Team Conflict, Integrative Conflict-Management Strategies, and Team Effectiveness: a Field Study

University of Tennessee, Knoxville TRACE: Tennessee Research and Creative Exchange

Doctoral Dissertations Graduate School

8-1999

Team Conflict, Integrative Conflict-Management Strategies, and Team Effectiveness: A Field Study

DonnaMaria Christina Vigil-King University of Tennessee, Knoxville

Follow this and additional works at: https://trace.tennessee.edu/utk_graddiss

Part of the Industrial and Organizational Psychology Commons

Recommended Citation Vigil-King, DonnaMaria Christina, "Team Conflict, Integrative Conflict-Management Strategies, and Team Effectiveness: A Field Study. " PhD diss., University of Tennessee, 1999. https://trace.tennessee.edu/utk_graddiss/4018

This Dissertation is brought to you for free and open access by the Graduate School at TRACE: Tennessee Research and Creative Exchange. It has been accepted for inclusion in Doctoral Dissertations by an authorized administrator of TRACE: Tennessee Research and Creative Exchange. For more information, please contact [email protected]. To the Graduate Council:

I am submitting herewith a dissertation written by DonnaMaria Christina Vigil-King entitled "Team Conflict, Integrative Conflict-Management Strategies, and Team Effectiveness: A Field Study." I have examined the final electronic copy of this dissertation for form and content and recommend that it be accepted in partial fulfillment of the equirr ements for the degree of Doctor of Philosophy, with a major in Industrial and Organizational Psychology.

Eric Sundstrom, Major Professor

We have read this dissertation and recommend its acceptance:

Joyce Russell, Dudley Dewhirst, Tom Ladd

Accepted for the Council:

Carolyn R. Hodges

Vice Provost and Dean of the Graduate School

(Original signatures are on file with official studentecor r ds.) To the GraduateCouncil:

I am submittingherewith a disseration written by DonnaMaria Christina Vigil-Kingentitled "TeamConflict, Conflict Management Strategies,and Team Effectiveness: A Field Study." I have examined the final copy of this dissertationfor form and content and recommend that it be accepted in partialfulfillment of the requirementsfor the degree of Doctor of Philosophy, with a major in Industrial/OrganizationalPsychology.

We have read this dissertation and recommend its acceptance:

Accepted forthe Council: �,(/ Associate Vice Chancellor and Dean of The Graduate School TEAMCONFLICT, INTEGRATIVECONFLICT-MANAGE1v.lENT STRATEGIES,AND TEAMEFFECTIVENESS: A FIELD STUDY

A Dissertation Presented forthe Doctor of Philosophy Degree The Universityof Tennessee,Knoxville

ii TO

Rosemary, Bill, and Neah, my beaconson a foggynight.

My father who took me to work before it was "cool. "

My God, my family, and the friendswho anchor me.

In Memoryof Dolores Vigil, My strongest supporter; I only wish you were here to see this in more than spirit.

iii ACKNOWLEDGMENTS

Thereare so many people who made this accomplishment possible. I would first

like to thank my wonderfulhusband, William A. Kingill, fornever allowing me to lose

sightof my goal. His unwaveringfaith in my abilities allowed me to believe in myself

even when the task seemed insurmountable. He is a constantsource of strength, love, and

devotion. Without him, I wouldn't smile as brightly, or laugh as loudly.

Thereare .many faculty, staff, and students at the Universityof Tennessee who had

a significantimpact on my graduatecareer. FirstI would like to thank my committee

members, Eric Sundstrom,Dudley Dewhirst, Tom Ladd, and Joyce Russell. I am forever gratefulto Eric Sundstrom. Since my firstyear in the programhe has stood by me, encouraged me, developed me, and challenged me. He has provided me with invaluable experiences and opportunities in both the research and the practice ofI/0 psychology. He is a great teacher, mentor, and friend. I owe much of my success to him.

I would also like to thank Dudley Dewhirst, who showed a genuine interest in this study when I firstproposed it. It was his initial interest that oftenkept me motivated while I was collecting data. Inaddition, Dudley's comments and suggestions throughout the. course of the project were insightfuland invaluable.

Tom Ladd deserves my gratitudefor holding me to a standard of excellence and foroffering support and counselwhen economic hardship almost necessitated me leaving the program. He gave me the "breathingroom" I needed to start a business that would support my family for several years.

iv I also cannotsay enoughabout Joyce Russell. From the beginning, she provided

me with the opportunityto collect meaningfulteam data. Without her support on my first

team research project, my dissertationwould never have come to :fruition. Her comments

andfeedback have always been insightfuland thought provoking.

I would like to thank members of the ManagementDepartment, the IvlBA

program, the MAcc program, and the Engineering program. Without the support of Gary

Dicer, BruceBehn, andElaine Seat,I would not have obtained a sample forthis study.

All threewere verygenerous with their support andencouragement.

SeveralIndustrial andOrganizational Psychology graduatestudents were also instrumentalin thesuccess of this project andI would like to thank them: Danielle

Adams, Kate Atchley, Chad Roedder, andDebrah Zegelbone Migetz. Additional faculty too numerous to list fromall of the mentioned programswere also instrumental in this study, as were the engineering facilitators. I would also like to thank the students of the above mentioned programsfor taking the time to complete surveys. Without their responses, there would have been no study.

In addition to my committeemembers and the variousparticipating departments, I would like to thankthe Department of Managementsecretaries, June Trbovich, Jackie

Cook, andCarolyn Alfred for all their support and encouragement. Their support, when it came time to distributegift certificates to the students, preventing me frompostponing the project. June is especially near anddear to my heart. From the moment I started applying to the Industrial andOrganizational Program until now, she has never wavered in her support. Her family is verylucky to have her, as is the 1/0"family."

V I would also like to thankAnn Lacava. Her willingnessto forgefor a solution that

would benefitall partiesinvolved, her tender and thoughtfulway of helping manage

stress, andher guidance in finalizingthis document are unforgettable.

Family andfriends have been my foundation. Without their support I could not

have succeeded or flourished. My heartfeltthanks to the friendsI met in graduateschool:

Danielle Adams, Joe Clark, LauraDavenport, Jackie DeMatteo,Lillian Eby, Stephen

Gaby, LauraGniatczyk, Todd Little, Sabine Maetzke, Paul andDebrah Migetz, Chad

Roedder, andDavid Vermillion. I also wantto thank those individuals that have been partof my foundationfor years: John andLyndia Dew, Cyndiand Randy Greenleaf,Alan andJanet Hecht, Jim andNorma Dar, Stephanie Myers, Steve, Karla, Eric, andMick

Polillo, PamThurman, andSteve Valdez. To those that have encouragedand mentored me: Bob Greenberg,Bob Maddox, Leigh Thompson,and Michael Passer many thanks.

Finally, I would like to express my greatestappreciation to my family. My sister

Rosemaryhas always made my life a joy. My brotherMatthew has always made my life a "hoot." My brother Danreminds me that success in lifeis a lot of luck. My fatherand

Marti, who have provided a sourceof comfort when things got rough. Adelaide Crain, who has always triedto do theright thing. And my extended family, who have helped me create fondmemories of laughter, roasting chiles, biscochitos, andshared superstitions.

vi ABSTRACT

A longitudinal fieldstudy examinedteam effectiveness, including both performanceand viability, in relation to teamconflict management strategies and three typesof teamconflict. Hypotheses predictedthat integrativeconflict-management strategieswould correlatewith teamperformance and viability, andthat these relationships would vary with the level andtype of team conflict. Teamsusing integrativeconflict managementstyles were expected to have higher performanceand viability thanteams using less integrativestyles. When higher levels of relationship conflictwere perceived by the team, the performanceand viability of teamsusing a more integrativestrategy would be higherthan teams using less a integrativestrategy. Finally, the performanceof teamsusing more integrative strategywould be higher thanteams usingless integrativestrategy, when higher levels of task conflict are perceived by the team.

Participants forthis study were 323 student members of 77 intact instructional teams. Individualmembers completed questionnairesmid-semester andlate semester.

Thequestionnaire was used to assess all variables expect performance: threeconflict managementstrategies - collaborative, compromising, andavoiding; threetypes of conflict- relationship, task, andprocess; andteam viability. Team performancewas assessed throughinstructors' grades, andwere standardizedwithin their respective courses. Participating teams were fromgraduate programsin business administrationand accounting, andundergraduate engineering andbusiness administration courses. From

50% to 100% of the students' gradeswere dependent on teamperformance.

vii Results showed that perceived relationship conflictand task conflictwere both significantlyand inversely related to team viability. Relationship conflictwas more

predictive of viability thantask conflict. The relationship between integrativeconflict managementstrategies, relationship conflictand viability was also significant. Similarly, when controllingfor the level of relationshipconflict or task conflict, integrativeconflict managementstrategies and performance were fowidto be significant. Theinteraction between relationship conflictand compromising conflict-managementstrategy was significantfor performance. hnplications are discussed.

viii TABLEOF CONTENTS

Chapter Page

1. INTRODUCTION...... 1 Typesof Conflict...... 3 ConflictTypolog)' ...... 5 Models of Group Effectiveness...... 8 Conflict-ManagementStrategies ...... 9 Hypoth.eses...... 13 Conflictan.d Viabilit y...... 13 Conflictand Performmce ...... 16 Conflict-ManagementStrategies ...... 19 Aggregationof Variables...... 21 2. l\1ETH0D...... 23 Research DesigI1...... 23 Participants...... 23 Time 1 Participants...... 28 Time 2 Participants...... 28 Procedure...... 29 Measures...... 32 ConflictStrategies ...... 3 2 ConflictMeasures ...... 34 Viability ...... 43 Individual-LevelVariables ...... 46 Group-Level Predictors andViability ...... 4 7 Performance...... 4 7 AggregatedVariables ...... 51 Power Analysis ...... 52

ix 3. RESULTS ...... 53 Data Analyses ...... 5 3 PreliminacyAnalyses ...... 54 Aggregation...... 57 Team Viability - Hypotltesis 1 ...... 57 Performance- Hypotltesis2 ...... 62 Tests of Accuracyof Assumptions...... 64 HierarchicalRegression - Hypotltesis3 ...... 67 Repeated Measures...... 7 5 4. DISCUSSION ...... 78 Limitations ...... 81 Generalizability ...... 81 MetltodVariance ...... 82 SampleSize ...... 82 Measurementof ConflictManagement...... 83 Measurementof ConflictTypes ...... 84 Measurementof Performance...... 84 Questions for FutureResearch ...... 86 Practical Implications ...... 90 Conclusion ...... 93 REFERENCES ...... 96 APPENDICES...... 108 APPENDIXA ...... 109 APPENDIX B...... 113 VITA ...... 125

X LIST OF TABLES

Table Page

2.1 Time 1 participant population demographicand return rate information...... 24 2.2 Summaryof performancemeasures at time 1 ...... 50 3.1 Individual-level descriptive statistics, tests formean differences, and test for equality of variancesfor participating students at time 1 ...... 56 3.2 Individual-levelstandardized descriptive statistics andzero-order correlationsfor all participants andreliabilities at time 1 ...... 58

3.3 Distributionof rwg for variables at time 1 ...... 59 3.4 Group-level descriptives forparticipating teamsat time 1 ...... 60 3.5 Group-level descriptive statistics forsub-group populations at time 1 ...... 61 3.6 Group-level descriptive statistics and zero-order correlationsfor all gro11psat time 1 ...... 63 3.7 Hierarchicalregression analysis for predictors of performancewith conflict- managementstrategies and task conflictat time 1 ...... 70 3.8 Hierarchical regression analysisfor predictors of performancewith conflict- management strategies andrelationship conflictat time 1 ...... 72 3.9 Correlations betweenconflict -managementstrategies and performance at tri-levelsof perceived relationship conflictat time 1 ...... 74 3 .10 Hierarchicalregression analysisfor predictors of viability withconflict- management strategies and relationship conflict at time 1 ...... 7 6 3.11 Summaryof supportfor hypotheses...... 77 A. I Factor analysis results andrevised scales measuring conflict-management styles ...... 110 A.2 Factor analysisresults and revised scales measuring the three typesof conflict...... 111

xi Table Page

A.3 Power an.alysis...... 112 B.l Time 2 participan.tpopulation an.dreturn rate information...... 114 B.2 Summaryof performance measures at time 2 ...... 115 B.3 Individual-level descriptive statistics, tests formean. differences, an.dtest forequality of varian.cesfor participating students at time 2 ...... 116 B.4 Individual-levelstandardized descriptive statisticsan.d zero-order correlationsfor all participan.tsan.d reliabilities at time 2 ...... 117

B.5 Distributionof rwg for variables at time2 ...... 118 B.6 Group-leveldescriptives for participatingteams at time 2 ...... 119 B. 7 Group-leveldescriptive statistics for sub-grouppopulations at time 2 ...... 120 B.8 Group-leveldescriptive statistics an.dzero-order correlationsfor all groupsat time 2 ...... 121 B.9 Hierarchical regressionanalysis for predictors of performan.cewith conflict- man.agementstrategies an.d task conflictat time 2 ...... 122 B.10 Hierarchicalregression analysis for predictors of performan.cewith conflict- man.agementstrategies and relationship conflictat time 2 ...... 123 B.11 Hierarchical regressionan.alysis for predictors of viability with conflict- man.agementstrategies an.d relationship conflictat time 2 ...... 124

xii LIST OF FIGURES

Figure Page

1.1 Framework forexamining conflict and team outcomes using conflict- managementstrategies ...... 4 1.2 Thomas' taxonomy of strategicintentions ...... 11 2.1 Itemsmeasuring the "collaborative" conflict-managementstyle ...... 35 2.2 Items measuring the "compromising" conflict-managementstyle ...... 36 . . . . 2. 3 Items measunngth e "avo1 d mg" co nfl1ct -managemen t s tyl e ...... 37 2.4 Instructionsfor completing the ROCI-Il...... 38 2.5 Items measuring task conflict...... 39 2.6 Items measuring relationship conflict...... 40 2. 7 Items measuringprocess conflict...... 41 2.8 Instructionsfor completing the items representing relationship, task, and process conflict...... 42 2.9 Items measuring team viability ...... 44 2.10 Instructions forcompleting group viability items ...... 45 3 .1 Samplescatterplot ...... 66

xiii CHAPTERl

INTRODUCTION

For the past several decades, the humanelement of organizationshas slowly shiftedfrom individuals performing single, complete, and intact assignmentsto groupings of individuals performinglarger, more complex tasks. Teams have become central to the structure of organizations (Hackman, 1990; Lawler, Mohrman, & Ledford, 1995). Nearly two decades ago, Jewell andReitz (1 981) observed: "A stronggroup of talented people can achieve what seems to beimpossible ... " A decade later, Gordon (1992) foundthat

82% of theU. S. organizationspolled had at least some teams. Sundstrom(1 999) postulated that by theyear 2000, most organizationswill use self-managedteams.

Conflictis common ingroups of all kinds, including work groups. For example,

Forsyth (1990) states, "groupconflict is as common as groupharmony'' (p.79). Stevens andCampion ( 1994) suggest thatmoderate levels of conflictmay be necessaryfor optimal team performance, and others have found that, when managed effectively, it can even be beneficial(Pruitt, 1981; Rahim, 1983; Thomas, 1990). Examples of positive outcomes of conflictin teams include increases in decision quality, strategicplanning, financialperformance, and organizationalgrowth (Bourgeois, 1985; Schweiger, Sandberg

& Rechner, 1989; Eisenhardt & Schoonhoven, 1990).

Conflictis a ubiquitous team dynamic, and it frequentlyfunctions as a catalyst for initiatingproblem-solving or discovering new and innovative solutions to complex situations (Pruitt, 1981). Even so, it is viewed as undesirableby some managers and employees (Losey, 1994; Stone, 1995). Previous research has indicateddetrimental effectsof conflicton performanceand satisfaction(Pondy, 1967; Blake & Mouton, 1984;

Gladstein, 1984; Wall & Nolan, 1986). However, more recent studies inthe team or work group context have foundconflict to be beneficial(Jehn, 1995, 1997; Amason and

Schweiger, 1994; Amason, 1996). For example,Jehn (1995) foundthat task-oriented conflictcan improve thequality of a group'sperformance. For these reasons, it is importantto continue investigating therole conflict plays in teamoutcomes.

Several researchers have suggested that conflictis an importantelement ingroup processes (Gladstein, 1984; Jewell & Reitz, 1981; Argote & McGrath, 1993). These processes are said to influencethe team's output(performance and effectiveness). Recent studies have begunto explore therole of conflict withinteams and its effect on performance (Jehn, 1995; Amason & Schweiger, 1994; Amason, 1996; Sessa, 1996,

Vigil-King & Rush, 1998). Prior to 1993, littleresearch had been done inthis arena, and the influenceof conflicton a team's level of outputhad been virtuallyignored (Argote &

McGrath, 1993).

Thepresent study examinesconflict within teams by investigatingthe integrative conflict-managementstyles usedby teams, andit expands on previous work examining the relationship between theuse of various conflict-managementstrategies and the team outcomes of performance andviability (Vigil-King & Rush, 1998). It incorporates the findingsof Jehn (1995, 1997) and her typologyof task, relationship, and process conflict.

Thetype of conflict-managementstrategy used by a groupis studied to determine it's

2 specificrelationship to the team's overall performanceand viability. Theframework

presented in Figure 1.1 servesas a guide.

This study expandson the existing researchin fourways. First, it attemptsto

examine conflict at the grouplevel instead of at the traditional individual or dyadic levels.

Second, this study is concernedwith cognitiveconflict tasks ( conflictsof viewpoints);

most research examiningconflict has focusedon mixed-motive tasks (conflicts of

interest). Third, it examines therelationship between integrativeconflict-management

strategiesand the groupoutcomes of viability andperformance. Finally, this study

explores threetypes of perceived conflict occurringwithin groupsand attempts to link

them to conflict-managementstrategies and group outcomes. To date, no published study

has examined the relationships between groupconflict, conflict-management strategies,

andgroup outcomes.

Theremainder of this chapter will review the typesof conflictand some of the models of group effectivenessidentified in the literature. A discussion of conflict managementstrategies as a tool forinvestigating conflict within groupswill follow, and the chapter will conclude by specifyingthe hypothesesfor this study.

Types of Conflict

To understandthe expected relationships between conflicttypes, conflict strategies, andgroup performance and viability, threetheoretical perspectives were considered. Thomas'(1 976) definitionof conflictand conflict-management strategies were integratedwith Jehn's (1997) conflicttypology and several input-process-output models of group effectiveness(Hackman & Morris, 1975; Gladstein, 1984;

3 Types of Conflict

Team Outcomes

Integrative Conflict-management Strategies

Figure 1. 1 Frameworkfor examining conflict and team outcomes using conflict managementstrategies.

4 Sundstrom& Altman, 1989). Thefollowing sections will firstpresent Jehn's conflict

typology,then briefly review groupeffectiveness models, and conclude with a discussion

of Thomas' conflict-managementstrategies and their applicability to thisstudy.

ConflictTypology

Jehn's (1997) conflicttypology provides a conceptual frameworkto explore three

types of conflict andtheir relationship to groupperformance and viability, andbuilds

upon thework of Guetzkow andGyr (1954), Wall andNolan (1 986), Priem andPrice

(1991), andPinkley (1 990). Guetzkow andGyr proposed two typesof conflict: affective conflict( conflictin interpersonalrelationships) and substantive conflict( conflict specificallyinvolving the group'stask). Wall andNolan expanded on theideas presented by Guetzkow andGyr and defined the types of conflictin termsof relationship-focused

"people conflicts" andsubstantive, content-oriented ''taskconflicts." More recently,

Priem andPrice suggested thatconflict emerging fromdisagreements indirectly related to thetask took on theforms of cognitive, task-related conflictsand social-emotional conflicts, while Pinkley uncovereda task-versus-relationship dimension while conducting a multidimensional scaling analysis of disputants' interpretationsof conflict. Jehn ( 1992) validated Pinkley's findingregarding the task-versus-relationship dimension and discovered that thesetwo types of conflictrelated differentlyto work groupoutcomes.

More recent work by Jehn(1 997) has foundevidence of a third type of conflict,

"process conflict,"in addition to task conflictand relationship conflict. Task conflict exists when group members disagreeabout the content of thetasks being performed; differencesmay be formulatedin termsof differingviewpoints, ideas, and/oropinions.

5 Displays of tension, animosity, and annoyanceamong groupmembers may signalthe

existence of relationship conflict, which involves interpersonal incompatibilitiesamong

group members. Over a 20-month period, Jehn observedsix organizational work teams

(two managementteams andfour production teams) in the international headquarters of a

household-goods-movingcompany and uncovered another type of conflict: process

conflict. Thisnewly identifiedconflict type focuses on how the work gets done, and

disagreementsabout assignmentsof duties or resources indicate its presence. To better understandprocess conflict, it is helpfulto examine McGrath's task circumplex, specificallythe negotiation quadrant.

McGrath (1984) provided a conceptual frameworkfor examining group-task performanceprocesses andclassified group tasks into fourgroup-performance functions : generatingalternatives, choosing between alternatives, negotiating conflict, andexecuting the behaviors necessaryto accomplish the agreed-upongoals. The function of interest in this study is the negotiation function, which can be furtherdivided intotwo task types, cognitiveconflict tasks (concerned with resolving conflictsof viewpoints)and mixed motive tasks (concernedwith resolving conflicts of interest). Mixed-motive tasks have beenthe most frequently-studied tasks (Levine& Moreland, 1990; Carnevale& Pruitt,

1992).

Cognitive conflictoccurs when members are in agreementas to the goal, task, or mission, but have determinedmultiple and competing means to the end. Mixed-motive conflictresults when two or more partiesor groupsof partieshave competing goals.

Jehn's (1995) task conflictis similar to cognitiveconflict, while her process conflictis

6 similarto mixed-motive conflictand other constructssuch as distributiveconflict

(Kabanoff, 1991) or proceduralcomplexity (Kramer, 1991). Process conflictis concernedwith team regulations or guidelines which governindividual team members' roles and responsibilities andthe allocation, distribution, andexchange of resources.

Relationship conflict, task conflict, and process conflictform the foundationof

Jehn's (1997) model of groupconflict and performance. Inaddition to establishing a new typologyof conflict, she proposes the existence of fourconflict dimensions which are each applicable to .all three typesof conflict: resolution potential, acceptability norms, emotionality, andimportance. Thesefour conflict dimensions moderate the three types of conflict and their relationship to group performance, andJehn suggests that the optimal profilefor a high-performanceteam includes "moderate task conflict, no relationship conflict, littleor no proceduralconflict, with norms that conflictis acceptable, perceptions thatconflict is resolvable, andwith little emotionality" (p. 552).

Jehn's identificationof the conflict dimension moderatorsis extremelysignificant, but the focus of thecurrent study is on conflict-managementstyles andtheir relationship to team performanceand viability, while exploring thetypes of conflictperceived by the team. Thispaper will expandon recent conflict researchexamining cognitive conflictat thegroup level by combining Jehn's (1997) typologyof conflictand Vigil-King and

Rush's (1998) work examining conflict-management styles and team performanceand viability. For this study, groupsor teams aredefined as two or more people who share responsibility foraccomplishing specificgroup goals.

7 Models of Group Effectiveness

Using Jehn's (1995) conflict typologyas a basic framework, conflictin a group

setting can best be examined using one of several models of groupeffectiveness which

have been presented in theliterature. Representativeof the current models are those of

Gladstein (1984), Hackmanand Morris (1 975), and Sundstromand Altman (1 989). All of

these models are variations of an input-process-outputmodel.

Gladstein's (1984) model is usefulbecause of its comprehensive nature and

because it builds on previous work by Hackman andMorris ( 197 5), expandingon their

model of group effectiveness by definingthe group interaction process in more detail.

Her model has been empirically tested on a largesample of work groups. Briefly, the

inputs of the model are at the grouplevel (groupcomposition and groupstructure) and at the organizationallevel (resources availableand organizational structure). Open communications, supportiveness,conflict, discussionof strategy,and weighing of individual inputsare included in the group processportion, which is definedas those behaviors which "build, strengthen, andregulate group life"and assist the groupin resolving the problems to which they are committed (Bales, 1958). Group effectiveness is theoutput portionof themodel, whichincludes groupperf ormanceand satisfaction of the groupmembers' needs.

Swidstromand Altman 's (1989) model is also a modifiedversion of the basic input-process-outputmodel, andtakes anecological perspective in analyzing work-team effectiveness. Organizational context andgroup bowidaries are the inputsof this model.

Team development equates to groupprocesses, andis inclusive of what has been

8 considered groupstructure and interpersonalprocesses. Theoutput is team effectiveness.

Incontrast to Gladstein (1984), Sundstrom,De Meuse, andFutrell (1990) identifygroup outputsas group performance and viability. Viability includes both the satisfaction of the groupmembers' needsand the group 's ability to exist over time.

All of these models of work-team effectiveness are input-process-outputmodel variations. This study incorporates these variations by focusingon conflict (identified by

Gladstein (1984) as a component of group process) andperformance and viability

(identifiedby Sundstromand Altman (1 989) as outputcomponents). Intheir review of groupprocesses in organizations,Argote andMcGrath (1993) stressedthat little research has examined the influenceof conflicton grouplevels of output. More recent work by

Jehn (1995) examined therelationship between conflictand satisfaction, liking, and intent to remainin thegroup at theindividual level, andthe relationshipbetween conflictand production at the group level. Within top management teams, Amason ( 1996) explored the impact of cognitiveconflict (task-based) and affectiveconflict (relationship-based) on strategic decisionquality. Vigil-King and Rush (1998) investigated the relationship between conflict-managementstrategies and performanceand viability, both at the group level. Thus,conflic t-managementstr ategies(which will be reviewed next) will be used to examine the influenceof task conflict, relationship conflict, and process conflicton group levels of output.

Conflict-Management Strategies

Thisstudy examines informalconflict within a groupand the influenceof the three types of conflict on thegroup' s overall performanceand viability, based on the

9 assumptionthat the conflict-managementstrategy used by the groupreflects their internal

group processes. Subjects' conflict-managementbehaviors were examined using

Thomas' (19 79) taxonomy of strategicintentions, shown in Figure 1.2. Thistaxonomy

definesan individual's decision to act in a specificway as an indicant of their intention

and shows that an intention is the culmination of the individual's thoughts and emotions

aboutthe given circumstances. According to Thomas' taxonomy, ''theparty' s strategic

intentions are classified and plottedalong two basic orthogonal dimensions of intent"

(1979, p. 667). Thesedimensions represent the extent to which a partyattempts to satisfy

their own concerns(the assertiveness dimension) andthe extent to which a partyattempts

to satisfythe concernsof the other party(the cooperativeness dimension). Within these axes are fivestrategic intentions:

1) A voiding: An attemptby a party not to satisfytheir own concernsor the concernsof others. Inthe literature, thisintention is also referredto as withdrawing, lose-lose, inaction, fatalistic, or isolation.

2) Accommodating: A partyattempts to satisfythe other party's concernswhile neglectingtheir own. Thisintention has also been referredto as smoothing over, yielding-losing, obliging, yielding,and lose-win.

3) Compromising: An attemptby a partyto satisfysome of their own concerns and some of the other party's concerns. Splitting the differenceor sharing areother common ways to refer to this intention.

10 Comp eting Collaborating

Integrative Axis

Distributh·e.Axis

A�"aiding Accomodating

Concernfar Others

Figure 1.2 Thomas' taxonomy of strategic intentions

11 4) Competing: An attemptby a partyto satisfytheir own concernswithout regard

for the otherparty's concerns. Otherterms for this intention include win-lose,

dominating, andcontending.

5) Collaborating: One partyis attemptingto satisfytheir concernsand the

concernsof the other party. Thegoal is anintegrative solution. Problem solving,

synergy, win-win, andintegrating are other termsfor thisintention.

Thomas(1 979) theorized that the fiveconflict-management styles lie on two

diagonals: theintegrative (major) diagonal, where the interests of both partiesare taken

into consideration when anagreement is made; andthe distributive(minor) diagonal,

where each party only looks at theirown interests. A voiding andcollaborating conflict

management stylesrepresent polar ends of the integrativedimension, while competing

andaccommodating represent polarends of thedistributive dimension. Compromising is

centered at the intersection of the diagonals andat the centers of the axes.

Thediagonals of Thomas' taxonomy have some similarities with cognitive

conflictand mixed-motive conflict. Integrativeconflict-management strategies are more

suited to the resolution of cognitiveconflict, in which team members agree as to what

their goal, task, or mission is but have differentideas about how to accomplish them. On the other hand,distributive conflict-management strategies are more related to mixed motive conflict,which involves competing goals. Most definitionsof teams incorporate the idea that the members are working towards a common goal (McGrath, 1984;

Campion, Medsker, andHiggs, 1993), so thisstudy focusedon cognitiveconflict and the integrativeconflict-management strategies that are suited to its resolution.

12 Previous work by Vigil-Kingand Rush (1998) explored how a group's commonly-usedintegrative conflict-management strategy is related to the group's performanceand viability. Using aninput-process-output model, they examined the relationshipbetween integrativeconflict-management strategies, group conflict, and groupperformance and viability. They found that a more integrativeconflict managementstrategy was associated with higher levels of groupviability and performance. Thisstudy will expandon Vigil-King andRus h's previous work by incorporating, in place of the single measure of conflict,the conflicttypology identified by Jehn (1997).

Hypotheses

Jewell andReitz (1981) discusstwo obstacles facedby newly-formedgroups and teams which can lead to conflict. Thefirst obstacle is uncertaintyand disagreement over power andauthority, andthe second is uncertaintyand disagreement over interpersonal relations. Conflictbetween team members over goals andthe meansused to achieve them is common (task conflict), as is conflictconcerning issues of groupstructure

(process conflict)or interpersonalrelationships ( relationship conflict). Regardless of the source, it is importantto investigate thethree identifiedtypes of conflictand their differing relationships to the group organizational outcomes of viability and performance.

Conflictand Viability

Building on the findingsof Surraand Longstreth (1 990), Walton andDutton

(1969), Peterson (1983), and Ross (1989), Jehn (1995) showed that an individual's level of satisfaction,liking of team members, andintent to remain in the teamwould be

13 negatively related to theindividual 's perceived level of relationship conflict. Viability,

which considers theteam 's capacity to continue operating togetherin the future, includes

member satisfaction, member participation, and capacityfor futureinteraction, and is

similar to thevariables examined by Jehn at the individual level. Thus, interpersonal

tensionand frustrationresulting from members ' feelingsof dissatisfaction,dislike, and

other negative affectiveresponses should be negativelyrelated to a team's viability

(measuredat thegroup level).

In addition, intense task andprocess conflictmay also cause tension, animosity,

and annoyanceamong group members. An individual's normalreaction to any

disagreementor questioning is frustration, dissatisfaction, andanxiety (Ross, 1989), and

within top managementteams, task conflictwas foundto lead to frustration and

dissatisfaction(Amason & Schweiger, 1994). Jehn (1995) foundthat higher levels of

perceived task conflictwere related to decreases in team member satisfactionand intent

to remain in the group. Conversely, members in groupswith hightask consensus have

expressed more satisfactionand desire to remain in the groupthan members in groups

with high task dissension (Schweiger, Sandberg, andRagan, 19 86).

Ambiguity and disagreements over roles andresources are :frequent sources of process conflictamong groupmembers (Jehn, 1997). Gladstein's (1984) input-process outputmodel would suggest that this formof conflict occurs at the group level (group structure)and at the organizationallevel (resources available), andthat these variables referto the degreeto which employee behavior is specifiedby routines, procedures, and prescribed roles (Pugh, Hickson,Hinings, and Turner, 19 68; Kiesler, 1978). Specifically,

14 normsand role clarity have been identifiedby Gladstein ( 1984) as components of group structure. Normsare rules forbehavior developed andused by the groupregarding the proper division of labor or activities ( e.g., "who should do what"). Behavior norms among team members are often discussed in termsof roles, role expectations, androle systems, androle expectations are normsthat specify"what should be done, andwho should do what, when andhow" (McGrath, 1984, p. 201). Role problems may be the consequence of unclearrole messages (role ambiguity) that may require differentand mutuallyconflicting actions (role conflict)and be beyond thecapabilities of the target individual(role overload).

At the individual level, Jackson andSc huler (1985) fowidthat role ambiguity and role conflictwere associated with lower levels of job satisfactionand higher levels of propensity to leave anorganization, as did Netemeyer, Johnston, andBurton (1990). At the grouplevel, Campion, Medsker, andHiggs (1993) fowid that components of "Job

Design"and "Context" (which map onto Gladstein's "Group Structure" and"Resources

Available," respectively) were correlatedwith employee satisfaction. In addition, process conflictcan occur when resources are fixed and limited. Theinvolved partiesmay desire to maximize their ownindividual gains (Walton & McKersie, 1965), resulting in outcomes which arenot favorableto all involved parties. This win-lose outcome of a distributivesituation or mixed-motive task canalso igniteprocess conflict,and is not recommendedif one desires to maintain a relationship in the futurewith the other party

(Greenhalgh, 1986; Lax& Sebnius, 1986).

15 Based on the above discussion, thefollowing four empirically-based hypotheses

were developed to investigate the relationships between the threetypes of conflict

identifiedby Jehn (1997) and their relationships withviability:

Hypothesis la: Perceived relationship conflictwill be related to team viability.

Hypothesis lb: Perceived task conflictwill be related to team viability.

Hypothesis le: Perceived process conflictwill be related to team viability.

Hypothesis ld: Relationship conflictwill be more predictive of the group's

viability than task conflictor process conflict.

Conflictand Performance Thebenefits of task conflictare numerous. Janis (1982) discovered that task

conflict decreased the groupthinkphenomenon by increasing the flowof ideas, criticism, and alternativesolutions, and Baron ( 1991) fowidthat task conflict within a group encouragedits members to develop new ideas. Putnam(1 994) determined thatissues were better widerstood and identifiedwhen task conflictwas present. Longitudinal research by Fiol (1994) demonstratedthat learningand accurate assessment of situations improved when groupmembers viewed the issues surroundingthe task's content differently. Withinthe decision-making research, Schwenk andValacich (1994) found that work groupsprovided higher qualitydecisions when they challenged the status quo, andAmason ( 1996) foundthat top management teams experiencing task conflictalso produced higher-qualitydecisions andhad a clearer understandingof these decisions. In short, conformity, complacency, and lack of understandingand innovation canbe the by products of too little task conflict.

16 Several researchers have suggested that there are optimal levels of task conflict, with too littleconflict resulting in low-qualityoutputs and too much conflictproducing low-quality outcomes or no outcomes (Pondy, 1967; Brown, 1983). Van de Vliertand

De Dreu ( 1994) found that conflictstimu lation enhancesperformance by stimulating activity, while too littleconflict can lead to a lack of urgency and inactivity. However,

Gersick (1989) found that groupsthat continued to hammerat issues well past the point of effectivenessand failed to reach consensus were unable to advanceto the next stage of fruitfulwork. Jehn (1995) unearthedsupport for a curvilinear relationship betweentask conflictand groupperformance, which was assessed using supervisorand production reports obtained at the grouplevel. Her analyses showed an optimal level of task conflict in non-routine-task groups.

Unlike task conflict, relationship conflicthas been found have a less thanpositive relationship to productivity and satisfactionin groups(Evan, 1965; Gladstein, 1984; Wall andNolan, 1986). Evan found that interpersonalattacks were negatively related to group performanceand productivity. Specifically, underconditions of relationship conflict, group resourceswere focusedon ignoringor resolving the interpersonal conflictsrather than on performingthe task. Baron ( 1991) foundthat when the components of interpersonalconflicts included anger or frustration, communicationand cooperation among group members were limited. In a studyof the ways that relationship conflict is associated witha group'sperformance, Pelled (1995) foundthree: 1) groupmembers are less able to consider new informationfrom other members due to more limited cognitive processing; 2) groupmembers are less open to ideas fromother group members; and

17 3) the group uses their resources working on resolving or ignoringthe conflictsinstead of

working on thetask at hand.

As mentioned before,process conflicthas its roots inthe variables of group

structureand resources available, as identifiedin Gladstein's (1984) input-process-output

model. Role issues such as role ambiguity, role conflict, and role overload, andresources

issues such as managerialsupport and training can spur process conflict. Role expectations (perceivedas role pressures) have also been foundto have anadverse relationship with team performance (Ross& Starke, 198 1; Katz& Kahn, 1978).

Campion, Medsker, andHiggs (1 993) foundthat managersrated teams with greaterjob assignmentflexibility as more effective. Processconflict resulting fromfixed or limited resources will also decrease performanceby creatinga distributive situation that pits one party's desires against another's. Researchexamining integrative and distributivebargaining has also foundthat more integrativesolutions generally improve the performanceof all partiesinvolved (Thompson, 1990). Thus, it is expected that process conflictwill decrease performance.

Groundedin the discussion above, the followinghypotheses were developed to explore the relationships between J ehn's ( 1997) three typesof conflictand their relationships with groupperformance:

Hypothesis2a : Teams perceiving moderate task conflict will have higherlevels of

performancethan teams perceivinghigh or low task conflict.

Hypothesis 2b: Perceived relationship conflictwill be related to team performance.

Hypothesis2c: Perceived process conflict will be related to team performance.

18 Hypothesis 2d: Task conflictwill be more predictive of the group's performance

thanrelationship conflictor process conflict.

Conflict-Management Strategies

Thethird set of hypothesesextends the firsttwo sets to include Thomas' conflict

managementstyles. According to Gladstein (1984) andArgote and McGrath(1 993),

conflictis a necessarycomponent of team development, andwhen teams learnto deal

with conflictthey can better focustheir energy on the group's tasks. Specifically, the

management of a team's conflicttakes the formof one of Thomas' five conflict

management styles, and integrative conflict-managementstrategies allow foran increase

in satisfactionand performance(Thompson, 1990). For example, when the negotiation

outcome was measured by the numberof points accumulatedduring a negotiation

exercise, it was foundthat the negotiator using a more integrativeconflict strategy

amassed the greaternumber of points (higher level of performance)(Pruitt & Rubin,

1986). Greater satisfactionhas also been associated with integrativetactics (Wall &

Nolan, 1986). Since satisfactionis an element of viability, it followsthat the level of

viability forthe groupwill be higherwhen integrativetactics are used.

Toe performanceand viability of teams using more integrativeconflict management strategiesshould also be higher than that of teams using less integrative conflict-managementstrategies, when higher levels of conflictare perceived by the team.

A previous test of these hypotheseslacked sufficientpower ( small number of teams), but marginally significantresults suggested furtherinvestigation was warranted(Vigil-King

& Rush, 1998). The currentstudy examined the relationships between performance, task

19 conflict, andthe use of integrativestrategies; between performance,relationship conflict,

andthe use of integrativestrategies; and between viability, relationship conflict, andthe

use of integrativestrategies. (Insitua tions where the team perceives low levels of

conflict, the team'schoice of a conflict-management style is not expected to be related to their performance or viability as much, because there is littleconflict to be managed.)

Hypotheses3a and3b address the relationship between conflict-managementstrategies and performance andviability. Hypotheses 3c, 3d, and3e are extensions of the first two hypotheses andreflect the interaction of conflict with conflict-managementstrategies:

Hypothesis 3a: Teams usinga more integrativestrategy will have higher

· performancethan teams using a less integrative strategy.

Hypothesis 3b: Teams using a more integrativestrategy will have higher group

viability than teams using a less integrative strategy.

Hypothesis 3c: The performanceof teams will be aninteractive functionof

conflict-managementstrategies andtask conflict, such that the

performanceof teams using a more integrativestrategy will be

higherthan teams using a less integrative strategy,when higher

levels of taskconflict are perceived by theteam.

Hypothesis 3d: Theperformance of teams will be an interactive functionof

conflict-managementstrategies and relationship conflict, such that

theperformance of teams using a more integrativestrategy will be

higher than teams using a less integrative strategy, when higher

levels of relationship conflictare perceived by the team.

20 Hypothesis 3e: The viability of teams will be an interactive functionof conflict

managementstrategies and relationship conflict, such that the

viability of teams using a more integrativestrategy will be higher

than teams using a less integrativestrategy, when higherlevels of

relationship conflictare perceived by the team.

Aggregation of Variables

Conceptually, the aggregation of individuals' perceptions of the team's level of

conflict,viability, andgroup conflict-management strategies to the grouplevel has to be

based on a sowidrationale anda construct which is equally meaningfulwhen analyzed at

a higherlevel (Roberts, Hulin & Rousseau, 1978; James, 1982). The items used in this

study to assess the level of conflictperceived by the team and the viability of the team directly addressed the relationship that group members had with one another, and the scales measuring conflict-managementstrategies addressed the way team members interacted with one another. Thus, the scalesmeet thefirst criterionto justify aggregation. Furtherconceptual justificationexists because the constructsof interest represent situations commonlyperceived by all group members; it is assumed that group members experiencing the same situation will describe it in similar ways (Jones & James,

1979). The participating teams were all performing similar grouptasks widergiven performance criteria, so this lends furthercredence to the decision to aggregate the data.

Inaddition to these conceptual justifications, there must also be statistical justificationto aggregate. Schneiderand Bowen (1985) suggested that a statistical test of

21 thehomogeneity within groupsis preferredover intraclasscorrelation coefficient indexes

(which require between-groupdifferences) when within-group agreement is desired. Past research conducted by George (1990) and George and Bettenhausen (1990) supportsthis, andJames, Demaree, andWolfs (1984) measure of within-group agreement(rwg(j)) for a multiple-item estimator (i.e., scale) was used to test for the homogeneity within groups.

Ther w g statistic is anindex of interrateragreement, andtests the proportion of systematic variation in judgmentsin relationship to the total variancein judgmentswithin each group. Provided that the scale to be examined has adequateinternal consistency

(coefficient alpha of . 70 andabove), the j items areassumed to be parallel indicators of the same construct. Kozlowski andHattrup (1992) interpreted the rwg statistic as an indicator of perceptual/statistical convergence, and values of . 70 or higher were generally viewed as indicative of agreementamong team members (George, 1990).

22 CHAPTER 2

METHOD

Research Design A fieldstudy examined 77 intact instructionalteams in I -semester university courses. The studentteam members were involved in team projects as partof course requirements. Members of participatingteams completed questionnairesat 2 times during thesemester - at week 8 to 10 andagain at week 14 to 16 - to provide assessments of team conflict,conflict management strategies,team viability, andteam performance.

(Thefirst survey distribution will be referredto as Time 1 andthe second survey distributionwill be referredto as Time 2.) Course instructors provided independent assessments of team performance. Participatingteams came fromthe Masters of

Business Administration(MBA) program, the Masters of Accountancy(MAcc) program, and undergraduateengineering and business administration programs. Thestudents participatedin the study voluntarily, and were provided giftcertificates or extracredit in return for their participation.

Participants A total of 323 student members of 77 intact instructionalteams lasting for one semester participatedin thestudy. Teamsconsisted of 3 to 12 students (mean = 4.26, standard deviation = 1.24). Table 2.1 provides informationregarding thenumber of

23 Table 2.1. Time 1 participantpopulation demographic andreturn rateinformation

Fall Spring Timein semester Gender A!e Ind Teams RR Ind Teams RR when measured o/oMale Mean SD MBA 77 17 84% 72 15 84% 8-10 weeks 72 25.4 3.54

MAcc 52 11 65% 8-10 weeks 43 23.9 6.28

Engineers 72 18 92% 8-lOweeks 91 23.7 3.66

Undergrad 1 26 5 100% 4 weeks 58 22.8 2.49

Undergrad2 19 5 83% 8-10 weeks 61 22.1 .94

Undergrad 3 22 5 85% 4 weeks 50 23.2 2.51

Totals 129 28 75% 211 48 88% 68 24.3 3.71

24 individuals and thenwnber of teams they comprised, the returnrate for each survey,and

the age and gender of thoseparticipating. 1

Therewas a varietyof approaches to teamwork used amongthe various

disciplines fromwhich the participant population was drawn. Thestudent participants

were involved in teamsthat were much like organizationalwork teams. Theywere

expected to develop and implement theirown strategies, andto create and present

multiple presentations to both facultyand external companies. Insome cases, they even

had to contact external vendors to obtain equipment. Between 50 - 100% of the participants' gradeswere dependent on team performance. Team projects spannedthe course of the semester, rangingfrom a single project to as manyas 7 projects.

:MBATeams. :M:BAstudents comprised two separate sets of participating teams, one during the fall semester and one during the spring semester. Prior to the onset of classes, students were divided into teams of fourto six members by the program coordinator, yielding 18 teams. To maximize the breadth of task-relevantexpertise, the programcoordinator took into consideration each individual's workbackground, expertise, andma jor concentration. Personalitymeasures and biographical data were also used to assist theprogram coordinator in shaping diverse teams.

Theabove method of dividing the MBAstudents into differentteams was followedfor spring semester as well. :MBA students are assignedto differentteams at the beginningof each new semester, so the :MBAteams surveyedduring the fall semester were not the same teams as those surveyed in the spring semester. Students were divided

1 See AppendixB, Table B.1, forTime 2 participantpopulation and return rate information. 25 into new teams of fouror fivestudents by the programcoordinator, yielding 17 teams.

(Attritionresulted fromstudents leaving the l\IBA programin both fall andspring

semesters).

:MBAcourse work was team-based, andthe teams worked together on the same

assignedprojects throughout the 16-week semester. Approximately 60% of the work

required of the :MBAstudents was team-based. 92 students were solicited in the falland

86 students were solicited in the spring.

MAcc Teams. Prior to the onset of classes, students were divided into teams of three or four members by the programcoordinator, yielding 20 teams. Biographicaldata was used to assist the programcoordinator in maximizingthe breadth of task-relevant expertise. Theprogram coordinator took into consideration each individual's work background,expertise, andtype of accountingspecialty when shapingdiverse teams.

MAcc coursework was team-based, andthe teams worked together on the same assignedprojects throughout the 16-week semester. Approximately 50% of the work required of the MAcc students was team-based. 80 students were solicited.

Engineering Teams. Courseinstructors assembled upper-level undergraduate students fromfour engineeringcourses into teams rangingin size from3 to 12 members:

7 Chemical Engineeringteams with 3 students per team; 12 Mechanical Engineering teams with 3 or 4 students per team; 1 Fluid Engineering Team with 5 students; and 1

Automotive Engineeringteam with 12 students. TheChemical Engineering and

Mechanical Engineeringcourse instructors randomlydivided the students into teams.

Students on the Fluid Engineeringand Automotive Engineering teams volunteeredfor

26 these projects, andthe projects lasted for the entiresemester. The Mechanical

Engineering teamseach had their own designatedproject to complete by the semester's end. The Chemical Engineeringstudents worked on fivedifferent projects during the semester, but differentgroups worked on the same projects at differingtimes. A total of twenty-one engineering teams,comprised of 78 students, were solicited.

One hundredpercent of the work required of the Engineering students was team based, andthe teamsworked together on assignedprojects throughoutthe 16-week semester. Depending on the instructor,between 50% and 100% of the students' grades were dependent on the outcome of their team project. Peer appraisals andinstructor ratingsof each individual determined the remainder of students' gradesfor teamshaving less than 100% of their course gradedependent on the team project.

TheAutomotive Engineering team stands out fromthe rest because it was the largest of all theteams surveyed, with 12 team members. Although it was much larger thanthe others, thisteam was included because all team members feltthat they were (and acted as) a single team. Theteam was tasked withthe responsibilityof creating a high performancecar to runin a race. Facultymembers andengineering students viewed and treatedthis team as similar to a high-performing sportsteam (i.e., a footballteam, a baseball team, the pit crew for a race car, etc.).

Business Teams. Upper-level undergraduatestudents were solicited from2 business courses, anddivided into teams of 5 or 6 members by the course instructors.

One instructorhad students working in thesame team for 16 weeks, yielding fiveteams.

Theother instructorhad students working in one team for the first8 weeks of the

27 semester. Students were then assignedto differentteams forthe last 8 weeks of the

semester. Regardlessof the instructor,teams withineach course completed thesame

projects. A total of fifteen teamsresulted fromthese courses. Approximately 50% of the

work required of the upper-level lllldergraduatebusiness students was team-based, and 49

students were solicited.

Time 1 Participants

Thefinal set of 323 surveys comprised 77 teams (32 :MBAteams, with a total of

143 student responses; 12 MAccteams, with a total of 42 student responses; 18

engineering teams, with a total of 71 student responses; and 15 tllldergraduate business

teams,with a total 67 student responses). Whenanalyzing data at the team level, 17

surveyswere eliminated because informationused to identifyteam membership was missing or because fewerthan three members of a given team completed thesurvey.

Time 2 Participants

Thefinal setof 270 surveys comprised65 teams (28 l\IBAteams, witha total of

126 student responses; 13 MAccteams, witha total of 4 7 student responses; 16 engineering teams, with a total of 64 student responses; and 8 tllldergraduatebusiness teams, witha total of33 student responses). Again, when analyzingdata at the team level, thirty-seven surveyswere eliminated because informationused to identifyteam membership was missing or because fewer thanthree members of a given team completed the survey.

28 Procedure

Participantsin this study were solicited in a variety of ways. Student participation

was petitioned fromthe MBA and MAcc programsthrough the respective department

deans. The undergraduatebusiness students were solicited throughtheir instructors,who

were graduate-studentinstructors from the Managementdepartment. Theengineering

student participantswere obtained throughthe Engineering department'steam facilitator.

The solicitation of the students was performedby arrangementwith the above

mentioned contacts. The students were given a brief explanationof the surveysthey

would be completing andthe incentives that would be available to them forparticipation.

At the beginning of courseworkand then periodically duringthe semester, thesurveys

were distributed during class visits, along with a reminder of the incentives for

participation. Furtherreminders were sent via mass e-mails, delivered throughthe deans

of theMBA and MAcc programs,the graduate-student instructors, andthe engineering

departmentteams facilitator. Thesurveys were returnedto secretaries in the Management

departmentor via campus mail in sealed envelopes provided to the participants.

Incentives were distributed at theend of each semester by Managementdepartment secretaries.

Incentives. Students were provided various incentives for participating. Any

MBA or MAcc student returning all surveysby the designateddeadlines received a $10 giftcertificate to Border's Bookstores. Thesestudents had the opportunityto earn an additional $5 giftcertificate if they andtheir teammates returnedall surveysby the designateddeadlines. Inaddition, teams that fullyparticipated in all surveyswere entered

29 into a drawingfor an additional giftcertificate to Border's Bookstores. Members of the winningteam each received an additional giftcertificate valued at $25. This procedure was followedagain duringthe spring semester forthe MBAstudents.

Engineering students were more motivated by food thanbooks. Any Engineering student returningall surveysby the designateddeadlines was given a $10 giftcertificate to Calhoun's restaurant. Students had the opportunityto earn anadditional $5 gift certificateif they and their teammatesreturn all the surveysby the designated deadlines.

Theundergraduate business instructors providedextra credit to participating students upon completion of the surveys. As anextra incentive, teams that fully participatedhad an opportunityto enter a drawing fora gift certificateto Border's

Bookstore. Members of the winningteam each received a giftcertificate valued at $15.

Studentswho were in two differentteams during the semester had two opportunitiesfor extra credit andan opportunity for a chancein both drawings.

Students completed two survey measuresduring the semester: approximately mid semester (Time 1; during weeks 8 through 10) and nearthe conclusion of the semester

(Time 2; duringweeks 14 through 16). Measures collected fromMBA students, MAcc students, undergraduate businessstudents, and half of the engineeringstudents corresponded with the completion of a major team project. Theremaining engineering students workedcontinuously throughout the semester on a single project, andonly the second surveycorresponded to the completion of a groupproject.

Several sets of participantswere measuredtwice. During spring semester, MBA students were again measured aftercompleting each of two milestones (major team

30 projects). Theseprojects occurredduring mid-semester and at theend of the semester.

(As a reminder, :MBAstudents are assignedto different teams at the beginningof each

new semester, so the MBAteams surveyedduring the fallsemester were not thesame

teamsas those surveyedin the spring semester.) Students fromthe second widergraduate

business course were measured during the fourthweek (mid-point of teamwork) and

eighth weekof classes (later part of team work). Theywere then assignedto different

teams andmeasured duringthe twelfth week (mid-pointof team work) and sixteenth

week of classes (later partof team work).

Performance measures were collected fromthe respective programcontacts.

MBA andMAcc instructors provided their ratings to the deansof their respective

departments. Thegraduate-student instructors provided their ratings directly. The engineeringdepartment teams facilitator collected theproject gradesfrom the engineering instructors.

Students were gatheredfor debriefing at the end of the semester. MBAstudents were not debriefedwitil spring semester, since they were measured during both fall and spring semesters. Thepurpose of this study andan explanation of the hypotheseswere provided to theparticipants, as well as some suggested "further reading" on teamsand conflictmanagement.

Toe variables used in this study were designedto assess the group's level of relationship, task, and process conflict, their viability, and the conflictstyle most often used by their group. Performance measures were gathered fromthe instructors.

31 Measures

Nearlyidentical four-pagesurveys were distributedtwice duringthe semester.

(Thedifferences in the surveyswere minor; forexample, demographicinformation was

only collected at Time 1.) Theseven scales used in this study were part of a largersurvey

effort. (Of the 126 items on each survey, all items addressed some formof group

interaction process such as group communication, groupgoal clarity, groupreflexivity,

etc.) Thegeneral instructions included at the top of each surveywere as follows:

"Completing this surveywill require approximately 15-20 minutes. No individual

responses will berevealed. All data will besummarized at the group-level. Participation

in this project is voluntary. Ifyou choose not to participate,you will not be penalized.

You may terminate participationat any time. Thereturn of thissurvey will constitute

informedconsent to participate."

Team conflictstrategies were assessed using a modifiedversion of the ROCI-II

(Rahim, 1983a). Task, relationship, andprocess conflictwere measured using scales

developed by Jehn (1995). Several additional items were also developed in order to

improve on Jehn's existing scales. Group viabilitywas assessed using a scale developed

by Sundstrom (George, Perkins, Sundstrom, & Myers, 1990). Teamgrades were collected fromthe instructors and used as the groupperformance measu re.

ConflictStrategies

The students reportedon theirteam's conflictstyle using a modifiedversion of the

ROCI-II scale (Rahim, 1983a). The original scale measuredan individual's situation specificconfli�t style. It consisted of 28 5-point Likert scale questions which measured

32 the fivestyles of handlinginterpersonal conflict, consistent with Thomas' taxonomy. The

Likert scaleresponses rangedfrom strongly agree (5) to strongly disagree(1 ). Evidence of the scale's validity is indicated by the ROCI-Il's ability to discriminate between groups having known,differing conflict styles andby the scale's relationship with other conflict constructs(Rahim, 1983a, 1983b, 1986; Weider-Hatfield, 1988). Other variationsof the original ROCI-11have been developed to assess the conflictstyle used by anindividual when interacting with their peers, supervisors, or subordinates (RahimMagner, & 1995).

Thisstudy used a groupvariation of the original ROCI-IIto assess the group's level andstyle of conflict-management, substituting the words ''team" or ''teammembers" where the original referredto either "supervisor," "peer," or "subordinate." For example, where the original ROCl-11item read: "I often went along withthe suggestions of my supervisor," the modified group variation reads: "Members of my teamoften went along with the suggestions of other team members." This group variation of the questionnaire asked individuals to consider the actions of their teammates during team exercises. The response options were stronglyagree (5), agree(4), neutral(3), disagree(2), andstrongly disagree(1 ). While students' responses were elicited for all fiveconflict-management styles,this study only focusedon thethree conflict-management strategies associated with the integrative axis ( collaborative, compromising, andavoiding). The collaborative, compromising andavoiding scales consisted of seven, four,and six items respectively, but one item each was dropped fromthe collaborative andavoiding scales. Theresponses to correspondingscale items were averaged to create each conflict-managementstrategy and average scores could rangebetween 1 and5. Thecoefficient alphas forthe scales

33 included inthe modifiedROC I-Il(collaborative, compromising, andavoiding) were .88,

.80, and .83, respectively. See Figures 2.1 through2.4 for theitems collected andthe

instructions given for completing these scales.

ConflictMeasur es Groupmembers were also asked to assess the level of conflictpresent in their respective groups. An eight-item measuredeveloped by Jehn (1995) was used to measure task and relationship conflict. Process conflictwas measuredusing three items from

Jehn's (1992) IntragroupConflict Scale. Nine additional items were added to augment

Jehn's three conflictscales, since past researchshowed thatseveral of the items loaded negatively on their factor(Jehn, 1993, 1995). The 5-pointLikert scale response options were: constant(5); a lot (4); some (3); a little (2); and none (1). See Figures2.5 through

2.8 for the fullset of items collected and the set of instructions given forcompleting these scales.

Task Conflict. Conflictassociated withthe group's task was assessed using

Jehn's (1995) four-itemmeasure of task conflict(e.g., "How frequentlyare there conflicts about ideas in your work unit?")and four newly developed items. Referencesto "work unit" inthe originalscale were changedto ''team." All eightitems were retainedand summed to create this scale (which could range between8 and 40), andthe coefficient alpha was .90.

Relationship Conflict. Conflictinvolving one 's affectiveresponses to individuals was assessed using Jehn's (1995) four-itemmeasure of relationship conflict(e.g., "How

34 1. Members ofmy team tried to integratetheir ideas to come up with a decision jointly.

2. Most membersof my teamtried to work together for a proper understanding of a problem.

3. Most members of myteam tried to investigate an issueto find a solution acceptableto us.

4. Most teammembers triedto work togetherto find solutionsto problems which satisfied ourexpectations.

5. Members ofmy team exchanged accurate info rmation with each other to solve a problem together.

6. Members of my team collaboratedwith each otherto come up with decisions acceptableto us.

7. My teammates triedto bringall our concerns out in theopen so that theissues could be resolved in the best possible way.

Note: Items are measured on a five-point scale (1 = strongly disagreeto 5 = strongly agree).

Figure2. 1 Items measuring the"collaborative" conflict-management style.

35 1. My teammatestried to finda middle courseto resolve an impasse. 2. My teammatesnegotiated with each otherso thata compromise couldbe reached. 3. Most membersof my team used "give andtake" so that a compromise could be made. 4. Members ofmy teamusually proposed a middle groundfor breaking deadlocks.

Note: Items are measuredon a five-pointscale (1 = stronglydisagree to 5 = strongly agree).

Figure2.2 Items measwingthe "compromising" conflict-management style.

36 1. My teammembers attemptedto avoid being "puton the spot" andtried to keep theirconflict with otherteam members to themselves.

2. Most members of my teamtried to keeptheir disagreements with each otherto themselvesin order to avoid hard feelings.

3. Members of my teamtried to avoid unpleasant exchanges with each other.

4. Most members of my teamusually avoided opendiscussions of theirdiff erenceswith each other.

5. Most members ofmy teamtried to stayaway from disagreements with each other.

6. My teammates avoidedencounters withother. each

Note: Items are measuredon a five-point scale (1 = strongly disagreeto 5 = stronglyagree).

Figme 2.3 Itemsmeasuring the "avoiding" confli ct-managementstyle.

37 SECTION I-Incompatibilities, disagreements, or differences (i.e., conflict) within teams occur regularly. Please carefullyread each of the following statements, then indicate YOUR EXPERIENCES andhow YOUR TEAM handled any conflictsituations that may haveoccurred from the beginning of the semester through the completion of:XXXproject name XXX. Using the following scale, indicate your responses by writing 1, 2, 3, 4, or 5 in the blank to the left of the item.

Figure 2.4 Instructions forcompleting the ROCI-II.

38 1. To what extent were there differencesof opinion in yourteam? 2. How often did yourteanunates disagreeabout opinions regardingthe work being done? 3. How much conflictabout thework you do was therein yourteam? 4. How frequently were thereconflicts about ideasin yourteam?

5. How often did yourteammates disagreeabout ideasregarding the task? * 6. To whatextent were differencesof opinion regardingtasks evident among yourteammates? * 7. To what extent where there disagreements aboutthe tasks you were working on with teammembers? • 8. How oftendid yourteammates disagreethe about work being done? •

Note: Items are measured ona fiv�point scale (1= none to 5 = constant).

• Newly developed items

Figure2.5 Items measwingtask conflict.

39 1. How much friction was there among members of yourteam?

2. How much tensionwas there amongteam members?

3. How much were personalityconflicts evident in yourteam?

4. How much emotional conflictwas there among teammembers?

5. How much anger was present in yourteam?

6. To what extentwere personalityclashes evidentin yourteam?

Note: ·Items aremeasured on a :five-point scale (1 = none to 5 = constant).

Figure 2.6 Items measuring relationship conflict.

40 1. How often did your teammates disagree about how thework should be divided? *

2. How frequently were there disagreements about who should do what in yourteam?

3. How much disagreement was there about procedures in yourteam?

4. To whatextent did you disagree aboutthe way to dothings in your team?

S. How often did your teammates disagree about whoseresponsibility it was to complete a task? *

6. To what extent did yourteammates disagree aboutthe processes used to complete tasks? *

Note: Items are measured on a five-point scale (1 = none to 5 = constant).

* Newly developed items

Figure 2. 7 Items measuring process conflict.

41 SECTION II - This sectionasks you to describe the ongoing activities in your team. Please read each item and indicate the response that best represents your evaluation of YOUR TEAM from the beginning of the semester through the completion ofXXX project nameXXX.

1 =none 2 = a little 3 = some 4 = a lot 5 = constant

Figure 2.8 Instructions forcompleting the items representing relationship, task, andprocess conflict. ·

42 much frictionis there among members in your work unit?") and two other items (Jehn,

1992). References to "work unit" were changedto ''team." Five of Jehn's items were retained andsummed to create a relationship conflictscale ( which could range between 5 and 25), andthe coefficientalpha was .95.

Process Conflict. Threeitems fromJehn's (1992) IntragroupConflict Scale (e.g.,

"How much disagreementwas there about proceduresin yourwork group?")and three additional items were developed to assess process conflict. References to "work group" were changedto ''team." Theitems reflecteddisagreements about roles andthe distributionof resources within the teams. Two newly developed items andone of Jehn's were summedto created thisfinal conflict scale (which could rangebetween 3 and 15), andthe coefficientalpha was .88.

Viability

Five items were used to measure the group's viability. Thisscale was deveioped by Swidstrom(George, Perkins, Swidstrom, & Myers, 1990), and covers aspects of satisfaction, participation,and capacity forfuture work (e.g., "Everyoneon my team wantsto continue working together in the future"). All items were scored using a 5-point

Likert scale, with thefollowing responseoptions: stronglyagree (5), agree(4) neutral

(3), disagree (2), stronglydisagree (1 ). Thescale score was calculated by averaging across all of the items, and could rangebetween 1 and5. The coefficientalpha for this scale was .86. See Figures 2.9 and2. 10 forthe fullset of items collected and the instructionsgiven for completing the scale.

43 1. Everyone on my team wants to continue working togetherin the future.

2. Everyone on myteam does their share of the work.

3. I find it personally satisfyingto be a member of my work-team.

4. Certainmembers of ourteam aren'tpulling their weight. (R)

5. I am proud to be a member of thiswork team.

Note: Items are measured on a five-pointscale (1 = strongly disagree to 5 = stronglyagree). R = Reverse-codeditem.

Figure2.9 Items measuringteam viability.

44 SECTION ID - This part of the questionnaire asks how you fe el about your team and about your team's ability to function. In the blanksprovided, please indicate how you currently fe el about your team TODA Y.

1 = stron 3 = neutral

Figure 2.10 Instructions for completing group viability items.

45 Individual-Level Variables

Collaborating Strategy. Individual scores were obtained by averagingresponses

to 6 items, all rangingfrom 1 to 5 where 5 indicated thegreatest collaboration. Individual

scores rangedfrom 1 to 5 with a mean of 4.04 and standard deviation of .60.

Compromising Strategy. Individual scores were obtained by averagingresponses

to 4 items, all rangingfrom 1 to 5 where 5 indicated thegreatest compromising.

Individualscores rangedfrom 1 to 5 witha mean of 3.63 andstandard deviation of .66.

A voiding Strategy. Individualscores were obtained by averaging responses to 5 items, all rangingfrom 1 to 5 where 5 indicated thegreatest avoidance. Individualscores rangedfrom 1 to 5 witha mean of 2.98 andstandard deviation of . 78.

Relationship conflict. Individualscores were obtained by summing responses to 5 items, all ranging from 1 to 5 where 5 indicated thegreatest conflict. Individualscores rangedfrom 5 to 25 with a meanof 10.19 andstandard deviation of 4.77.

Task conflict. Individualscores were obtained by summing responses to 8 items, all ranging from 1 to 5 where 5 indicated thegreatest conflict. Individual scores ranged from8 to 40 witha meanof 17 .68 andstandard deviation of 5. 70.

Process conflict. Individualscores were obtainedby summing responses to 3 items, all rangingfrom 1 to 5 where 5 indicated the greatestconflict. Individualscores rangedfrom 3 to 15 with a mean of 5.13 andstandard deviation of 2.34.

GroupViabili ty. Individualscores were obtained by averaging responses to 5 items, all rangingfrom 1 to 5 where 5 indicated thegreatest viability. Individualscores rangedfrom 1 to 5 witha mean of 3.49 and.standard deviation of .93.

46 Group-Level Predictors and Viability

To createthe group-levelvariables, individual scale scores forthe predictor

variables( collaborating, compromising andavoiding strategiesand relationship, task, and

process conflict) andthe criterion variable of groupviability were aggregated using the mean aggregationfunction in SPSS. (Individualscale scores were standardizedbefore aggregationto controlfor group differences. Thegroup-level scores forthe 77 teams rangedfrom -1 .97 to 2.19. See thePreliminary Analysis Section of Chapter 3 for individual-level unstandardized andgro up-level descriptives.) Individual team member scale scores were aggregatedto create a team scale score forthe item when there were at least threeindividual scale scores per team. There was no need to aggregateperformance since team scores were obtained fromthe course instructor.

Performance

Performancewas measured throughoutthe semester by thecourse instructorsand departmentfaculty. Manyof the performance measures correspondedto thecompletion of a team project andto thedistribution of a survey. For Time I, teamperf ormance scores were thesum of all team project gradesup to the completionof the firstsurvey.

For Time2, teamperformance scores were theswn of all teamproject gradesfrom the completion of the first surveyuntil the end of the semester.

l\IBAstudent performancewas measuredusing threelevels: Below Expectations

(BE), Meets Expectations (ME), andExceeds Expectations (EE). For Time I, instructors provided four performancemeasures forthe fall l\IBA teams. The instructorsassigned numericvalues of -1, 0, or 1 forthe performancescores, representing BE, ME, andEE,

47 2 respectively. Thus,the performance scores at Time 1 could rangebetween -4 and4.

Two performancemeasures were obtainedfor the spring MBA teamsat Time 1. Numeric 3 values of-1, 0, and 1 were used again, andthe scores could range from -2 to 2. Toe

combined performance scores are presented in Table 2.2.

MAccstudent performance was also assessed by courseinstructors and

departmentfaculty. Dependingon the project, their performancewas measuredusing

eithera three-level (BE, :ME,and EE) or a seven-level ("A+" through "C") scale.

Correspondingnumeric representations used by the accounting facultywere: BE (76);

:tv1E(8 8); EE (100); or A+ (98); A(95); A- (92); B+ (88); B (85); B- (82); C (78 or 75).

MAccpe rformancewas calculated for Time 1 by combining two team project 4 scores. Whilethe firstpro ject was completed earlierthan the mid-po int of thesemester,

facultymembers feltthat it was a better representation of team performancethan the

second project. (They said that students were more attentive duringthe firstpro ject

because during the period of the second project, the students were consumedwith

interviewingfor posi tions theywould take followinggraduati on.) Therefore,the first

project score was doubled, then summed withthe second project score. These scores can

be fonnd inTable 2.2, with theirrange being between 228 and 295. (The firstpro ject

occurredduring the fourth week of classes, and was worth 10% of the students' grades.

The second project occurredduring the eighth week of classes, and wasworth only 5% of the students' grades.)

2 Performance at Time 2 could rangebetween -3 and 3 because 3 performancemeasures were collected. 3 Performance at Time 2 could rangebetween -2 and 2 because 2 performancemeasures were collected. 4 Performance at Time 2 was assessedusing a single measure ofperformance. 48 Performancefor the undergraduatechemical engineering students was measured

using a typicalpercentage scale of O - 100%. Lettergrades were not assigneduntil the

end of the semester, but students understood that 90 -100% was an "A;" 80 - 89% was a

"B;" 70 - 79% was a "C;" 60 - 69% was a "D;" and below 60% was an "F." For Time 1,

the performancescores forthe chemical engineeringstudents consisted of the summation of threepro ject scores, andthe rangeof these scores was between O and300. 5

Performancefor the remainingengineering teams (mechanical,fluid, and automotive) was a one-time performancegrade at the end of the semester on an "A" through"F" scale.

Theengineering faculty assigned the followingnumeric values to the letter grades:

6 (A+); 5 (A); 4 (B+); 3 (B); 2 ( B-); 1 (C+); 0 (C).

Undergraduatebusiness student performancewas measured using a typical percentage scale. Students understoodthat 90 -100% was an "A;" 80 - 89% was a "B;"

70 - 79% was a "C;" 60 - 69% was a "D;" andbelow 60% was an"F/ ' For Time 1, two of the three business courses had only one performance measure. Thetwo performance measures forthe remaining course were summedto create a performancescore. The numberof performancemeasures collected for a given period andall the methods used to access performanceare summarizedin Table 2.2.6

Differencesin performancerating systems used by the MBA,MAcc, and undergraduate engineeringand business facultywere controlledfor by standardizing

5 Performance at Time 2 was thesum of 2 performancemeasures. 6 See Appendix B, Table B.2, forTime 2 summary of performance measures.

49 Table 2.2. Summary of performance measures at time 1.

MBA MAcc Under2raduate Business 2 Semester Performance Week 5 to Week 3 to Week 8to Period Week 8 Week 8 Week 12 Perf. Scale EE, ME, BE EE, ME, BE Percenta2e Percenta2e AtoF Percenta2e Percenta2e Percenta2e 4 2 2 3 2 -2 to 2 227 to 295 0 to 100

292 1 NIA 1 192.5 1 288 1 ( only measured 1 190 1 285 2 at Time 2) 1 183 1 275 1 1 175 2 271 2 1 IJl 268 2 0 261 5 247 1 244 1 240 1 237 3

;�ffl�'i�;;f4:;s�ll';;,1,�iM&iwt�t�Sl11;S f'[MlffifJ,:(1.i{Sll't f;rSijffliJi;!rfSIJ1i� j�;';�1mr:im!1:i#D.0i: .44 1.20 .35 . 79 ifllm265.4�iAiJiit 25.Sl11:�tw± fifu:ffliffl;�;;Jiit�;ifa! 80.0 10.0 84.6 5.86 183.1 8.17 performancescores within each of the respective populations andcourses. Ninety-one teamswere solicited forthis study, and while only 77 teams participatedat Time 1 and

65 teams participatedat Time 2, all performancescores were utilized when standardizing performancewithin their respective populations or courses. lMBAand MAcc performancescores were standardized within their respective groups. Undergraduate chemical engineering and business student performancescores were standardized within their respective course populations (performancewas measured atTime 1 andTime 2).

Theperformance scores of the remaining 14 undergraduateengineering teams were standardizedwithin their respective population (performancewas measuredat Time 2).

Aggregated Variables

All of thepredictor variablesand the criterion variableof groupviability were aggregatedto the grouplevel. A performancescore foreach team was obtained fromthe courseinstructor, so aggregationof this variable was not necessary. Justificationfor the use of aggregated variables is dependent on meeting certain conceptual andstatistical conditions, as discussed in Chapter 1.

Using the James, Demareeand Wolf (1984) procedure, 77 estimates (i.e., number of groups) were calculated foreach of thefollowing scales: relationship conflict, task conflict, process conflict, collaborative strategy,compromise strategy,and avoiding strategy, andgr oup viability.

James (1982) also suggested that beforeaggregating the variables, the differences between groupsshould be shown,so between-groupdifferences were examined using the

George andBettenhausen ( 1990) technique. One-way ANOV A's were calculated for

51 each of thescales. If the obtained F ratio was greaterthan 1.0, group differences were

deemed significant based on Hays ( 1981). Theresults obtained using the rwg statistic and

the ANOV A proceduresupported thedecision to aggregatevariables to thegroup level.

Power Analysis

A power analysiswith respect to multiple regressionrevealed thatthere was a

50-60% probability of detecting a medium effectsize (12 = .15).7 However, it was not

possible to achieve a power of .80 or greater(which would have required at least 104

teams), giventhe limits inthe availability of teamsand funding. Thepower analysis calculation suggests thatthe limited samplesize may make it difficultto determinesome

of thehypothesized relationships, especially thoseinvolving interactions (i.e., hypotheses

3c, 3d, and 3e).

7 See AppendixA, Table A.3, for anticipatedsample sizesfor powers .80, .85, and .90.

52 CHAPTER 3

RESULTS

Data Analyses Beforeaggregation, several preliminaryanalyses were performedusing

individual-leveldata. Principal components exploratoryfactor analyses were used to

evaluate both the ROCI-Iland the scales that measured thethree types of conflict and

scale reliabilities were calculated.

Tests forgroup differenceswere conducted, andz-scores were computed for each

of the variables since differenceswere found. Individual standardized correlations were

also examined.

All of the variables except performance were aggregated. Performance, a group

level variable, was collected fromcourse instructors. Sincethere were differences in the

performancescales used by the instructors, the scores were standardized withinthe

respective populations. Hypotheses 1 a through I c and2a through 2c were tested using

correlational analyses. Hypotheses 1d and2d were tested using simple regression.

Beforetesting hypothesis3, tests were conducted to examine whether the assumptions were met forhierarchical regression,including tests formulticollinearity andoutliers, and examination of residual scatterplots. Finally, tests of the study' s hypotheses were conducted using theaggregated data.

53 PreliminaryAnalyses

Threeprincipal components exploratoryfactor analyses were used to evaluate the

psychometricadequacy of thescales measuringthe three conflict-managementstyles and

the threeconflict types. The modifiedROCI-Il scales (collaborative, compromising, and avoiding strategies)were examined using the first principal components analysis

(varimaxrotation). Theseven items measuring the collaborativestrategy loaded on the firstfactor. Thesix items measuring the avoiding strategyloaded on the second factor and the four items measuringthe compromisingstrategy loaded on the third factor.

Based on the items' loadings on their respective factors(factor loading of less than.60), previous research (which used confirmatoryfactor analysis to examine the fiveconflict management strategies, modifiedto reflect the group(Vigil-King, 1999)), and known problems with the compromising scale (see discussion in limitations), one item measuring the avoiding strategyand one item measuringthe collaboratingstrategy were eliminated. Theitems measuringeach of the conflictmanagement strategies were then averaged to create their respectivestrategies: avoiding, collaborating, compromising.

Theresults of the principalcomponents factoranalysis are shownin Appendix A,

Table A.I.

Thesecond principal components analysis( also a varimax rotation) examined the scales that measured the threeconflict types. These scales included Jehn's original 13 items andthe 7 newly developed items. Thisanalysis failed to indicate separate factors foreach of the three conflictscales, so a third principal component analysis (4-factor, quartmax rotation) was performed. Theresults of this analysis (see Table A.2 in

54 Appendix A) showed a general conflictfactor and two separate independent factors. All

8 task conflictitems clearly loaded on thegeneral conflictfactor. Five items identified as

relationship conflictloaded on thesecond factor. Threeitems identified as process

conflictloaded on the thirdfactor. Based on this analysis, one relationship conflictitem

andthree process conflict items were dropped. Theconflict items were then summed

togetherto create theirrespective conflictscales.

Scale reliabilities were also examined. Using Cronbach's alpha, it was

determined that all scale reliabilities were above .80.

Since fourdifferent student populations were used in this study, tests forgroup

differences were conducted in addition to testing thepsychometric adequacy of thescale

items. (Different subpopulations have been foundto perceive andrate conflictand other

variablesdifferently.) Significantmean differences and differences in varianceswere

foundfor many of thevariables; Table 3.1 shows these differences.8 Z-scores were

computed foreach of the variables to controlfor these group differences.

Individual-level standardizedcorrelations between integrativeconflict managementstrategies showed that the collaborative and compromising strategieswere significantlyrelated (.52, p < .01) andthat the compromising andavoiding strategieswere significantlyrelated (.11, p < .05). All threetypes of conflicthad significant negative relationships with thecollaborative andcompromising strategies, andthe largest

8 See Appendix B, Table B.3, for Time2 descriptivestatistics by sub-populations (I\,IBA,MAcc, Engineer, Undergraduate Business).

55 Table 3.1. Individual-level descriptive statistics, tests formean differences, and test forequality of variances forparticipating students at time 1.

l0 Variables I MBA I MAcc I Ene:ineerine: I Business I F-value 1 I Levene Test F-value SD N N -·�j}, Collaborative Strategy 52 Compromise Strategy 3.53 .66 52 3.80 .59 72 3.57 .80 67 3.79 .49 .77 , I I .77 Avoiding Strategy 149 2.95 52 2.97 .70 I 72 2.88 I .84 I 67 I 3.17 I t ]1 5. Relationship Co� 149li.06 52- 4".63" s'.'u"'3.26 - 7 2 • 1:r1·-5.38 ·' 67 7.75 3.46 19.54••• . 8.02••• 6. Task Conflict 149 19.85 5.32 52 14.72 4.20 72 17.71 6.17 67 15.10 4.94 18.99••• 1.74 7. Process Conflict 148 5.43 2.34 51 4.37 1.89 72 5.50 2.86 67 4.63 1.79 4.37•• 4.42••

VI °' Note: Viability and Strategy Predictors were measured using a 5-point Likert scale where1 = "Strongly Disagree"and 5 = "Strongly Agree." ConflictPredictors were measured using a 5-point Likert scale where 1 = ''None" and 5 = "Constant." •n � .o5, ••n � .01, *** p � .001.

(a) > Perfo rmancewas measured at the group level andstandardized within the MBA, MAcc,engineering andundergraduate business studentpopulations. (b> Tests formean difference (ANOVA) between the MBA,MAcc, Engineer, and Business student populations respectively.

Means, standard deviations, correlations, andreliabilities forthe individual-level

variables (beforeaggregation) are shown in Table 3.2.9

Aggregation

Aggregation of standardized individual responsesto thegroup level was only

performedwhen therewere at least threeindividual responses per team. As explained in

the"Participants" section, a total of 17 surveyswere eliminated because the information

used to identifyteam membership was missing, or because fewer than three members of a

given team completed the survey. 10 Withthe exception of process conflict, 86% of the

estimates were in theacceptable range (i.e., greaterthan .70). Table 3.3 contains the

results of aggregation. 11 Theseresults suggest that process conflict does not meet

traditionalcriteria fora group-levelmeasure · andwas thereforeeliminated from further

analyses. Overallgroup-level means, standarddeviations, andranges for aggregated

variablesare in Tables 3.4 and3.5. 12 (Group-level correlationscan be found laterin the text.)

Team Viability - Hypothesis 1

Correlationswere computed to test forHypotheses 1 a, 1 b, and1 c. Hypothesis1 a predicted relationship conflict would be related to team viability. Congruentwith the prediction, a negative, statistically significantcorrelation was foundbetween relationship conflict and viability (-.82, p < .00 I). Hypothesis I b predicted that perceived task

9 SeeAppendix B, Table B.4, for Time 2 standardized descriptive statisticsand zero-order correlations. 10 At Time 2, 37 surveys were eliminated. 11 See Appendix B, Table B.5, for theresults of aggregation at Time 2. 12 See AppendixB, Tables B.6 and B.7, forgrou p-level descriptive statisticsat Time 2.

57 Table 3.2. Individual-level standardizeddescriptive statistics andzero-order correlations for all participants and reliabilities at time 1.

Variables M SD 1 2 3 4 5 6 7

Relationship Conflict .00 1.00 -.62•• -.39·· .02 (.95) TaskConflict .00 1.00 -.44•• -.45.. -.21•• -.07 .69•• (. 90) ProcessConflict .00 1.00 -.49. . -.so•• -.30 .. .04 .61•• .63 .. (.88) Note: Allstatistics were computedat theindividual level; N = 340.

.05, .01, p .001. *R s **R s *** s <•> Performance was measured atthe group level and standardizedwithin the MBA, MAcc,engineering and undergraduate business studentpopulations.

58 Table 3.3. Distribution ofr"S forvariables at time I.

Variables s.S9 .

.08 .11 .20 .54 .82 1.83 .04 .04 .22 .65 .91 .87 1.56 .09 .09 .08 .11 .28 .31 1.23

With the exceptionof process conflict, overall 86% of these

<•> Performancewas measured atthe group level.

59 Table 3.4. Group-level descriptives forparticipating teams at time 1.

Variables Team Descriptives

<•> Performancewas measured at the group level.

60 Table 3.5. Group-level descriptive statistics forsub -group populations at time 1.

Variables MBA MAcc Engineering Teams SD

-.07 5. 32 .01 .66 -1.3 to l.2 12 -.02 ,69 -1.0 to 1.5 18 .79 -1.0 to 1.7 .03 -.9 to 1.4 , r 1 11s 6. 32 .04 .64 -.9 to 1.3 12 -.06 .56 -1.1 to .9 18 -.04 .75 -.9 to 2.2 15 .02 .63 ,I -t.4 to 1.0

<•> Performancewas measured at the group level. conflictwould be related to team viability, and a negative, statistically significant

correlationwas also f01md between task conflictand viability (-.67, p < .001). Table 3.6

contains the zero-order correlationsfor all variables. 13 Hypothesis1 c ( dealing with

process conflict)was eliminatedfrom the analysis due to problems cited earlier.

Hypothesis 1 d stated that relationship conflictwould be more predictive of group

viability than either task conflictor process conflict. A test forthe significanceof the

difference between dependent r's was conducted (Cohen & Cohen, 1983; p. 56) andwas foundto be significant(t = -3.35, p < .01).14 (Process conflictwas eliminated fromthe analysisdue to problems cited earlier in the text.)

Performance- Hypothesis 2

Correlations werecomputed to test forHypotheses 2a, 2b, and 2c. Hypothesis2a predicted that teamsperceiving moderate task conflict wouldhave higherlevels of performancethan teams perceiving highor low task conflict. To test forthe expected curvilinear relationship, the task conflictwas divided into 3 groups(using SPSS ranking method-ntiles, which created 3 groupswith approximately the same numberof cases).

Contraryto prediction, no significantrelationship was foundbetween moderate levels of task conflictand performance ( .10, p = ns). 15 Hypothesis2b predictedthat perceived relationship conflictwould be related to team performance. Again, contraryto prediction, no significantrelationship was foundbetween relationship conflictand performance ( .12,

13 See AppendixB, Table B.8, for group-level descriptive statistics and zero-order correlations at Time 2. 14 For Time 2, test forthe significance of the differencebetween dependent r's was f01md to be significant (t = -3.76, p < .01). 15 For Time 2, no significant relationship was fowidbetween moderate levels of task conflict and 15 performance (.20, p = ns).

62 Table 3.6. Group-level descriptivesta tisticsand zero-order correlations forall groupsat time 1. Ca)

Variables M SD 1 2 3 4 5 6 7

-.1s•• -.67•• -.63** -.48** .78** Note: Allstatistics were computedat thegroup level;N = 77. *Rs;.05 , **Rs;.01

<•> Because of significant groupdifferen ces, all variableswere standardized within theI\1BA, MAcc, engineering, and undergraduate business studentpopulations.

63 p = ns). (See correlations,Table 3.6). Hypothesis 2c (dealing with process conflict)was

eliminated fromthe analysisdue to problems cited earlier.

As with viability, simple linear regressionwas usedto test the relationship

between the conflicttypology and performance. Hypothesis 2d stated that task conflict

would be more predictive of group performance than relationship conflictor process

conflict.. A test forthe significance of the differencebetween dependent r's was

conducted (Cohen & Cohen, 1983; p. 56) and was not significant(t = -.53, p < ns). 16

(Process conflictwas eliminatedfrom the analysisdue to problems cited earlier in the text.)

Tests of Accuracy of Assumptions Multicollinearity. Several of the correlationcoefficients between the independent variables were relatively high(p > .50), so multicollinearity diagnosticswere conducted with respect to the relevantregressions. Pedhazur(1 982) suggests that multicollinearity can be definedby veryhigh intercorrelationsamong the independent variables, but he makes it clear that there is no consensus as to what the termmeans or agreement on what constitutes ''very high" intercorrelation. Tabachnick and Fidell (1989) suggest that correlations above .90 aredangerously high andthat variableswith correlations greater than. 70 should be carefullyexamined before including them in regressionanalysis.

Collinearvariables provide verysimilar information, and their effectsare difficult to separatebecause of this. Regarding regression, highcorrelations between independent variableshave been foundto affectthe stability of theregression weights while R itself

64 remains unaffected(Pedhazur, 1982). Thus,minor changesin the data canproduce

substantialchanges in a beta weightbecause the significanceof thebeta weight is based

on the ratio of its absolute value to its standard error.

Two diagnostic measures of collinearity were used: tolerance and variance

inflationfactor (VIF). Thetolerance of a variable has beendefined as 1.0 minus the

squared multiple correlationcoefficient. A low tolerancein dicates that the variable is

almost a linear combination of the other independent variables. Closely related to

toleranceis the VIF,which can be best definedas the reciprocal of the tolerance. A high

VIFalso indicates that the variable is almosta linear combination of the other

independent variables. Tolerancesfor all independent variables were greater than .37 and

the highestVIF was 2. 70 (for the collaborative conflict-managementstrategy), so

multicollinearity was not foundto be a problem.17

Outliers. Univariate andmultivariate tests for outlierswere conducted.

Scatterplotsbetween each of the predictor variablesand each of thecriterion variables

(univariate) and Mahalanobisdistance (multivariate) were used to determine outliers.

Theunivariate scatterplots indicated thepossibility of outliers, but no values were in excess of 3 standard deviations fromthe mean. (Figure3. 1 shows a sample scatterplot forthe relationship between the compromising conflict-management strategy and performance.) Using Mahalanobisdistance (x, 2 = 24.322; p < .001) with seven degreesof freedom,three teams were identifiedas outliers. However, these threeteams were not

16 For Time 2, test forthe significanceof the difference betweendependent r's was foundto be significant (t = -.13, p < ns). 17 For Time 2 tolerancesfor all independent variables were greater than .24 and the highestVIF was 4.14. 65 3.0

2.0 I

0 0 0 0 1.0 I Q) 0 [:cl 0 E 0 0 0 0 8 0 i= 0 0 0 0 0 0 0 o 0 -as 0 0.0 I qao 0 >, D 0 C) 0 � Do. 0 0 8 o Do 0 Q) o C 0 0 0 -as J 0 0 (/)- •1.0 I 0 C) 0 C ·en

.E •2,0 I E u -3.0 - - -3 -2 -1 0 2 3

Performance at Time 1

Figme3.1 Samplescatterplot

66 eliminatedfrom the regressionanalyses because they had previouslybeen identifiedby

. facultyas problematic, and 87% of the rwgS (level of agreement) foreach of the aggregatedvariables for these three teams were above . 70.

Examination of Residuals Scatterplots. Residual scatterplotswith respect to

multiple regression were used to test the assumptions of normality, linearity, and

homoscedasticity between predicted criterion scores and errorsof prediction. When these

assumptionsare met, theresiduals are normally distributed about the predicted criterion

andhave a straightline relationship withthe predicted criterion scores, andthe variance

of the residuals about the predicted criterion scores are thesame forall predicted scores.

Plots of the predicted values of groupviability andperformance withresiduals indicated that the assumptions forregression were met. Thescatterplots of thepredicted criterion scores anderrors of prediction were rectangularin shape, withno values exceeding three standarddeviations fromthe mean. Thebest fit line was a straight horizontal line positionedat zero, which was thestandardized mean.

Hierarchical Regression - Hypothesis 3

Hierarchical regressionanalysis was used to test the relationships associated with

Hypotheses3a and 3b. It was hypothesizedthat teams usingmore integrativeconflict managementstrategies would have higher performanceand viability than teams adopting less integrativestrategies. For Hypotheses3c, 3d, and3e, moderated regressionwas used to test themoderating influenceof conflicttypes on the relationships associated with the first two hypotheses. Interaction terms were created and added to the model by multiplying each of the three conflict-managementstrategy variables by the type of

67 conflictpreviously stated in thecorresponding hypothesis. Specifically,for Hypotheses

3a and3b therole of conflict-management styles was examined in Step 1 foreach

criterion variable. Step 2 examinedthe simultaneous effect of task conflict andconflict

management strategieson the criterion variables. Theinteraction of task conflict with each of theconflict-management strategies was tested in Step 3 (a test of Hypotheses 3c,

3d, and3e).

Performanceand Task Conflict. A significantrelationship between conflict managementstyles and performance was foundin Step 1 (Hypothesis3a). Theoverall model (F = 3.34, p < .05) and the coefficientsassociated with the relationships between the collaborative strategyand performance (.35, p < .01) andbetween thecompromising strategyand performance (-.43, p < .01) were significant. Thebeta weight forthe avoiding strategy was not significant(-.02, p = -.15). Specifically, the beta weight forthe collaborative strategy was positive in signand the beta weight forthe compromising strategywas negative in sign.

For Step 2, the analysisof therelationship between conflict-management styles andperformance, controlling for levels of task conflict,was significant(F = 4.12, p < .01). Thebeta weight for the collaborativestrategy increased (.55, p < .01) when task conflict was added to the model. Thebeta weightsfor the compromising andavoiding strategiesremained unchanged. For Step 3, Hypothesis3c was not supported by the data.

Whilethe overall regressionwas significant(F = 2.42, p < .05), none of the coefficients

68 for the interactions were significant.18 Theresults of the hierarchical regressionanalysis

forthe predictors of performanceare found in Table 3.7.

Performanceand Relations hip Conflict. Similar testingwas performed for

Hypothesis3d, except that relationship conflict replaced task conflictin Step 2 and

beyond. The simultaneousentry of the threeconflict strategieswas analyzedabove, so an

explanationof Step 1 will not be repeated. The analysisof therelationship between

conflict-management stylesand performance, controllingfor levels of relationship conflict,was significantin Step 2 (F = 3.90, p < .01), andthe beta weight forthe collaborativestrategy increased when relationship conflictwas added to the model. As before, thebeta weightsfor the compromising andavoiding strategies remained relatively unchanged.

For Step 3, Hypothesis3d was supported by thedata. Thisstep tested the interactions between the conflict-managementstrategies and relationship conflictwith performance. Theoverall regressionwas significant(F = 2.96, p < .01). Thecoefficient for theinte raction of collaboration (more integrativestrategy ) with relationship conflict was significant(.38, p < .06). Thisresult suggests thathigher performance levels were associated withteams using more integrativestrate gies, when higherlevels of relationshipconflict were perceived. Thecoefficient for theinte raction of compromise with relationship conflictwas significant(-.38, p < .05), andthis result suggests that lower levels of performancewere associated with this (less integrative) strategy, when

18 Non-significantresults were found in all three steps forthe Time 2 data See Appendix B, Table B.9. 69 Table 3.7. Hierarchical regression analysis forpredictors of performance withconflict-manag ement strategies and task conflict at time 1.

Step Predictor Variable t R F dfs

'.2tl8J�ll �li111fillllllfl1*��1l, CompromiseStrategy -.43 -2.96** Avoiding Strategy -.02 -.15 IIIJflflllllllilWB lll§J,i IIIBD l�lfBll�qj CompromiseStrategy -.42 -2.99** Avoiding Strategy -.03 -.26 Task Conflict .34 2.38**

,,,. Compromise Strategy -.38 -2.55* Avoiding Strategy -.04 -.30 Task Conflict .34 2.28* CollaborativeX Conflict .15 .87 CompromiseX Conflict -.15 -.92 AvoidingX Conflict .01 .08

Note: Step 1 is a test of thelinear relationship between ConflictStrategies andPerf ormance. Step 2 addsTask Conflict to theequation. Step 3 tests theinteraction between Conflict Strategies andTask Conflict.

+ p :s; .10; * p :s;.05; •• p .01; ••• p .001. s s

70 higherlevels of relationship conflictwere perceived. See Table 3.8 forresults of

hierarchicalregression analysis of theviability predictors. 19

Additional analyseswere performedto more clearly examine the interactions

identifiedbetween conflictmanagement strategies, relationship conflict, and

performance. Relationship conflictwas partitionedinto three equal groupsrepresenting

three levels of relationship conflict: low, moderate, and high (using SPSS ranking

method-ntiles, which created 3 groupswith approximately the same number of cases).

Correlationswere calculated between each of the conflict-managementstrategies and

performancefor each level of relationship conflict.

The correlationbetween thecollaborative strategyand performance was positive

andsignificant when relationship conflictwas high(.49, p < .05), andthe correlations

between the compromising strategyand performance and between the avoiding strategy

and performancewere negative (althoughnot significant). Thislends support to the

previously noted association betweenhigher levels of performanceand the more

integrativeconflict-management strategy at high levels of relationship conflict.

On the other hand, correlationsbetween the compromising strategyand performance and theavoiding strategyand performance werenegative andsignificant (-.45, p < .05; -.52, p

< .05; respectively) at moderate levels of relationship conflict,while the correlation between thecollaborating strategy andperformance were · positive ( although not significant). Thislends support to the previously noted association between

19 For the Time 2 data, no significantresults were found. See Appendix B, Table B.10. 71 Table 3.8. Hierarchical regressionanalysis for predictors of performance withconflic t-management strategies and relationship conflict at time 1.

Step Predictor Variable '3 t R F dfs

····-· �.�.� ,,.. , !.' CompromiseStrategy -.43 -2.96** Avoiding Strategy -.02 -.15

' :i::: l�!,�� ;'' Compromise Strategy -.39 -2.70** Avoiding Strategy -.06 -.48 Relationship Conflict .38 2.22*

·,.iit f· ' ....i :..� flfflll1t,J�!,fj i CompromiseStrategy -.33 -2.26* Avoiding Strategy -.08 -.64* Relationship Conflict .44 2.53* CollaborativeX Conflict .38 1.96+ Compromise X Conflict -.38 -2.04* Avoiding X Conflict .05 .39

Note: Step 1 is a test of thelinear relationship between ConflictStrategies and Perf ormance. Step 2 adds Task Conflictto theequation. Step 3 teststhe interaction between Conflict Strategies andTask Conflict.

+ p :S.1 0; * p :S.05 ; ** p :S.0 1; *** p :S.0 01.

72 degradedperformance and the less integrative conflict-management strategies.

(Thisassociation seems to be more pronouncedat moderate levels of relationship

conflict.)

A summaryof all thecorrelations between conflict-managementstrategies and

performancefor each level ofrelationship conflict is shown in Table 3.9.

Viabilityand Relationship Conflict. Similar testing was performedfor

Hypotheses3b and3e, with relationship conflict replacingtask conflictin Step 2 and beyond. Hierarchicalregression analysis was used to test the moderating relationship between conflictstrategies, relationship conflict,and viability. The interaction of relationship conflictwith each of the conflict-managementstrategies was tested in Step 3.

Hypothesis3b predicted that teams using a more integrativestrategy would have higherlevels of groupviability. This was tested in Step 1 of theanalyses. The overall model (F = 32.19, p < .001) andthe coefficientassociated with therelationship between collaborative strategyand group viability (.71, p < .001) were foundto be significant. All conflict-managementcoefficients were in the hypothesizeddirection, but beta weightsfor the compromising (.05, p = .61) andavoiding (-.05, p = .49) strategiesand group viabilitywere not significant.

Thepositive relationship between the collaborative strategyand viability was also evident in Step 2, controlledfor the level of conflict. When conflictwas added to theregression equation, therelationship between conflict-managementstyles, relationship conflict, andviability was also significant(F = 45.93, p < .001).

Relationship conflicthad a counteradditive (negative) association with thecollaborative

73 Table 3.9. Correlationsbetween conflict-management strategiesand performance at tri-levels of perceived relationship conflict at time 1.

Relationship Conflict Level Conflict-Management Strategy Performance

High Collaborative .49*

N=26 Compromising -.21

Avoiding -.04

Moderate Collaborative .26

N=25 Compromising -.45*

Avoiding -.52*

Low Collaborative -.09

N=26 Compromising .14

Avoiding .11

*RS.05

74 strategyin its relationship to viability (i.e., higherviability when conflictassociated with

collaboration). Beta weightsfor the compromising strategyand the avoiding strategy

remained unchanged. Hypothesis3e (Step 3) was not supported by thedata. While the

F-test associated withthe overallmodel was significant(F =26.72, p < .001), the

coefficientsof the interaction termswere not. 20 Theresults of the hierarchical regression

analysis forthe predictors of viability are foundin Table 3 .10.

Repeated Measures The General Linear Model Repeated Measuresprocedure was conducted to check forresponse differencesover time, because participantswere measuredtwice during the semester. Withthe exception of theavoiding strategy(F = 5.87, p < .05), no time differences were noted forconflict-management strategies, conflict types, viability, or performance.

20 Similarresults were foundfor Time 2 data. See AppendixB, Table B.11. 75 Table 3.10. Hierarchicalregression analysis forpredictors of viabilitywith conflict-m anagement strategies andrelationship conflict attime 1.

Step Predictor Variable t R F dfs �F

Compromise Strategy .05 .61 Avoiding Strategy -.05 .49

�:- 1fi}, �; � 1-1 11 liiill1JB lilaa�1: Compromise Strategy -.06 -.76 Avoiding Strategy -.02 -.24 Relationship Conflict -.60 -6.17*** 1 · ·• -1111119rllli&i rllJllll��i�t�l�lli!111li Compromise Strategy -.08 -.99 Avoiding Strategy .00 .04 Relationship Conflict -.63 -6.39*** CollaborativeX Conflict -.20 -1.59 Compromise X Conflict .12 .99 Avoiding X Conflict .00 -.04

Note: Step 1 is a test of the linearrelationship between Conflict Strategies andViability. Step 2 adds Relationship Conflict to theequation. Step 3 tests theinteraction between Conflict Strategies and Relationship Conflict.

+ p s .10; • p s .05; •• p s .01; ••• p s .001.

76 Table 3.11. Summaryof support forhyp otheses.

la) Relationship Conflictand Viability (correlation) Yes

1 b) Task Conflictand Viability ( correlation) Yes

1 c) Process Conflictand Viability (correlation) Dropped : fromanalysis ld) Relationship Conflict + Task Conflictand Viability Yes (test of differencesbetween correlations)

2a) Relationship Conflictand Performance (correlation) No

2b) Task Conflictand Performance ( correlation) No

2c) Process Conflictand Performance ( correlation) Dropped fromanalysis 2d) Task Conflict+ Relationship Conflictand Performance No (test of differences between correlations)

3a) Collaborating + Compromising +Avoiding Strategieswith Yes Performance( simpleregression) 3b) Collaborating+ Compromising + AvoidingStrategies with Viability Yes (simple regression) 3c) Conflict-Management Strategies+ Task Conflict+ Interactionswith Performance(hierarchical regression) No 3d) Conflict-ManagementStrategies + Relationship Conflict + Yes Interactionswith Performance (hierarchical regression); Collaborating, Compromising,A voidingStrategies, and Performance (correlations)controlling for level of relationship conflict 3e) Conflict-ManagementStrategies + Relationship Conflict + No Interactions with Viabili ierarchical re · ess1on

77 CHAPTER 4

DISCUSSION

Thisexploratory study weaves conflict, conflict-management strategies, and team

outcomes together. To date, no published study has examined the relationship between

the typesof conflictperceived by the team, the conflict-management strategiesreportedly

used by the team,and team performanceand viability. Thisstudy of 77 instructional

groupsyielded fourmain findings: as predicted, members' aggregatedreports of group I) viability correlatedwith two types of reportedconflict, task andrelationship; reported 2) use of integrativestrategies for managing conflict( collaboration andcompromise) both correlatedpositively with viability; contraryto predictions, group performancedid not 3) correlatewith the two types of reported conflict;and the collaborative strategywas 4) positively correlatedwith teamperformance when perceived relationship conflict was high.

Thenegative correlationof relationship conflictwith groupviability (the team's capacity to continue operating together in the future), which includes member satisfaction andmember participation (Sundstrom, DeMeuse, Futrell, 1990), is consistent with a & study by Jehn (1995) (Hypothesis la). She foundthat satisfaction, liking of team members, and intent to remain in the team were negatively related to relationship conflict in freighttransportation groups.

Thefinding of a negative relationship between task conflictand viabilityis consistent with a study by Amason andSchweiger ( 1994) which foundthat task conflict

78 was associated with frustration, dissatisfaction, andanxiety within top managementteams

(Hypothesislb ). Similarly, Jehn (1995) founddissatisfaction and intent leave the group

related to task conflict.

Tests confirmedthe prediction that relationship conflictwould have a stronger

association with groupviability thantask conflict(Hypothesis Id). While Jehn (1995)

foundrelationships between satisfactionand intent to remain in the groupfor both

relationship andtask conflicts, no study has examined which of these types of conflictis

more predictive of viability.

Thisstudy foundno supportfor hypothesespredicting relationships between

performanceand reported types of groupconflict (Hypotheses 2a, 2b, and 2d). Earlier

studies foundevidence for optimal levels of task conflictfor performance (Pondy, 1967;

Brown, 1983; Jehn, 1995). Otherresearch foundan inverse relationship betweenconflict andperformance (Evan, 1965; Amason, 1996). Thelack of a relationship with performancein this study could reflectany of at least three factors: 1) unreliability of performancecriteria; 2) restrictedvariability in performance; 3) invalid measure of conflict. Thesefactors are discussed at length in the limitations section.

Support was foundfor more integrative conflict-management strategiesbeing related to higher viability and performance(Hypotheses 3a and3b ). Past empirical research foundthat negotiators using more integrative conflict-management strategies have increases in satisfaction(Wall & Nolan, 1986) andperformance (Thompson, 1990), and teams reportingmore integrativeconflict-management strategies were foundto have

79 higherviability and performance than teamsreporting less integrativestrategies

(Vigil-King & Rush, 1998).

Supportwas not foundfor the prediction that performancein teamsusing a more integrativeconflict-management strategy would be higherthan in those using a less integrativestrategy when higherlevels of task conflictwere perceived by the team

(Hypothesis3c ). Increases in performancehave been linked to the use of integrative conflictstrategies (Thompson, 1990) andtask conflicthas been linked with performance

(Brehmer, 1976; Jehn, 1995), but this study examined these relationships in combination andthe predictions made were not supported.

On the other hand, supportwas foundfor the prediction that performance of teams usinga more integrativeconflict-management strategy would be higherwhen higher levels of relationship conflictwere perceived (Hypothesis3d). While this study examinedthe relationships between integrativeconflict-management strategies and relationshipconflict with performance, increases in performance have been linked to the use of integrative conflict strategies(Thompson, 1990) and relationship conflicthas been linked with performance(Jehn, 1995; Amason, 1996).

These�dings were further examinedby dividing relationship conflict into three almost equal groups(l ow, moderate, andhigh levels). Under conditions of high relationship conflict, the collaborative strategyhad a significant positive correlation with performance( suggesting a positive trendin performanceassociated with the most integrativestrategy), while the compromising andavoiding strategieshad non-significant negative relationships with performance. Under moderate levels of relationship conflict,

80 both the compromising andthe avoiding strategieshad significant negative relationships withperformance ( suggesting a negative trendin performance associated with theless integrativestrategies), while the collaborating strategy had a non-significant positive relationship with performance. Theseadditional analysesadded some support to the hierarchical regression results.

No support was foundfor the prediction that the viability of teamsusing more integrativestrategies would be higherthan of thoseusing less integrative strategies when higher levels of relationship conflict were perceived by the team (Hypothesis3e ).

Relationshipconflict appeared to have a counter additive (negative) association with collaborationin its relationship to viability. Increasesin satisfactionand intent to remain in thegroup have been linked to theuse of integrativeconflict strategies (Thompson,

1990) andrelationship conflicthas been linkedwith viability(Jehn, 1995).

Limitations

Generalizability

Theexternal validity of studies using student samples is periodically debated in the literature (Dobbins, Lane,& Steiner, 1988; Gordon, Slade, & Schmitt, 1986) andcalls intoquestion the generalizabilityof results to other populations. However, unlikemany other studies examiningteams, these 77 intact instructional teams were working towards a common goal (thegrade on a groupproject) andworked together fora minimumof

8 weeks withminimal interference from their instructors. Their team-based projects had highfidelity to anorganizational setting. Students were expected to create and deliver

81 multiple business presentations to facultyand members of externalcompanies, to contact externalvendors, andto implementstrategies for completing projects.

Multiple populations of student participantswere also included in this study. The

77 intact teams were fromthree separate programs, representing three different disciplines. Specifically,the majority of the teams were fromthe :MBA, MAcc, and engineering programs, andall were fromsenior-level undergraduateor graduate-level courses. Itis felt thatthe structure of the team projects andthe composition of the teams contribute to the generalizability of thisstudy.

Method Variance

All of the variables assessing conflict-managementstyles, conflicttypes, and viabilitywere collected twiceduring the semester, using a self-reportsurvey instrument.

Theseself-report surveys were the primarysource of data for this study, and their use may result in the inflationof some of the relationships. However, subjective performance measures (and in manycases multiple subjective measuresof performance)were gathered from the instructorsduring the semester to minimize these effects.

Sample Size

Ideally, it would have been desirable to increase thenwnber of participants.

Whiledata regarding 77 teams was collected, more teams are necessary. A power analysiswas conducted on the data for this study which yielded a 50-60% probability of

21 detecting a medium effect size (12 = .15). Since four variables and their interactions were of interest, a sample size of 104 or greaterwould be ideal (.80 probability of

21 See Appendix A, Table A.3, foranticipated sample sizes forpowers .80, .85, and .90. 82 detection). Approximately 90 teams were solicited, and monetary opportunitieswere offeredas incentives. Still, some teammembers chose not to participate, andthe members of theirteams whodid so usually indicated that their team was performing poorly andwas experiencing a great deal of conflict. Therewas usually a member of the non-participatingteams who was interested in participating, but this member was unable to encouragethe remaining members to do so.

Measurement of ConflictManagement

TheROCI-11 {Rahim, 1983) was the conflict-managementscale chosen forthis study. It was selected over thesimilar MODE scale, developed by Thomasand Kilman

(1974), forseveral reasons: 1) it has received considerable attentionfrom theresearch community; 2) it is used andevaluated regularly; 3) it has been successfullymodified to address multiple targets (e.g., supervisor, peer, subordinate). However, VanDe Vliert andKabono:ff (1990) recently examined bothof these scales empirically and concluded that each of the scales has some problems: 1) the MODE differentiatespoorly (both theoretically andpractically) between thecollaborative andcompeting styles; 2) the

ROCI-11 discriminates poorly between the compromising andcollaborative styles; 3) both scales failto distinguishbetween theavoiding andaccommodating styles. Theyconclude thatboth the MODE andthe ROCI-II are "moderately valid measurementsof theconflict managementtheory underlying their construction" with"considerable room for improvement" (VanDe Vliertand Kabonoff, 1990; p.206). Thus,it is possible that improved measures might increase the magnitudeof support forseveral of thehypotheses of this study.

83 Measurement of ConflictTypes Whiletask conflictand relationship conflicthave been predominately studied in the literature (see Guetzkow andGyr (1 954), Pinkley (1990), Priem andPrice (1991), and

J ehn ( 1992) forsome of thesestudies), recently a third typeof conflicthas been hypothesized(Jehn, 1992; Jehn, 1997). Jehn's relationship, task, andprocess conflict scales were used to assess the frequencyof theseconflict types within teams. Based on a published factoranalysis (Jehn, 1992), it was thought that additional items mighthelp to improve the existing scales.

However, when using theconflict scales with theadditional items in this study, problems still existed. For example,while the coefficientalphas forthese scales were above.85, a principalcomponents factoranalysis with varimax rotation failedto separate the items intotheir respectivescales. Inaddition, task conflictand relationship conflict were successfullyaggregated to the group level, but only 28% of the teams agreedon the level of perceived process conflictthey were experiencing. While J ehn' s task and relationship conflictscales have been used in multiple studies (Jehn, 1992; Jehn, 1995;

Amason, 1996), her process conflictscale has not been used as much. Future research mightinclude refinement of the process conflictscale to elicit better agreementbetween team members for group-level analyses.

Measurement of Performance Team performancemeasures were subjective ratings of performancegathered frominstructors. While the performanceof several teams was measured only once at the end of thesemester, most team measures included multiple ratings of performance

84 collected at varyingtimes throughoutthe semester. Still, there was little variancein

performance, which may have contributedto thedifficulty in finding relationships

between performanceand othervariables such as relationship andtask conflict. Also, participants were upper-level students who were expected to performwell. Theresulting

lack of poorly-performingstudents may have contributed to the inability to find

significantrelationships between performanceand other variables. Statistical support for restrictionin rangeis shownby thenon-significant correlation between performance at

Time I andperformance at Time 2 (-.03, p = .80).

Overall, it is feltthat threelimitations made it difficultto discover and confirm relationships between variables. Sample size was a problem; thepower (.80 ) needed to detect a medium effectsize requireda minimum of I 04 teams, andonly 77 teams participated in this study. Several of the of scales were also in need of refinement; while the coefficientalphas forthe scales measuring the threetypes of conflict were solid, some indications of possible problems include: thefactor analysis; the number of low rwgS for process conflict; andthe sometimes awkwardstructure of theitems. Finally, gradeswere used as measures of performance. While the use of gradesby instructorsto rate performanceparallels the ratings givenby supervisorsin the workplace,it is likely there wasn't enough variancein performanceto detect significantrelationships between variables. Specifically,the use of upper-levelstu dents restrictsthe range in performance because few(if any) students/teams are expected to performpoorly. It could be suggested that the performancemeasure be augmentedby some other measure of performance in future studies.

85 Questionsfor Future Research What are the typ es of conflict? Jehn (1997) suggests that conflictcan be separated into threetypes of conflict, and her conflicttypology provides a conceptual frameworkto guide in the study of conflict. However, her measures of relationship, task, andprocess conflicthave not been as rigorously tested as the ROCl-11,and problems occurredwith theprocess conflictscale when aggregatingthis measure to the group level. Specifically, only 28% of the teams surveyedwere in agreementconcerning the amountof process conflictthey were experiencing. In contrast, the teams were much more in agreementon task conflict(9 1 % agreement)and relationship conflict (82% agreement), andpast empirical research supportsthe idea of only two types of conflict(Wall & Nolan, 1986;

Priem & Price, 1991 ). Futureresearch needs to clarify typesthe of conflictthat exist.

Whatcon flict-management strategywill actuall y increaseperf ormance in situationsof high conflictwithin a team? Takingthe research by Jehn, Amason, and others a step further,Thomas' taxonomy of strategic intentions was used to examine internalgroup processes. Thisstudy foundrelationships between types of conflict, conflict-management strategies, groupand performance and viability, but no causal links.

While manyof thefindings in this study were consistent withthose fowid at the individual anddyadic levels (e.g ., more integrativestrategies were associated with higher viability), others were not (e.g., more integrativestrategies were not necessarily related to higherperformance). Future researchshould therefore continue examininginformal conflictat the grouplevel andattempt to findcausal links between the type of conflict, groupconflict-management strategies, andgroup out comes.

86 Whatconflict-management styles are most preferred by team members? The

conflictliterature has linked individual conflict-managementstrategies to personality

dispositions. For example, previous research has foundthat individuals with a high need

for affiliation tend to use the accommodating strategyand not the competing strategy

(Jones, & Melcher, 1982) andindividuals with a high need for achievement tended to use

the collaborative strategy(Bell & Blakeney, 1977). However, the teamconflict

management compositionwas not examined in thisstudy. Futureresearch should

examinewhich conflict-management styles are most preferredby team members and how

they are related to the group-levelstrategy reported by the team.

Could rwg be usedas an independent surrogate measure of groupconflict within

teams? A significant andnegative correlationwas foundbetween the rwg for the three types of conflictand the actual scale measures of these types of conflict (-.52, p <.001;

-.32, p <.01; -.31, p < .001 forrelationship, task, and process conflict, respectively).

Thesecorrelations suggest thatthe aggregate measure of conflictmight be a good group

level surrogatemeasure of conflict. Inthe current study, groupswith low agreement

about the level of conflicthad the highest levels of reportedconflict, while the groups with the lowest reported levels of conflict showed the highest agreement. Conflictis naturallyabout disagreement, so the lack of agreementamong groupmembers regarding the threeconflict types is itself a typeof conflict. Future researchshould examine the possibilityof using rwg as a group-levelmeasure of conflict. Also, the elimination of variables in conflictstudies based on rwg should be reconsidered.

87 How does the perceived intensityof conflictwithin a team relate to the group 's outcomes? The conflictmeasures used in this study examinedthe perceived frequencyof relationship andtask conflictwithin the groups. Another dimension of conflictthat may bearexamination would be theperceived intensity of the conflict andhow this perception is associated with the performanceand viability of the groups. Future research could include the development of conflictscales that incorporate intensity as well as frequency into the measure.

Are teams concurrently using more than one type of conflict-management strategy? Typically,the study of conflict-managementstrategies to date has focusedon identifyingthe one strategyused most by the individual (Van de Vliert& Kabanoff,

1990). Both the ROCI-IIand the MODE scales are scored such that one conflict managementstrategy emerges as the preferredone (Rahim, 1983a; Thomas& Kilman,

1974). It is possible that several strategiesmay be employed at the same time. For example, team members might employ a combination of both the collaborative strategy

(because at the individuallevel it has been found to improve performanceand satisfaction)and the avoiding strategy(because it is a good strategyto use when trying to prevent the escalation of conflict) in situations of highconflict. Future research should examine the possibility of two or threestrategies being employed in combination.

How do the distributiveconflict-management strategiesrelate to team outcomes?

Thethree integrativeconflict-management strategies of collaborating, compromising,and avoiding were the focusof this study. These three strategies are based on the idea that each party involved has a near-equal level of concernfor self as they have concernfor

88 others. Integrativeconflict-management strategies are characterized by trustand openness, andare more congruentwith teamwork. On theother hand,distributive conflict-managementstrategies, which are characterized by coercion, entrenchment,and manipulation, are grotmdedin theidea that one party's gainwill be at the expense of the other party (Lewicki & Litterer, 1985). Theyare not congruentwith teamwork, andtheir use increases the likelihood of mistmderstandingand hostility (Pruitt, 1971; Thompson&

Hastie, 1990). Thismistrust and hostility can lead to bitternessand the desire to avoid working with others in thefuture. While distributive conflict-management strategiesare not conducive to teamwork, team members may utilize them when dealing with others.

Futureresearch should examine the role of distributiveconflict- managementstrategies withinteams and their impact on groupviability andperformance.

Whatother things should be considered when conducting groupcon flict research? While this studyfotmd relationships betweenconflict-management styles and viability andbetween conflict-management styles and performance, tests of the hypothesizedinteractions yielded mixed results. A replication of thisstudy using additional teams would help to clarifythese relationships. An organizationalfield study examiningthese relationships could address the issues surroundingthe use of student populations andallow for theuse of more objective measures of performance (such as production rates, scrap rates, and other quality assurance measures) instead of subjective measures of performance(i.e., supervisorratings).

More longitudinal research would also help improve this study; not because there wasn't a longitudinal element, but because participantswere measured twiceduring the

89 semester. Many of the findingsfor viability at Time 1 remained at Time 2. However, the performancefinding at Time 2 was inconsistent with Time 1, andwas also inconsistent with the predicted/hypothesizedrelationships. More researchis needed to understandthe minute changesin performanceover the courseof a project. Futurelongitudinal research should focuson measuringconflict, confli ct-managementstrategies, andperformance at multiple times during the course of a project prior to its conclusion.

Method variance was anothercited limitation. Since team facilitatorsare becomingmore commonplacein organizations,it might be usefulfor future research to have these facilitatorsrate their team(s) on the amountand typeof conflictthey feel the teamis experiencing, as well as the typesof conflict-managementstrategies they are using. This would allow foradditional (and possibly less biased) measures of conflict andconflict-management strategies.

Problems with the scales used to measureconflict-management strategiesand conflicttypes were also discussed as limitations. VanDe Vliert andKabonoff (1 990) mentioned in their review of theROCI-II that theconflict-management scales suffered fromsome problems andcould be improved. One of the reasons that Rahim's conflict managementscale was selected is because it has been continually used, studied, and revised to reflectdifferent targets over the 25 yearsof its existence. Thus, while thescale is not a perfect measure of conflict-management strategies, it is a relatively good one.

Practical Implications

If replicated, theresults of this study have implications formanaging conflict within teams, andsuggest thata delicate balanceexists within a group or teani. For

90 example, underhigh levels of conflict, the collaborative strategywas positively correlated

with performance, when controllingfor the level of relationship conflict. Also, wider

moderate levels of conflict, the compromising andavoiding strategieswere negatively

correlatedwith performance,when controllingfor the level of relationship conflict. Thus,

the conflict-managementstrategy reportedby a team related differentlyto performance, depending on the level of conflict perceivedby the team members. Furtherstudy is needed to determine if there is a causal link(this study only shows anassociation), and if team members may be able to improve team performanceby monitoring ( and possibly altering)their conflict-managementstrategy based on the perceived level of conflict.

Thisstudy also has organizational implications forteamwork. While the collaborating and compromising strategieswere positively correlated withviability (the team members' satisfactionand the te am'scapacity to continue working together in the future),the collaborative conflict-managementstrategy was more predictive of the team's viabilitythan either the compromising or avoiding strategies. Even when controllingfor relationship conflict, the collaborativestrategy was still positively related to viability.

Thesefindings are consistent with previousresearch, which foundthat greater satisfaction was associated withthe use of more integrativeconflict-management strategies (Wall &

Nolan, 1986). A causal link (if one can be established) between the collaborative strategy and viabilitymight provide a method for improving the "health" of a team. Given that teams arebecoming more centralto the structure of organizations(Hackman, 1990;

Lawler, Mohrman,& Ledford, 1995), informationof this naturecould prove invaluable to managers who are interested in findingout what they can do thatwould allow successful

91 teams to continue working together in the future, thus potentially improving their

organizationaloutcomes.

Conflict-managementstrategies are also a proactive method fordealing with

conflictwithin organizationalgroups, because they are a resource that is readily available

to all disputants; thechallenge comes in using themeffectively to reconcile differences

(Kilmann& Thomas, 1978; Brett, Goldberg, & Ury, 1990). Another possible business

use of thefindings of thisstudy mightbe to teach members of organizationalteams to

recognizethe typeand level of conflicttheir team is experiencing andwhich of thefive

conflict-managementstyles (or combinations of them) could assist themin maximizing

their teamoutcomes, which in turnmay be positivelyrelated to organizational outcomes.

In a recent surveyof U.S. firms, it was foundthat the use of interpersonalskills

training(which oftenincludes conflict-management training)has increased by 30%-40%

over thepast several years(Filipczak, 1994). Withan estimated $50.6 billion being spent

on formaltraining programs, managerswant to know that their trainingdollars are being

spent wisely. The results of thisstudy tie theintegrative conflict-management strategies

presented in the trainingto team outcomes like viability and performance. However, the

effectivenessof thesetraining interventions ( as withany interpersonal skills training) will

be dependent on themaintenance mechanisms inherent in the training process. For example,post-training interventions that focus on self-managementand goal-setting have emerged as successfulmethods of trainingmaintenance (Wexley and Nemeroff, 1975).

While relationships between conflict-managementstyles and viability and between conflict-managementstyles and performancewere foundin groupsinteracting

92 face-to-face,these relationships maynot hold truefor virtual teams, whose members communicate via electronic mail ( e-mail) from separate locations, cities, countries, etc.

For example, McGuire, Keisler, and Siegel (1987) examined risk-taking amongcorporate managers and foundthat electronic communication increased risk-taking behavior among corporatemanagers, which in tum could alter levels of performance.

Electronic communication (and especially e-mail) opens the door to miscommunicationand misinterpretation because it creates an impoverished social environment (Thompson, 1998). The social cues and sharedexperien ces that bind individuals to each other are absent, and a social context that creates and fuels feelings of anonymity is fostered. In addition, the ephemeral nature of electroniccommunication must not be ignored. E-mail messages are frequently fired back and forth rapidly among the recipients, so the ease and speed of e-mail makes it easy for individuals to forget their audience and ignore typical face-to-face courtesies such as social boundaries, politeness rituals, andthe acknowledgment of other's viewpoints andcontributi ons (Sproull and

Keisler, 1991). This impoverished social environment provides a foundation for conflict to growand makes it more difficult for collaboration (which is based on shared viewpoints and acknowledgement of all team members' contributions) to minimize conflict within teams.

Conclusion

Inconclusion, this study of 77 instructional groups considered conflict, conflict management strategies, and team outcomes together, something no published study has previously examined. As predicted, team members' aggregated reports of group viability

93 correlatedwith reported task andrelationship conflict. Thereported use of the collaborative andcompromising conflict-managementstrategies correlated positively with team viability. Groupperformance did not correlatewith task or relationship conflict,contrary to predictions. Finally, when perceived relationshipconflict was high, thecollaborative strategycorrelated positively withteam performance.

94 REFERENCES

95 REFERENCES

Amason, A. C. ( 1996). Distinguishingthe effectsof functionaland dysfunctional

conflicton strategicdecision making: Resolving a paradox fortop management

teams. Academy of Management Journal, 39, 123-253.

Amason, A. C. & Schweiger, D. M. (1994). Resolvingthe paradox of conflict, strategic

decision makingand organizationalperformance. InternationalJournal of

ConflictManagement, �, 239-253.

Argote, L., & McGrath, J. E. (1993). Groupprocesses inorganizations: Continuity and

change. In C. L. Cooper & I. T. Robertson(Eds .), Internationalreview of

industrialand organizational psychology, (Vol. 8, pp. 333-389). John Wiley &

Sons Ltd.

Bales, R. F. (1958). Task roles and social roles in problem-solving groups. InE. E.

Maccoby, T. M. Newcomb, & E. L. Hartley(Eds .), Social Psychology, (Vol. 3,

pp. 437-447) New York: Holt, Rinehart& Winston.

Baron,R. A. ( 1991 ). Positive effectsof conflict: A cognitiveperspective. Employee

Responsibilities and Rights Journal, � 25-36. ' Bell, E., & Blakeney, R. (1977). Personality correlatesof conflictresolutions modes.

Human Relations, 30, 849-857.

Blake, R., & Mouton, J. (1984). TheManagerial Grid. Houston, TX: Gulf Publishing.

96 Bourgeious,L. J. (1985). Strategicgoals, environmentaluncertainty, andeconomic

performancein volatile environments. Academyof ManagementJournal, 28,

548-573.

Brehmer, B. (1976). Social judgmenttheory and the an alysis of interpersonalconflict.

Psychological Bulletin,..fil_, 985-1003.

Brett, J. M., Goldberg, S. B., & Ury,W. L. (1990). Designingsystems forresolving

disputes inorganizations. AmericanPsyc hologist, 45, 162-170.

Brown,L. D. (1983). ManagingConflict at Organizational Interfaces. Reading, MA:

Addison-Wesley.

Campion, M. A., Medsker, G. J. & Higgs, A. C. (1993). Relations between work group

characteristics and effectiveness: Implicationsfor designing effective work

groups. Personnel Psychology, 46, 823-850.

Carnevale,P. J., andPruitt, D. G. (1992). Negotiation andmediation. Annual Review of

Psychology, 43, 531-582.

Conger, A. J. (1974). A revised definitionfor suppressor variables: A guide to their

identificationand in terpretation. Educational andPsychological Measurement,

34, 35-46.

Eisenhardt, K. & Schoonhoven, C. (1990). Organizational growth: Linking founding

team, strategy,envi ronment, and growthamong U.S. semiconductor ventures,

1978-1988. Administrative Science Quarterly, 35, 504-529.

Evan, W. (1965). Conflictand performance in R & D organizations. Industrial

ManagementReview, 1, 37-46.

97 Fiol, C. M. (1994). Consensus, diversity, and learningin organizations. Organization

Science, �' 403-420. Filipczak, B. (1993). Armedand dangerous at work. Training, .Q, 39-43. Filipczak,B. (1994). Looking past the numbers. Training,.ll, 67-72, 74.

Forsyth, D. (1990). Group Dynamics. Needham Heights, MA: Allyn& Bacon.

George, J. M. (1990). Personality, affect,and behavior in groups. Journalof Applied

Psychology,75, 107-116.

George, J. M., & Bettenhausen, K. (1990). Understanding prosocial behavior, sales

performance, and turnover: A group-levelanalysis in a service context. Journal

of AppliedPsychology, 75, 698-709.

George, J. M., Perkins, M., Sundstrom,E., & Myers, S. D., (1990). Work team context

development andeffect in a manufacturingorganization. Proceedings of the 1990

InternationalConf erence on Self-managed Work Teams, Denton, TX.

Gersick, C. J. G. (1989). Markingtime: Predictable transitionsin task groups. Academy

of Management Journal, 32, 274-309.

Gladstein, D. (1984). Groups in contest: A model of task groupeffectiveness.

AdministrativeSc ience Quarterly, 29, 499-517.

Gordon, J. (1992). Work teams: How far have they come? Training, 29, 59-65.

Greenhalgh,L. (1986). Managingconflict. SloanManagement Review, 27, 45-51.

Guetzkow, H. & Gyr,J. (1954). Ananalysis of conflictin decision-making groups.

Human Relations, 1, 367-381.

98 Hackman, J. R. (1990). Groups that work {and those that don't): Creating conditions for

effectiveteamwork. San Francisco, CA: Jossey-Bass.

Hackman, J. R. & Morris, C. G. (1975). Group tasks, groupinteraction process, and

groupperformance effectiveness: A review and proposed integration. In L.

Berkowitz (Ed.), Advancesin Experimental Social Psychology, (Vol. 8, pp. 45-

99). New York: Academic Press.

Hackman,J. R., & Walton, R. E. (1986). Leading groupsin organizations. In P. S.

Goodman(Ed.), Designingeffective work groups. San Francisco: Jossey-Bass.

Hammond, K. R., Todd, F. J., Wilkins, M., & Mitchell, T. 0. (1966). Cognitiveconflict

between persons: Application of the "lens model" paradigm. Journal of

Experimental Social Psychology, � 343-360. ' Hannon, J., & Rohrbaugh, J. (1990). Social judgmentanalysis and small groupdecision

making: Cognitivefeedback effects on individual and collective performance.

OrganizationalBehavior andHuman Decision Processes, 46, 34-54.

Hays, W. J. (1981). Statistics. New York: Holt, Rinehart& Winston.

Hill, G. W. (1982). Group versus individual performance: Are N + 1 heads better than

one? Psychological Bulletin, 91, 517-539.

Jackson, S. E. & Schuler, R. S. (1985). A meta-analysis and conceptual critique of

research on role ambiguity and role conflictin work settings. Organizational

Behavior andHuman Decision Processes, 36, 16-78.

James, L. R. (1982). Aggregationbias in estimates of perceptual agreement. Journal of

AppliedPsychology, 67, 219-229.

99 James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimating within-group interrater

reliability with andwithout response bias. Journal of AppliedPsychology, 69, 85-

98.

Janis,I. L. (1982). Groupthink. Boston: Houghton-Mifflin.

Jehn, K. A. (1992). Theimpact of intragroupconflict on effectiveness: A multimethod

examination of the benefitsand detriments of conflict. Unpublished doctoral

dissertation, NorthwesternUniversity.

Jehn, K. A. (1995). A multimethod examination of the benefitsand detriments of

intragroup conflict. AdministrativeScience Quarterly, 40, 256-282.

Jehn, K. A. (1997). A qualitative analysisof conflicttypes and dimensions in

organizational groups.Administrative Science Quarterly, 42, 530-557.

Jones, A. P., & James,L. R. (1979). Psychological climate: Dimensions and

relationships of individual andaggregated work environment perceptions.

Organizational Behavior andHuman Decision Processes, 23, 201-229.

Jones, R. E.,& Melcher, B. H., (1982). Personality andthe preference formodes of

conflictresolution. HumanRelations, ,li, 649-658.

Jewell, L. N. & Reitz, H. J. (1981). Group Effectivenessin Organizations. Scott,

Foresmanand Company.

Kabanoff,B. (1991). Equity, equality, power, and conflict. Academyof Management

Review, 16, 416-441.

Katz, D., & Kahn, R. L. (1978). Thesocial psychologyof organizations(4th ed.). New

York: Wiley.

100 Kiesler, S. (1978). InterpersonalProcesses in Groups andOrganizations. Arlington

Heights,IL: AHMPublishing.

Kihnann, R. H., & Thomas,K. W. (1978). Four perspectives on conflictmanagement:

An attributionalframework for organizing descriptive andnormative theory.

Academy of ManagementReview, � 59-68. ' Kozlowski, S. W. J., & Hattrup, K. (1992). A disagreementabout within-group

agreement: Disentanglingissues of consistency versus consensus. Journalof

Applied Psychology, 77, 161-167.

Kramer, R. M. (1991). Intergrouprelations andorganizational dilemmas: Therole

categorization processes. InL.L. Cummings andB. M. Staw (Eds.), Research in

OrganizationalBehavior, .U.,191-228. Greenwich, CT: JAI Press. Lawler, E. E., Mohrman, S. A., & Ledford, G. E. (1995). Employee Involvement and

Total QualityManagement: Practices andResults in Fortune 1000 Companies.

San Francisco: Jossey-Bass.

Lax, D., & Sebenius, J. (1986). TheManager as Negotiator: Bargaining forCooperation

andCompetitive Gain. New York: Free Press.

Levine, J. M., & Moreland, R. L. (1990). Progress in small group research. Annual

Review of Psychology, 41, 585-634.

Lewicki, R. J. & Litterer, J. A. (1 985). Negotiation. Homewood, IL:Irwin.

Losey, M. R. (1994). Managingin an era workplace violence. ManagingOffice

Technology, 39, 27-28.

101 McGrath, J. E. (1984). Groups: Interactionand Performance,Englewood Cliffs, NJ:

Prentice-Hall.

McGuire, T., Keisler, S., & Siegel, J. (1987). Groupand computer-mediated discussion

effectsin risk decision-making. Journal of Personalityand Social Psychology, 52,

917-930.

Netemeyer, R. G., Johnston,M. W., & Burton, S. (1990). Analysis of role conflictand

role ambiguityin a structural equations framework. Journalof Applied

Psychology, 75, 148-157.

Nunnally, J. C., & Bernstein,I. H. (1994). PsychometricTheory, New York, NY:

McGraw-Hill, Inc.

Pedhazur,E. J. (1982). Multiple Regressionin Behavioral Research. Orlando,FL: Holt,

Rinehart and Winston, Inc.

Pelled, L. H. (1995). Demographicdiversity, conflict andwork groupoutcomes: An

intervening process theory. Organization Science, §, 615-631.

Peterson, D. R. (1983). Conflict. InH.H. Kelley et al (Eds.), Close Relationships (pp.

360 396). New York: W. H. Freeman.

Pinkley, R. L. (1990). Dimensions of conflictframe: Disputantinterpre tations of

conflict. Journal of Applied Psychology,75, 117-126.

Pondy, L. R. ( 1967). Organizationalconflict: Concepts and models. Administrative

Science Quarterly, 12, 296-320.

102 Priem, R., & Price, K. (1991). Process and outcome expectations forthe dialectical

inquiry, devil's advocacy, andconsensus techniques of strategicdecision making.

Groupand Organization Studies, 16, 206-225.

Pruitt, D. (1981). Negotiation behavior. New York: Academic Press.

Pruitt,D. G. (1971). Indirect communicationand the search for integrativeagreement in

negotiation. Journal of AppliedSocial Psychology,1, 205-239.

Pruitt,D. G. (1983). Strategicchoice in negotiation. AmericanBehavioral Scientist, 27,

167-194.

Pruitt, D., & Rubin, J. (1986). Social conflict: Escalation, stalemate, andsettlement.

New York: Random House.

Pugh,D. S., Hickson, D. J., Hinings, C. R., & Turner, C. (1968). Dimensions of organizationstructure. AdministrativeScience Quarterly, .U,65- 105. Putnam,L. L. (1994). Productive conflict: Negotiationas implicit coordination. International Journalof Conflict Management,.2., 285-2 99.

Rahim, M. A. (1983a). Rahim organizational conflictinventories. Palo Alto, CA:

Consulting Psychologists Press.

Rahim, M. A. (1983b). A measure of styles of handling interpersonalconflict. Academy

of ManagementJournal, 26, 368-376.

Rahim, M. A. (1986). Managingconflict in organizations. New York: Praeger.

Rahim, M. A., & Magner,N. R. (1995). Confirmatoryfactor analysis of the styles of

handlinginterpersonal confli ct: First-order factormodel and its invarianceacross

groups. Journal of AppliedPsychology, 80, 122-132.

103 Rigdon, J. E. (1994). Companies see more workplaceviolence. Wall StreetJournal,

April 12, BI.

Roberts,K. H., Hulin, C. L., & Rousseau,D. M. (1978). Developing an interdisciplinary

science of organizations. San Francisco, CA: Jossey-Bass.

Rohrbaugh, J. (1979). Improving thequality of groupjudgment: Social judgment

analysisand the delphi technique. OrganizationalBehavior andHuman

Performance, 24, 73-92.

Rohrbaugh,J. (19 81 ). Improving thequality of groupjudgment: Social judgment

analysisand the nominal grouptechnique. Organizational Behavior andHuman

Performance, 28, 272-288.

Ross, R. S. (1989). Conflict. fuR. Ross & J. Ross (Eds.), Small Groups in

Organizational Settings (pp. 139-178). Englewood Cliffs, NJ: Prentice-Hall.

Schneider, B. & Bowen, A. E. (1985). Employee andcustomer perceptions of service in

banks: Replication andextension. Journalof Applied Psychology, 70, 423-433.

Schweiger, D. M. & Sandberg, W. R. (1991). A team approach to top management

strategicdecisions. fuH. E. Glass (Ed), Handbookof BusinessStrate gy, .Q, 1-20.

New York: Warren, Gorham, and Lamont.

Schweiger, D., Sandberg, W., & Rechner, P. (1989). Experiential effects of dialectical

inquiry, devil's advocacy, andconsensus approaches to strategicdecisions

making. Academy of Management Journal,32 , 745-772.

Schwenk, C. & Valacich, J. S. (1993). Effectsof consensus and devil's advocacy on

strategicdecision-making. Journalof Applied Social Psychology, 23, 126-139.

104 Sessa, V. I. (1996). Using perspective taking to manage conflictin teams. Journalof

Applied Behavioral Science, 32, 101-115.

Shah,P.P., & Jehn, K. A. (1993). ·no friendsperform better than acquaintances? The

interactionof friendship,conflict, andtask. Group Decision and Negotiations, 2,

149-165.

Sproull, L. & Keisler, S. (1991). Connections: New ways of working in the networked

organization. Cambridge, MA: MIT Press.

Steven, M. J., & Campion, M. A. (1994). Theknowledge, skill, andability requirements

forteamwork: Implications forhuman resource management. J oumal of

Management, 20,

503-530.

Stone, R. A. (1995). Workplacehomicide: A timefor action. Business Horizon, 34, 17-

20.

Sundstrom,E. (1998). Challenges of supportingeffective work teams. In E. Sundstrom

(Ed.), SupportingWork Team Effectiveness: Best ManagementPractices for

Fostering HighPerformance, SanFrancisco, CA: Jossey-Bass.

Sundstrom,E., & Altman, I. (1 989). Physical environmentsand work-group

effectiveness.In L. L. Cummings& B. Staw (Eds.), Researchin organizational

behavior, (Vol. 11, pp. 175-209). Greenwich: JAi Press.

Sundstrom, E., De Meuse, K. P., & Futrell, D. (1990). Work teams: Applications and

effectiveness. AmericanPsychologist, 45, 120-133.

105 Surra,C., & Longstreth, M. (1990). Similarity of outcomes, interdependence, and

conflictin dating relationships. Journalof Personalityand Social Psychology, 59,

501-516.

Tabachnick, B. G. & Fidell, L. S. (1989). Using Multivariate Statistics. New York, NY:

HarperCollinsPublishers, Inc.

Thomas, K. W. ( 1976). Conflictand conflictmanagement. InM. D. Dwmette(Ed.),

Handbookof Industrialand Organizational Psychology, (pp. 889-935). Chicago:

RandMcNally.

Thomas, K. W. (1979). Conflict.In S. Kerr(Ed.), Organizational behavior (pp. 151-181).

Columbus, OH: GridPublications.

Thomas,K. W. (1992). Conflictand negotiation processes in organizations. In M. D.

Dlll1Dette& L. M. Hough(Eds. ), Handbook of Industrial andOrganizational

Psychology, (Vol. 3, pp. 651-717). Palo Alto, CA: ConsultingPsychologists

Press.

Thompson,L. (1990). Negotiation behavior andoutcomes: Empirical evidence and

theoretical issues. PsychologicalBulletin, 108, 515-532.

Thompson,L. (1998). TheMind andHeart of the Negotiator, New Jersey, Prentice-Hall.

Thompson,L. & Hastie, R. (1990). Social perception andnegotiation. Organizational

Behavior andHuman Decision Processes, 4 7, 98-123.

Tzelgov, J., & Henik, A. (1991). Suppression situations in psychological research:

Definitions, implications, andapplications. Psychological Bulletin, 109, 524-536.

106 Van de Vliert, E. & De Dreu, C. K. W. (1994). Optimizing performance by conflict

stimulation. International Journalof ConflictManagement, � ' 211-222. Vigil-King, D. C., (1999) Confirmatoryfactor analysis of interpersonalconflict handling

styles at the team level. Presented at the 14th Annual Conference of the Society

forIndustrial and Organizational Psychology.

Vigil-King, D. C., & Rush, M. C. (1998). Conflictstrategies: Theirrelationship to group

performanceand group viability. Presented at the 13th Annual Conference of the

Society forIndustrial and Organizational Psychology.

Wall, V. D., & Nolan, L., (1986). Perception of inequity, satisfaction, andconflict in

task-oriented groups. HumanRelations, 39, 1033-1052.

Walton, R. E., & Dutton,J. M. (1969). Themanagement of interdepartment conflict: A

model andreview. Administrative Science Quarterly, 14, 73-84.

Walton, R. E., & McKersie, R. B. (1965). A behavioral theoryof labor negations: An

analysisof a social interactionsystem. New York: McGraw-Hill.

Weider-Hatfield, D. (1988). Assessing the Rahim organizationalconflict inventory -II (ROCl-11). ManagementCommunication Quarterly, 1, 350-366. Wexley, K. N. & Nemeroff, W. (1975). Effectivenessof positive reinforcement and goal

setting as methods of managementdevelopment. Journalof AppliedPsychology,

60, 446-450.

107 APPENDICES

108 APPENDIXA

109 Table A.I. Factor analysis results andrevised scales measuringconflict-management styles.

Component Item 1 2 3 Avoiding 1. My team members attemptedto avoid being"put on the spot" andtried to .746 .103 keep theirconflict with other teammembers to themselves. 2. Most membersof my team triedto keeptheir disagreementswith each other .782 to themselves in order to avoid hardfee lings. 3. Members of my teamtried to avoid unpleasant exchanges eachwith other. * .280 .580 .204 4. Most membersof my team usuallyavoided open discussionsof their -.102 .785 differences with each other. 5. Mostmembers of my team triedto stay away from disagreements with each other. .806 .106 6. My teammates avoidedencounters with each other. -.168 .670

Collaborative I. Members of my teamtried to integratetheir ideasto comeup with a decision .768 jointly. .279 2. Mostmembers of my teamtried to work together for a proper understanding .758 .123 of a problem. 3. Most membersof my teamtried to investigate anissue to finda solution .642 .247 acceptableto us. 4. Most teammembers tried to work togetherto findsolutions to problems .807 which satisfied ourexpectations. .162 . 5. Members of my teamexchanged accurate informationwith each other to .779 .107 solve a problem together. 6. Members of my team collaborated with each other to come up with decisions .781 .187 acceptable to us. 7. My teammatestried tobring all our concernsout in the openso that the .571 -.176 .299 issuescould be resolved in the bestpossible way. *

Compromise I. My teammates triedto find amiddle course to resolvean impasse. .189 .105 .752 2. My teammatesnegotiated with each other so that a compromise could be 3 reached. .39 .673 3. Most members ofmy teamused "giveand take" so that a compromise could .241 .761 be made. 4. Members of my team usuallyproposed a middle ground for breaking .197 .175 deadlocks. .769

Note: Analysiswas principal components factoranalysis with varimax rotation. Only factorloadings greaterthan or equal to .10 are presentedin the table. Factors with eigenvaluesgreater than 1.0 were retained. N = 340. * = items eliminated from respectivescales.

110 Table A.2. Factor analysis results andrevised scales measuringthe three typesof conflict.

Component Item 1 2 3 4 Task Conflict 1. To what extent were there differences of opinion in your team? .729 -.320 -.153 .219 2. How often did your teammates disagree about opinions regarding the work being done? .826 3. How much conflict about the work you do wasthere in your team? .742 -.422 4. How frequentlywere there conflicts about ideas in yourteam? .725 -.258 -.227 .287 s. How often did yourteammates disagree about ideas regarding the task? .740 -.262 -.256 .229 6. To what extent were differences of opinion regardingtasks evident among your teammates? .779 .361 7. To what extent where there disagreements about the tasks you were working on with team members? .755 .218 -.232 8. How often did your teammates disagree about the work being done? .777 -.174 Relationship Conflict 1. How much frictionwas there among members of yourteam? 2. .813 -.128 .402 How muchtension was there amongteam members? .843 .343 3. How much were personality conflicts evident in yourteam? .828 .424 4. How muchemotional conflictwas there among teammembers? .806 .375 s. How much angerwas present in your team? • .796 .286 -.106 6. To what extent were personality clashes evident in your team? .789 .450 .102 Process Conflict 1. How often did your teammates disagree about how the work should be divided? .685 .550 2. How frequentlywere there disagreements aboutwho should do what in your team? .674 .530 3. How much disagreement wasthere about procedures in your team? • .810 .113 -.114 -.113 4. To what extent did you disagree about the way to do things in your team?• .762 -.261 s. How often didyour teammates disagreeabout whose responsibility it was to complete a task? .676 .591 .122 6. To what extent did your teammates disagree about the processes used to complete tasks? • .817 -.234

Note: Analysis wasprincipal components factor analysiswith varimaxrotation. Only factorloadings greaterthan or equal to .10 are presented inthe table. Factors witheigenvalues greaterthan 1.0 were retained. N = 340. * = items eliminated from respective scales.

111 Table A.3. Power analysis

Significance criterion a = .05

Desired power .80 .85 .90

14.35 forpower of .SO 16.04 forpower of .85 18.28 forpower of .90

2 2 I ; effect size 1 = .15 C

n * ; nwnber of subj ects d 104 forpower of .80 115 forpower of .85 130 forpower of .90

a based on the structural equation forh 1

6 fromTable E.2 in Cohen and Cohen (1983) c based on a medium effect size as definedby Cohen ( 1977) where R1 = .13 d basedon the following equation: n* = L -2 +k+I 1

112 APPENDIX B

113 Table B.1. Time 2 participantpopulation and return rateinform ation

Fall Sprine Time in semester Ind Teams RR Ind Teams RR when measured MBA 74 16 80% 63 12 73% 16 weeks

MAcc 58 13 69% 16 weeks

Engineers 68 17 85% 16 weeks

Undergrad 1 2 5 84% 8weeks

Undergrad 2 12 1 50% 16 weeks

Undergrad 3 10 2 38% 16 weeks

Totals 132 29 75% 175 37 72%

114 Table B.2. Summaryof perfonnancemeasures at time 2.

Population MBA MAcc En_gineers Undergraduate Business SubpopulaH�n Chemical I Other 2 3 1t,••ot:Teilm$i.;f;'.:, : ::!_�i 11- .. , Semester I Fall Spring Fall I Spring I Spring I Spring J�r!J!& Spring Performance Week 9to Week 9 to Week 16 I Week 9 to Week 16 Week 5 to I Week 15 Week 13 to Period Week 16 Week 16 Week 16 I I Week 8 Week 16 Perf. Scale EE, ME, BE EE1 ME, BE Percentage Percentage AtoF Percentl!_ge Percent�e Percentage # of Projects 3 2 2 2 2 Score Range -3 to 3 -2 to 2 0 to 100 0 to 200 -5 to 6 0 to 200 0 to 100 0 to 200 iffS�tlt;�&KiN

>..:M:emi· ;. so:> t •Mean.·- . ·.·.SD: >• I Mean,.:}� 0i.:1St>t.,Ah Mefili+":�,;/4/SFf);,,J.;��,. . $Rht1'�l·!ilvl�·. · ··.. -..: $t)y!j.:1Meun , ;;/t2SJ3)/,:l�>�

Variables MBA MAcc Levene Test

Collaborative Strategy . Compromise Strategy 3.55 3 62 .72 45 3.87 .66 3.78• .82 Avoiding Strategy I .81 I 45 I 3.15 I .81 1 1.18 I .44

'&-1, ••• 5. Relationship Conflict 137 11.32 4.32 58 1.16 3.53 68 9.35 4.33 44 8.16 3.93 13.49 1.19 137 5. 5.19 .. 6. Task Conflict 19.65 26 58 14.64 4.95 68 17.43 44 14.95 6.17 15.56* .28 7. Process Conflict 137 5.71 2.33 58 4.62 2.10 68 5.37 2.49 44 4.80 2.19 3.82* .65

Note: Viability and StrategyPredictors were measured using a 5-point Likert scale where 1 = "Strongly Disagree"and 5 = "Strongly Agree." ConflictPredictors were measured using a 5-point Likert scale where 1 = "None" and 5 = "Constant." *Q $ .05, 012 $ .01, *** p $ .001.

(a) Performancewas measured at the group level and standardized within the MBA,MAcc, Engineer, and Business student populations respectively. (b) Tests formean difference (ANOVA) between the MBA, MAcc, Engineer, and Business student populations respectively. (c) Tests for thehomogeneity of variance between the MBAstudent and Business student populations. Table B.4. Individual-level standardizeddescriptive statistics andzero-order correlations forall participantsand reliabilities at time2.

Variables 1 2 3 4 5 6 7

(.83)

5. Relationship Conflict .00 1.00 -.67** -.58** -.42** . 05 (.93) 6. Task Conflict .00 1.00 -.49** -.46** -.33** -.07 .72** (.92) 7. Process Conflict .00 1.00 -.47** -.49** -.33** .03 .68** .68** (.89) Note: All statistics were computed at the individual level; N = 308.

*Rs.05, **R s .01, ••• p S .001.

<•> Performancewas measured atthe group level and standardized within the MBA student andBusiness studentpopulations.

117 Table B.5. Distribution ofrwg forvariables at time2.

Variables S.59

.89 .03 .08 .18 .94 1.28 . 60 . OS .18 .42 Process Conflict .11 .06 .35 1.66 Note: AN OVA = analysisof variance. Total numberof groupsis 65. Estimates > . 70 areconsidered acceptabl e. With the exceptionof process conflict, overall 87% of these estimates arewithin acceptablerange. *R s .05.

118 Table B.6. Group-level descriptives for participating teams at time 2.

Variables · Team Descriptives

2. -1.24 to 1.4 7 3. -.92 to 1.20 4. -.87 to .97

5. -1.36 to 2.04 6. -1.39 to 1.45

119 Table B.7. Group-level descriptive statistics forsub-group populations at time 2. - - Variables MBA MAcc Engineering Teams I SD

28 I -.02 I .61 I -t.3 to t.5 I 13 I -.03 -.9 to 1.1 I 16 I -.09 I .70 -t.4 to 1.3 I 8 I -.03 I .57 I -.5 to 1.1 ..... N 0 Table B.8. Group-level descriptive statistics and zero-order correlations for all groupsat time2. <•>

Variables M ·, SD 1 2 3 4 5 6

6. Relationship Conflict -.05 .69 -.83••• -.12••• -.62••• .14 J.. TaskConflict -.04 .62 1 -.14 -.61••• -.6s••• -.56*** .03 .s2••• Note: All statisticswere computedat the group level; N = 65. *Rs;.OS, **Rs;.01, ***Rs;.00 1

<•> Becauseof significant diffegroup rences, all variables were standardized within the MBA student and business student populations.

121 Table B.9. Hierarchical regression analysis forpredictors of performance with conflict-management strategies and task conflict at time 2.

2 Step Predictor Variable t R F dfs AR AF

Compromise Strategy -.05 -.25 Avoiding Strategy .14 1.07

Compromise Strategy -.06 -.25 Avoiding Strategy -.13 1.03 Task Conflict -.04 -.22

Compromise Strategy -.05 -.22 Avoiding Strategy .13 .96 TaskConflict -.04 -.22 CollaborativeX Conflict .14 .53 Compromise X Conflict -.04 -.18 Avoiding X Conflict .12 .80

Note: Step 1 is a test of the linearrelationship between Conflict Strategies andPerf ormance. Step 2 adds Task Conflictto the equation. Step 3tests the interaction between ConflictStra tegies and TaskConflict.

+ p .10; * p .05; ** p .01; *** p .001. s s s s

122 Table B. l 0. Hierarchicalregression analysis for predictors of performancewith confli ct-management strategies and relationship conflictat time2.

Step Predictor Variable t R F dfs

I< CompromiseStrategy -.05 -.25 Avoiding Strategy .14 1.07 fd�illllllJ- •�1:r· CompromiseStrategy -.06 -.27 Avoiding Strategy .14 1.07 Relationship Conflict -.06 -.33

CompromiseStrategy -.01 -.05 Avoiding Strategy .11 .82 Relationship Conflict .00 .05 CollaborativeX Conflict .19 .74 CompromiseX Conflict .01 -.02 Avoiding X Conflict .11 .71 Note: Step 1 is a test ofthe linear relationship between ConflictStrategies andPerf ormance. Step2 adds Relationship Conflictto theequa tion. Step 3 tests the interaction betweenConflict Strategies and Relationship Conflict.

+ p S .10; * p S .05; ** p S .01; *** p S .001.

123 Table B.11. Hierarchicalregression analysis forpredictors of viabilitywith conflict-management strategies andrelationship conflict at time 2.

Step Predictor Variable t R F dfs �F

CompromiseStrategy .08 .65 Avoiding Strategy -.01 -.15 m�IBl![11i �111-lll(llllnilllBlli� Compromise Strategy -.04 .37 Avoiding Strategy -.01 -.11 Relationship Conflict -.48 -5.67*** �1a11. -1 �1Ja1a1iuJE.1�Jf.( ijJ:.. ,1 CompromiseStrategy .01 .06 Avoiding Strategy .01 .14 Relationship Conflict -.52 -5.72*** Collaborative X Conflict -.15 -1.36 CompromiseX Conflict .08 .85 Avoiding X Conflict -.11 -1.63

Note: Step 1 is a test of thelinear relationship between Conflict Strategies andViability. Step2 adds Relationship Conflict to theequation. Step 3 tests theinteraction betweenConflict Strategies and Relationship Conflict.

+ p s .10; * p � .05; ** p s .01; *** p � .001.

124 VITA

DonnaMariaC. Vigil-King was born in Sacramento, Californiaon March 22,

1964, a windyPalm Sunday. Along with two youngerbrothers anda sister, she grew up

in Albuquerque, New Mexico, also a very windy place. Her favoritepastimes, besides

arm-wrestlingthe boys at school for their lunch money, were playing tennis andsewing.

She attendedQueen of Heaven Catholic School, Montgomery Elementary,

ClevelandJunior HighSchool, andDel Norte High School, all in Albuquerque. Because

of a paper snafushe was accepted to the Universityof New Mexico in 1981 as anearly

admissioncandidate but never "officially'' graduatedfrom high school. Not one for

leaving things unfinished, she received her G.E.D. in Juneof 1982.

Afteran internship at Peanut Butterand Jelly Preschool, her focuschanged from

child psychologyto "?". She quit school for a period andput her husbandBill through

school ( since he knew what he wantedto do). When he graduatedfrom college in 1986, they left NewMexico forFlorida. For the next two years, they traveledthe United States

courtesy of the US Navy.

InJanuary 1988, she began commuting4 hours a day to the Universityof

Washington in Seattle. She did aninternship with BattelleNorthwest andin December

1989 received her Bachelors degreein Psychology. Aftertwo-year stint as a computer programmer, she andher husbandmoved to Clinton, Tennessee, and she enrolled in the

Ph.D. programin Industrial andOrganizational Psychology at the Universityof

Tennessee, Knoxville.

125 As a graduate student, she was primarily involved in teamresearch andworked withDr. Eric Sundstrom. Inher free time, she also worked at the TennesseeValley

Authority and started Sierra TECH, a computer consulting business. Her dissertation was supported by the Bonham Dissertation Fellowship, her husband Bill, and Neah-Bay (their collie/border collie).

Donnacurrently lives in Clinton, Tennessee, with her husbandand dog andworks for Covenant HealthCare in Knoxville, Tennessee.

126