable propositions, and public testing, are also crucial to our approach. These features are designed to create challenging tests that may disconfirm our ideas. The criteria for validity must be rigorous because we are studying difficult, threatening issues that affect people's lives.

We have written this book with three purpose mindn si ; these correspon e threth eo de t book partth e firs f Th o s.t pur- posidentifo t s ei y primare somth f eo y issuephilosophe th n si y of science that relate to action science and have been discussed through the years. In Part One we describe the major positions taken by some of the key protagonists in this dialogue. We in- troduce our position and conclude with a statement of our the- oretical perspective. In doing so, we neither suggest that we have foun e answeth d theso t r e age-ol- im d questione w o d r sno y thae answerpl th t e providw s e completear e readee Th .- fa r miliar wit e literatur th he philosoph th n o e f sciencyo e knows that these issues have a long and distinguished history. We show wher believe ew e action science fitthin i s s dialogu ordeen i o rt stage th r furthe t efo se r inquir clarificationd yan . The second purpose is to identify similarities and differ- ences in the methodology of normal science and action science and to examine the implications of these for the skills that researcher e actioneey b o ma sdt n - scientistsex e w o Parn I . Tw t plore three research approaches use contemporarn i d y social science and compare them to action science. We identify the norms and rules that guide inquiry in each of these approaches, and we discuss how each may be self-limiting. We then describe e methodth f actioso n science designe overcomo dt e these limi- skille tationth sd thaan s t researcher sucse neeus h o dmethodst . These skills buil thosn do e that researchers have already learned e methodologith n y courses presently taugh mosn i t t universi- ties. Preface xv

e thir d probablTh an d y most important purpos f thio e s book is to show that a community of inquiry can be created in whic skille hth s neede conduco dt t action scienctaughte b n eca . Action science cannot become a science unless its skills can be made explicit and taught, so that successful action science research is more science than art. In Part Three we illustrate how we are teaching the skills of action science. Our approach is not necessarily the best, and we intend to continue our inquiry into modes of teaching action science skills. Our hope is to provide some guideline r thosfo s e researcher y wis o learma t h o n wh s an teaco dt h these skills and, more importantly wisy ma h o wh , o conduct t empirical researcthew ho y n migho h t teace th h skill youno st g researchers. Action Science is a product of genuine cooperation among the three authors. We designed and executed the book as equal partners. We acknowledg e hel th f eDianno p e Argyris, Donald Schon d Emilan , y Souvain readinn ei g e manuscript partth f o s . e greatlWar e y indebte o Marint d a Mihalakist onlno y o wh , typed and retyped chapters, but did so with speed, competence, and with a careful eye to what statements made or did not make sense. Marina is a great team member.

Cambridge, Massachusetts Chris Argyris August 1985 Robert Putnam Diana McLain Smith

Contents

Preface IX

Authore Th s xix

Part One: Designing a Science of Human Action 1

1. Philosophical and Methodological Issues 4

2. Action Science: Promoting Learning for Action and Change 36

0 8 . 3 Theorie f Actioso n

Part Two: Practices, Methods Resultd an , s of Normal Science and Action Science 103

4. Beyond the Limitations of Normal Science: Comparing Laboratory Experimentd san Action Experiments 105

XVII xviii Contents

5. Organizational Assessment Research: Filling in Gaps That Normal Science Overlooks 139

Ethnographie Th 6. c Approac Interventioo ht n and Fundamental Change 158 Dianne Argyris

7. The Social Scientist as Practitioner: Barrier Translatino st g Scientific Knowledge into Practical Knowledge 190

8. Practicing Action Science: Methods of Inquiry and Intervention 225

Part Three: Developing Skills for Useful Research and Effective Intervention 267

96 . 27 Engagin e Learningth g Process

10. Promoting Reflection and Experimentation 319

11. Expanding and Deepening the Learning 368

. Developin12 4 39 Framew gNe f Referencso e

References 451

7 46 Index The Authors

Chris Argyris is James Bryant Conant Professor of Education and Organizational Behavior at Harvard University. He was awarde e A.Bth d . degre n psychologi e y from Clark University (1947) e M.Ath ; . degre n economici e d psychologan s y from Kansas University (1949)Ph.De th d . an ;degre organizationan ei l behavior from Cornell University (1951). From 195 o 19711t , a facult s wa y e h membe t Yala r e University, servin s Beacga h Professo f Administrativo r e Science s chairpersoa d an e s th f no Administrative Sciences department durin e latteth g r parf o t this period. Argyris's early research focuse unintendee th n do d conse- quence r individualfo s f formao s l organizational structures, executive leadership, control systems, and management infor- mation systems—an individualw ho n o d s adapte o changt d e those consequences (Personality and Organization, 1957; Inte- grating e Individual e th Organization, th d an 1964) e theH . n turne s attentiohi d o wayt n f changino s g organizations, espe- ciall e behavioyth f executiveo r e uppeth t sa r level f organizao s - tion (Interpersonal Competence and Organizational Effective- ness, 1962; Organization Innovation,d an 1965).

xix xx The Authors

This line of inquiry led him to focus on the role of the social scientist as a researcher and interventionist (Intervention Theory Method,d an 1970; Inner Contradictions f Rigorouso Research, 1980). During the past decade he has also been developing, with Donald Schon theora , f individuayo organizad an l - tional learning in which human reasoning—not just behavior- become basie r diagnosith s fo s actiod an s n (Theory Practice,n i 1974; Increasing Leadership Effectiveness, 1976; Organiza- tional Learning, 1978). Argyris is currently working on a project that will relate e perspectivth e presente thin di se idea boo th f othe o so k t - rre searcher d practitionersan s . Argyri s earneha s d honorary doc- torates from the Stockholm School of Economics (1979), the University of Leuven, Belgium (1978), and McGill University (1977).

Robert Putna doctoraa ms i l studen instructon a d an t n ri counselin consultind gan g psycholog t Harvarya d Universitye H . received the A.B. degree with honors in political science from Syracuse University (1970) and was designated a Woodrow Wil- son Fellow. He studied political economy and government at Harvard University, servePeace th n dei Corp Costn si a Ricad an , worked as a counselor and training director at a human service agenc n Bostonyi e receiveH . e Ed.Mth d . degre counselinn ei g and consulting psychology from Harvard University (1980).

Diana McLain Smith is a doctoral student and teaching fellow in counseling and consulting psychology at Harvard Uni- versity. She received a B.A. degree with honors in political writ- ing from Boston University (1973a master d an ) s degren i e counselin consultind gan g psychology from Harvard University workes (1980) aree ha th f communite ao n di Sh . y mental health as an administrator and a psychotherapist and more recently as a consultant to organizations in both the private and public sec- tors.