teams in organizations: recent research on performance...

0066-4308/96/0201-0307$08.00 307

Annu.Rev. Psychol. 1996.47:307–38Copyright© 1996by AnnualReviewsInc. All rightsreserved

TEAMSIN ORGANIZATIONS: RecentResearch on PerformanceandEffectiveness

Richard A. Guzzoand Marcus W. Dickson

Psychology Department, University of Maryland, CollegePark, Maryland20742

KEY WORDS: group dynamics, organizational change, autonomous workgroups, computerassistedgroups,cockpit resourcemanagement

ABSTRACT

This review examines recent research on groups and teams, giving specialemphasis to research investigating factors that influence the effectivenessofteamsat work in organizations.Several performance-relevant factors arecon-sidered, including group composition, cohesiveness, and motivation, althoughcertain topics(e.g. composition)havebeenmoreactively researchedthanothersin recent yearsandsoare addressedin greater depth. Also actively researchedarecertain typesof teams, including flight crews,computer-supported groups,andvariousformsof autonomouswork groups.Evidenceonbasicprocesses inandtheperformance effectivenessof such groups is reviewed.Also reviewedarefindingsfrom studiesof organizational redesign involving theimplementa-tion of teams. Findingsfromthesestudiesprovidesomeof thestrongestsupportfor thevalueof teamsto organizational effectiveness. Thereview concludesbybriefly considering selected open questionsandemerging directions in groupresearch.

CONTENTSINTRODUCTION..................................................................................................................... 308

Scopeand Objectives........................................................................................................... 308Definitional Struggles.......................................................................................................... 308Framework for theReview.................................................................................................. 309

NEW LOOKSAT LONG-STANDING ISSUESIN GROUPPERFORMANCE .................. 310Cohesiveness........................................................................................................................ 310GroupComposition............................................................................................................. 310Leadershipand Group Performance................................................................................... 313

Motivation and Group Performance................................................................................... 313GroupGoals........................................................................................................................ 314Other Issues......................................................................................................................... 315

KINDS OFGROUPS................................................................................................................ 315Flight Crews:Teamsin the Cockpit.................................................................................... 316Computer-AssistedGroups.................................................................................................. 320DefinedProblem-Solving Groups....................................................................................... 323

TEAMSAND CHANGE IN ORGANIZATIONAL SYSTEMS............................................. 326DISCUSSION............................................................................................................................ 331

OpenQuestions ................................................................................................................... 331NewWaves,NewDirections ............................................................................................... 333Points ofLeverage............................................................................................................... 334

INTRODUCTION

ScopeandObjectives

For morethana decadenow, psychology hasenjoyeda rekindledinterestingroupsandteams.Chaptersin previousAnnualReviewof Psychologyvolumeshaveconsideredgroupresearch(e.g.Levine& Moreland1990)andorganiza-tional behavior(e.g.Wilpert 1995),but this chapteris uniquebecauseof itsspecialfocus on teamperformancein organizationalcontexts,especiallyinwork organizations.

The literature reviewedconsiders,amongother emphases,researchcon-ductedin organizationalsettingswith groupsor teamsthat must meet thedemands ofproducinggoods ordeliveringservices.Although wereviewsomeresearchconductedin other thanorganizationalsettings,we emphasize studiesin whichthedependent variables wereclearly indicative of performanceeffec-tivenessratherthanstudieson intragroupor interpersonalprocessesin groups(e.g.studiesof conformity,opinion change,conflict). We alsoincludestudiesof interventions madeto test the efficacy of techniquesintendedto improveteam effectiveness.Such interventions may be targetedat individual teammembers(e.g.enhancingmemberskills that are importantto teamperform-ance),at teamsasperformingunits (e.g.teamdevelopmentinterventions),orat the organizationsin which teamswork. Thus, researchon larger-scaleorganizationalchangeeffortsof which theimplementationor enhancementofteamsareonepartof anoverallchangestrategyis included.Lastly,weempha-size research in the1990s,thoughwe doreferto earlier works.

DefinitionalStruggles

WORK GROUP/TEAM Whatis awork group?A varietyof definitionshavebeenoffered(Guzzo& Shea1992),butoneweadoptowesits originsto thework ofAlderfer(1977)andHackman(1987).A “work group”ismadeupof individualswho seethemselvesand who are seenby othersas a social entity, who areinterdependentbecauseof thetaskstheyperformasmembersof a group,who

308 GUZZO& DICKSON

are embeddedin oneor morelargersocialsystems (e.g. community, organiza-tion),andwhoperformtasksthataffectothers(suchascustomersorcoworkers).

“Team” haslargelyreplaced“group” in theargotof organizationalpsychol-ogy. Is this a mere matter of wording or are there substantivedifferencesbetweengroupsand teams?For many,“team” connotesmore than “group.”Katzenbach& Smith (1993), for example,assertthat groupsbecometeamswhen they develop a senseof sharedcommitment and strive for synergyamongmembers.Thedefinition of work groupspresentedabove,we believe,accommodatesthe usesof the many labelsfor teamsand groups,includingempoweredteams,autonomouswork groups,semi-autonomouswork groups,self-managing teams, self-determiningteams,self-designingteams,crews,cross-functionalteams,quality circles,project teams,task forces,emergencyresponseteams,andcommittees—alist that represents,but doesnot exhaust,availablelabels.Consequently, we usethe labels“team” and “group” inter-changeably inthis review,recognizingthattheremaybedegreesof difference,ratherthanfundamentaldivergences,in themeaningsimplied by theseterms.We usethe termsinterchangeablyas a convenience.The word “group” pre-dominatesin the researchliterature—intergrouprelations,group incentives,groupdynamics—andthoughit uses“group” asits root word, we believetheliteraturehasgreatrelevancefor understandingvirtually all formsof teamsinorganizations,too.

EFFECTIVENESS Thereis no singular,uniform measureof performanceeffec-tivenessfor groups.Weprefertodefineit broadly,ashaveHackman(1987)andSundstrometal (1990).Accordingly,effectivenessin groupsis indicatedby (a)group-producedoutputs(quantityor quality,speed,customersatisfaction, andsoon),(b) theconsequencesagrouphasfor itsmembers,or (c) theenhancementof ateam’s capabilitytoperformeffectivelyin thefuture.Researchthatassessesone or moreof these threeaspects ofeffectiveness isof primary interest in thisreview.

Frameworkfor theReview

We begin with recent researchon severallong-standing issuesrelevanttowork-group effectiveness,including teamcohesiveness, teamcompositionandperformance,leadership,motivation,andgroupgoals.Theyaregenericissuesin the sensethat they pertainto almostall teamsdoing almostall kinds ofwork. Althoughnot theonly performance-relevantresearchtopics,theyaretheones mostactivelyinvestigatedin recent years.

We thenconsiderresearchon theperformanceof differentkindsof groups,includingcockpitcrewsandelectronicallymediatedgroups,aswell asgroupscreatedto solveproblems(quality circles,taskforces)andautonomousworkgroups.The next sectionexplicitly addressesteamsand the organizational

TEAM EFFECTIVENESSIN ORGANIZATIONS 309

systemsin which they are embeddedand focuseson the interconnectionsbetween teamand organization.

The final sectionoffers selectedconclusionsandflags openquestionsandnewdirectionsfor future research.Thesectionconcludeswith a brief discus-sionof pointsof leverage for effecting change inteams.

NEW LOOKSAT LONG-STANDING ISSUESIN GROUPPERFORMANCE

Cohesiveness

Reviewsof cohesivenessresearchhaveappeared in recent years (e.g.Evans &Dion 1991, Guzzo & Shea1992). The former review found a substantialpositiveassociationbetweencohesionandperformancewhile thelatteroffereda morequalified conclusion.Smith et al (1994) reporta positive correlationbetweena cohesiveness-like measureof top managementteamsin smallhigh-technologyfirms and firm financial performance.Zaccarroet al (1995) re-portedthat highly task-cohesivemili tary teamsunderhigh temporalurgencyperformedaswell on a decisiontaskasdid eitherhigh task-cohesiveor lowtask-cohesiveteamsunderlow temporalurgency,suggesting that taskcohe-sion can improve team decisionmaking under time pressure.The topic ofcohesivenessis still very much an unsettled concernin the literature. It iscertainlyrelatedto issuesof familiarity, which arediscussedat otherpointsinthe chapter.

Group Composition

Groupcomposition refersto thenatureandattributesof groupmembers,anditis one of the most frequently studiedgroup designvariables.Most of theempirical researchon composition andwork-group performance inrecentyearshas investigated variablesassociatedwith teameffectivenesswithoutinterveningor experimenting to affect thosevariables.The typical modelofstudy has beento assessthe performanceof existing groupsor teamsinorganizationsover time andto relatethatperformanceto measuredaspectsofgroup composition.

Other studiesinvestigatedgroup composition as one of severalpossibledesignvariablesfor groups.Groupdesignrefersto issuesof staffing(who is inthe group, what the group size shouldbe), specifying the group’s task andmembers’ roles, and creatingorganizational support systems(e.g. trainingopportunities) for groups.Studiesconductedwith teamsin organizationalsettingsare of particularinteresthere.

Onestudythatrelatedteameffectivenessto composition andotherpotentialdesignvariableswasreportedby Campion et al (1993).Theystudied80 work

310 GUZZO& DICKSON

groupsin a financial servicesfirm andfound broadevidenceof relationshipsbetweeneffectivenessand19 designvariablesclusteredinto five categories:teamjob design(e.g.amountof self-managementin the team),interdepend-ence among team members, composition (especially the heterogeneityofmembers),intragroupprocesses,andcontextualfactors(e.g.managerialsup-port). Campion et al found teamsizeto be positively relatedto effectivenessand found heterogeneityof members’ backgroundand expertiseto be unre-latedor negativelyrelatedto effectiveness,dependingon thespecificcriterionmeasure.

Another study examining some of the same issues was reported byMagjuka& Baldwin (1991).Herethe focuswason factorsthat contributetothesuccessfulimplementationof team-basedemployee-involvementprogramsand the longer-term effective performance of teams in such programs.Throughteamsemployeeshavevoice in organizationalaffairs,gainaccesstoinformation and addressproblemspreviously reservedfor management,andtake on new and varied responsibilities. On the basisof resultsfrom theirnationalsurvey,Magjuka& Baldwin identified factorsthoughtto contributetothe effectivenesswith which employeeinvolvementteamsaredesignedandimplemented.They thenobtainedadditionaldataandexaminedrelationshipsbetweenthesefactorsand effectivenessfor 72 teamsin two manufacturingfirms. They found that largerteamsize,greaterwithin-teamheterogeneity(intermsof thekindsof jobsteammembersheld),andgreateraccessto informa-tion werepositively associatedwith teameffectiveness.The implicationsofthesefindings for designing and implementingemployeeinvolvementteamsarestraightforward.Otherfactorssuchashoursspentin meetings andmem-bers’ wages didnot relate to effectiveness.

HETEROGENEITY AND PERFORMANCE Theextentto which teameffectivenessis affectedby the heterogeneityamongmembersis a complicatedmatter.Magjuka& Baldwin (1991)andCampionet al (1993),asnotedabove,offerseeminglycontradictoryfindings. Jackson et al(1995), in their paperon diver-sity in organizations,reviewedand summarizedempirical evidencefrom anumberof relateddisciplinesaboutthelink betweendiversity(that is, within-groupheterogeneity)andteameffectiveness.Their readingof the literatureisthatheterogeneityis positively relatedto thecreativityandthedecision-makingeffectivenessof teams.Notethatheterogeneityisbroadlydefinedhereandrefersto the mix of personalities,gender,attitudes,and backgroundor experiencefactors.For example,Bantel& Jackson(1989) foundthatorganizationalinno-vationsin thebankingindustrywerepositively associatedwith heterogeneityof functionalexpertiseamongmembersof thetop managementteamsof firmsin thatindustry. Watson etal (1993)reportedthat,overtime(15weeks),initialperformancedifferencesbetweennewly formedculturally homogeneousand


culturallydiversegroups disappearedand eventually“crossed-over,” such thatculturally heterogeneous groups that initially performedpoorly relative tohomogeneousgroups later performedbetter than homogeneous groupsonselectedaspectsof taskperformance(namely,generatingalternativesolutionsandapplyinga rangeof perspectivesin analyzingbusinesscases).Overall,theCampionet al (1993) finding of a nil or negativeassociationbetweentheheterogeneityof groupmembers’ backgroundsandteameffectiveness appearsto be morethe exceptionthanthe rule (Jacksonet al 1991),thoughevidencesupporting thevalueof memberheterogeneityfor teamperformanceis clearestin the domainsof creativeand intellective tasks.The processes(cognitive,social) throughwhich heterogenousgroupcompositionshavetheir effect onteamperformanceare far from fully specified,thoughJacksonet al (1995)explore possible mediatingprocesses.

Heterogeneityof membersalsoappearsto haveother,performance-relatedconsequences.Jacksonet al (1991) reportedthat heterogeneityamongmem-bers of top managementteamsin bank holding companieswas positivelyrelatedto turnoverin thoseteams.Wiersema& Bird (1993)found similar, ifstronger,resultsin a sampleof Japanesefirms. Turnoveris usuallythoughtofasdysfunctionalfor teameffectiveness,thoughit is possiblethat the conse-quencesof losing and replacingmemberscould work to the advantageofteamsin somecircumstances.

FAMILIARIT Y AND PERFORMANCE Anotheraspectof groupcomposition thathasrecentlybeenstudiedfor its relationship to teamperformanceis that offamiliarity amongmembers.Goodman& Leyden(1991)examined,over thecourseof 15 months, theproductivity (in tonspershift) of coal-mining crewswhodifferedin theextenttowhichmemberswerefamiliarwith eachother,theirjobs, and their mining environment. Results indicatedthat lower levels offamiliarity were associatedwith lower levels of productivity. Watson et al(1991)studiedgroupswho spentmorethan30 hoursin decision-making tasksand found that group decision-making effectiveness(relative to individualdecision-making effectiveness)roseovertime,a finding theyattributeat leastin part to the effectsof increasedfamiliarity amongmembers.Dubnicki &Limburg (1991)foundthat olderhealth-careteamstendto bemoreeffectiveincertainways,thoughnewerteamsexpressmorevitality. Thus,someevidenceindicatesthatteamscomposedof individualswhoarefamiliar with oneanothercarryout theirwork with greatereffectiveness than teams composed ofstrang-ers.However,oneshouldbearin mind thatsomeolderevidenceindicatesthattheremaybeapoint,perhapstwoor threeyearsafteragroupis formed,atwhichgrouplongevity andmemberfamiliarity becomedetrimentsto groupperform-ance(Katz 1982). In the later sectionon cockpit crewswe provide furtherdiscussionof team memberfamiliarity.

312 GUZZO& DICKSON

Leadership and GroupPerformance

Theeffectson groupperformanceof leaders’ expectationsof groupperform-ancewerestudiedin a field experimentby Eden(1990a).Thepurposeof theinterventionwasto raise,throughinformationprovidedby an“expert,” groupleaders’ expectationsof their group’s performancein a training setting.Thegroupswereplatoonsin the Israeli DefenseForcesin training that lasted11weeks.Platoonstrainingunderleaderswho heldhigh expectationsperformedbetteron physicalandcognitive testsat theendof training thandid compari-sonplatoons. This researchextendsprior work on theeffectsof expectationson performance(Eden 1990b) andindicatesthat suchexpectancy effectsoccurin theabsence of any loweredexpectationsfor comparisongroups.

Jacobs& Singell (1993) offer a different perspectiveon how individualleaderscanaffect teamperformance.They examinedthe effectsof managers(after controlling for other variables)on the won-lost recordof professionalbaseballteamsovertwo decades and foundit was possibleto identify superiormanagers.Superiormanagerswere effective through at least two possibleprocesses:by exercisingexcellenttacticalskills or by improving theindividualperformances of teammembers.

George& Bettenhausen(1990)studiedgroupsof salesassociatesreportingto a store managerand found that the favorability of leaders’ moodswasinverselyrelatedto employeeturnover.Another study in businessorganiza-tionsexaminedtheposition-basedpowerdominanceof firms’ chief executiveofficers (CEOs) and their top-managementteam size as predictorsof firmperformance(Haleblian & Finkelstein 1993). The study found that firms’performancewasworsein turbulentenvironments whenthe CEO wasdomi-nant andbetter whentop-managementteamsize wasgreater.

Motivation and GroupPerformance

In recentyearsmotivation in groupshasreceivedmoretheoreticalratherthanempiricalattention. Muchof thisattention is devoted tounderstanding motiva-tion at a collective (group, team) level rather than to strictly confining themotivation constructto an individual level of analysis.For example,Shamir(1990)analyzedthreedifferentformsof collectivistic work motivation:calcu-lation (rewardsor sanctionsare anticipatedto follow from group perform-ance), identification (one’s self-conceptis influencedby membershipin agroup),andinternalization(acceptanceof groupbeliefsandnormsasa basisfor motivatedbehavior).Each orientation is consideredviable in differentcircumstances.Guzzoet al (1993) introduced the conceptof group potencyand definedit as the group’s collective belief that it can be effective.Theydifferentiatedthe constructfrom otherrelatedconstructs(e.g.collectiveeffi-cacy)andreviewedevidencethat the strengthof this motivationalbelief sig-


nificantly predictedgroup effectivenessin customerserviceand other do-mains.Guzzoet al (1993) maintainedan interestin motivation at the grouplevel ofanalysis,notat theindividual level ofanalysis.

Individual motivation within groupsalsohasreceivedattention, especiallyas individual motivation is relatedto group-levelfactors.Earley(1994)pro-videdempiricalevidenceon therole of individualism-collectivism (a culture-basedindividual difference)in shapingthe impactof motivational (self-effi-cacy) training for individuals. Group-focusedtraining was found to haveastronger impact on collectivist individuals, and self- focused training wasfoundto havea greaterimpacton individualists.For Earley,a centralresearchquestionwashow individual motivation is affectedby the matchof motiva-tional trainingto theindividual valuesof trainees.Sheppard(1993)offeredaninterpretationof individual task-performancemotivation in groupsthat drewheavily on expectancytheory (e.g. Vroom 1964), reinterpretingwithin theexpectancytheoryframeworkevidenceon individual motivational deficits inthe formof socialloafingand free-riding ingroups.

Group Goals

Relatedto issuesof groupmotivation areissuesof groupgoalsandgoal-set-ting. Goals for group performancecan take many forms: quantity, speed,accuracy,serviceto others,andsoon (seeBrawleyet al 1992for anexplora-tion of the typesof goalssetby sportsteams).And theevidenceis clearthat,comparedwith the absenceof goals (or the presenceof ill-defined goals),specific,difficult goalsfor groupsraisegroupperformanceon thosedimen-sionsreflecting the contentof the goal (Weldon& Weingart1993).That is,goalsfor quantity tendto raisequantity,goalsfor speedtendto raisespeed,and soon.

Thereareoccasionalreportsof failuresof groupgoalsto induceperform-anceeffects(seeFandtet al 1990 for an example).Despitethe exceptions,theredoesappearto bea strongevidentiarybasisfor theperformanceeffectsof goals. Inlight of this, researchhasbeenredirectedtowardunderstanding theprocessesthroughwhich goalshavetheir effects.Weingart (1992),for exam-ple, examinedin a laboratoryexperimentmembereffort and planning,twopossiblemediatorsof goaleffects,andfoundevidenceindicating thatmembereffort mediatedthe impactof goal difficulty on performance.The quality oftheplanning processalsoaffectedgroupperformancein theexpecteddirectionbut was not observedto be a result of goal levels.Weldonet al (1991)andWeingart & Weldon (1991) also provide evidencethat group goals raisemembereffort, but only in the former study did that effort translateintoincreasedgroupperformance.Otherpossiblemediatorsof theeffectsof groupgoalsincludethedegreeof cooperationandcommunication theystimulateingroups(Weldon &Weingart 1993;seealsoLee1989,Locke& Latham1990).

314 GUZZO& DICKSON

Goalsfor group performanceoften coexistwith goalsfor individual per-formance.Whengroupandindividual goalsconflict, dysfunctions canresult.However,it is not necessarilythe casethat evenwhengroupand individualgoals are compatible the presenceof both resultsin levels of performancehigherthanwheneithergoal typeexistsalone.Specifically,Mitchell & Silver(1990)found that thepresenceof both individual andgroupgoalsresultedinperformanceno greaterthanthatattainedin thepresence of group goalsalone.Self-efficacyhasalsobeenexploredin this context,with Lee(1989)showingthatteamgoal-setting mediated therelationship betweenteam-memberself-ef-ficacy and winningpercentage amongseveral female fieldhockey teams.

OtherIssues

Other issuesof long-standing interestbecauseof their relationshipto groupperformanceeffectivenessinclude feedbackand communication in groups.For example,in a studyof a collegiatevolleyball team,de ArmasParedes&Riera-Milian (1987) found won-lost recordsto be relatedto the quality ofintrateamcommunication.Theperformanceeffectsof feedbackwereinvesti-gatedin a studyof railway work crewsby Pearson(1991),who found smallbut statistically significant increasesin productivity over time as a conse-quenceof receiving performancefeedback.The effect of task-performancefeedbackalsowasinvestigatedby McLeodet al (1992).However,theyfoundno significant changein task performanceeffectivenessattributableto suchgoal-referencedfeedback.They alsoinvestigatedthe effectsof feedbackthatconcernedinterpersonalprocessesin groupsand did detecta changein thedominancebehaviorof individuals attributable to it.

KINDS OF GROUPS

The precedingsectionreviewedrecentresearchon long-standingissuesofrelevanceto groupperformance.Issuessuchascomposition, motivation, andleadershipare of near-universalimportanceto groups.They are relevanttomany typesof teamsin many kinds of settings.In this sectionwe considerrecent researchon particular typesof groups.

Many classifications of groupsinto typeshavebeenoffered.Hackman’s(1990)book,for example,organizesits reportsof groupsinto categoriessuchasservice(e.g.delivery) andperforming(e.g.symphonic) teams.In this sec-tion we,too,specifydifferentkindsof groupson thebasis ofthework theydo.We do not offer thefollowing categoriesasa typology thatwe expectto havevalueoutsideof the confinesof this review. Instead,the categorizationsde-fined belowarea matterof conveniencefor organizingrecentresearchlitera-ture.


Flight Crews:Teams in theCockpit

“The crewconcept”in airlines has hadmany names overtheyears.Thephrase“Cockpit ResourceManagement”initially tookhold.More recently, thisfocushascometo be known as “Crew ResourceManagement”(CRM) owing, inpart, to the recognitionof the importanceof including personsnot actuallyinthecockpit (e.g.controllers,flight attendants,etc)aspartof theteam(Lauber1993).

CRM has been defined as“using all available resources—information,equipment, and people—to achieve safe and eff icient fl ight operations”(Lauber1984).The practicalimportanceof sucha programis shownin thefact that over70% ofall severeaircraftaccidentsbetween1959and1989wereat leastpartiallyattributable to flightcrewbehavior.

In general,CRM trainingincludes“not only optimizingtheperson-machineinterfaceandtheacquisition of timely, appropriateinformation, but alsointer-personalactivitiesincluding leadership,effectiveteamformationandmainte-nance,problem solving anddecision making, andmaintainingsituationaware-ness.…Itrepresentsanewfocusoncrew-level(asopposedto individual-level)aspectsof trainingandoperations”(Helmreich& Foushee1993,p. 4). Helm-reich& Wilh elm (1991)noted that CRMtraining is generally wellreceivedbytraineesandleadsto positivechangesin crew members’ attitudesaboutbothcrew coordinationandpersonalcapabilities(or self-efficacy).However,theyalsoacknowledgethat in a smallpercentageof traineesthereis a “boomerangeffect” in which attitudesbecomeless positive.

Related toCRM trainingis Line-OrientedFlight Training(LOFT),which isa broadcategoryencompassingflight simulationsconductedfor severalpur-poses (e.g.to qualifyasapilot, for training). Butler (1993)asserted that LOFTis most importantasa training methodology to reinforceCRM conceptsandtraining.This typeof LOFT is calledCRM LOFT, andit is ongoing, system-atic flight simulation of realistic problemsituationsthat requirethe type ofdecision-making skills andcrewcommunicationthataretaughtin CRM train-ing. Wieneret al (1993) provide an excellentreview of literatureon CRMtrainingand LOFT.

CRM AND CREW COMMUNICATION Communicationis oneof themajorareascoveredin CRM training (Orlady & Foushee1987). In the contextof CRMtraining, communication includes such things as “polite assertivenessandparticipation,activelistening,andfeedback”(Orlady& Foushee1987,p.199).Thougheffectivecommunication is almostuniversallyrecognizedascrucialtoeffective flight crew performance,and CRM training is generallyseenasimprovingcommunicationskills of flight crewmembers,thereis little experi-mentalor quasi-experimentalresearchon theeffectivenessof CRM’s commu-

316 GUZZO& DICKSON

nicationtrainingfor improving outcomes.Instead,the majorityof the researchexaminesthe effects of CRMtrainingon process variables.

Effective crew coordination is in large part a function of effective crewcommunication, and so we note researchby Stout et al (1994), thoughnotquitea CRM-basedstudy.Theirpreliminary investigationsuseda low-fidelityflight simulator,andtheyexaminedthe interactionsamongtwo-personteamsof undergraduatevolunteers.They found that,whenteammembersmustactinterdependentlyto performeffectively,increasedlevelsof suchteamprocessand communication behaviorsas providing information before it is needed,planning,askingfor input, andsteppingin to help otherswereall relatedtoincreased effectiveness.Urban etal (1995) had similar results in anothernon-CRM laboratorystudy in which they examinedthe impactof workloadand teamstructure oneffectiveness.

CRM AND DECISION MA KING Diehl (1991) suggested that 50% of all acci-dent-relatederrorsareerrorsof decision.Thus,thequestionof whetherCRMcan enhance the qualityof decisionmakingin thecockpitis an important one.

Flight crewsarein somewayslike manyothertypesof groupsthat makedecisions.Powerdynamicsarepresent,andtraditionalgroupdecision-makingpitfalls (e.g.groupthink, risky shift) mustbeavoided.Flight crewsaresimilarto othergroupsin that they determinewhat the situation is, assessavailableoptions,and choose amongthem.

In other ways,though, decision making in the cockpit is unlike othergroup-decisionsituations. One significantdifference is that crewdecisionmakingis hierarchicallymanageddecisionmaking:Eachmemberof thecrewcontributeshis or her knowledgeand opinions,and the captainis the finaldecision-maker.Finally, there is a great variety of expertiseavailablein aflight crew,makingflight crewsperhapsmoreheterogeneousthanmanyothertypesof decision-making groups(Orasanu 1993).

CONTEXTUAL VARIAB LES Thereareseveralcontextualvariablesthat play arole in airline crewperformanceandprocess.Oneof themostsignificantis thelimiteddurationof flight crews’ existenceasa unit. In thecommercialairlineindustry,a givenflight crewwill probablyonly work togetherfor atmostfourdays,andsometimeswill betogetherfor only partof oneday.Indeed,commer-cial airline flight crewsperhapsmostclosely resembleproject teamsor taskforcesin thattheyarecomposedof personswith expertisein aspecificarea(e.g.navigator,captain)andwork togetherfor a limi tedperiodof time,afterwhichmembersarereassignedto other flightcrews.

Becauseof this,CRM trainingandLOFT areconductedin thecontextof ateam (allof themembers of aCRM or LOFT flight crewaretrainees).Further,the training is not done with the intention of strengtheningthat particular


team,but ratherwith the goal of making the individuals more effective inwhatever team/flight crew they findthemselves.

Crewslearnto developrelationships quickly (Bowerset al 1993a,Fousheeetal 1986).Thisprocesscanbefacilitatedby thestandardpreflightbriefing.Inthismeeting, thecaptainlaysouthisor herexpectationsfor thecrewandstatesthe goalsof theflight (Ginnett1993).

Finally, and most significantly, Fousheeet al (1986) found that newlyformedcrewscommunicatelesseffectivelyandaremorelikely to haveacci-dentsthanarecrewsthathavebeenintact for at leasta shorttime. This is theprimary reasonthatHackman(1993)recommendedthat thesystemof sched-uling flight crews be modified, though he recognizedthat there would bestrongresistanceto this ideaby flight crew personnel.Note that this mirrorsthestudiescitedearliersuggesting thatteamscomposedof individualswhoarefamiliar with eachotherwill in generalbemoreeffectivethanteamscomposedof peoplewho do not know eachother at all, as is often the casein newlyformed cockpit crews. Indeed,the United StatesArmy embracedthis viewwhentheymandated“battle-rostering” of crews(assigningaviationcrewswhowork togetherfor extendedperiodsof time). However, recentresearchbyLeedom& Simon(1995)suggested that battle-rosteringfor the long-term maylead tooverconfidence—anderrors—amongaviators.

Leedom& Simon(1995)alsonotedthat theunderlyingpurposeof battle-rosteringandothertacticsto increaseteammemberfamiliarity is to increasepredictability of behaviorin theteamsetting.Theyexploredtheeffectivenessof standardizedbehavior-basedtraining to improve team coordinationandfunctioningandfound that this approachled to higherlevelsof performancethan didbattle-rostering and that it did sowithout the potentialoverconfidenceeffectsfound with battle-rostering.Thus, the issueof crew structureand fa-miliarity remainsopen.

A secondcontextualissue is the increasinglevel of automation in thecockpit.With new aircraft designsandthe emergenceof the “glasscockpit,”crewsface new issuesof communication, interaction,and decisionmaking.One reasonfor the emergenceof new automation is the attemptby aircraftmanufacturersto reducehumandecisionmaking as much as possible—be-causepeople too often make bad decisions(Billings 1991). Bowers et al(1993b)foundin a simulatortestthat theadditionof automationdecreasedtheperceivedworkload,but this decreasein workloaddid notnecessarily resultinincreasedperformance.In fact, in difficult situationsthenonautomatedcrewsmadebetterdecisionsthantheautomatedcrews.Further,Costley et al (1989)found that therewerelower communication ratesin moreautomatedaircraft,thoughthere was nodecreasein operationalactions.

318 GUZZO& DICKSON

MIL ITARY FLIGHT CREWS Although there are of coursemany similaritiesbetweenmilitary flight crewsandcommercialflight crews,therearealsosomesignificantdifferencesbetweenthe two. Mili tary flight crewsmay be signifi-cantly larger,for example,andtheyarelikely to remaintogetherasa unit formuch longer periods of time than arecommercialflight crews, owing tobattle-rostering(describedin theprecedingsection).Further,issuesof rankofpersonnelmayplay a greaterrole in themilitary flight crews,andthis maybeat odds with the assertivenesstaught in most CRM-type training. Finally,military flights in peacetimearealmostalwaystraining flights of somekind,whereascommercialflights arefor thepurposeof transportationof cargoandpassengers rather thanfor training(Prince &Salas 1993).

Despitethosedifferences,CRM and LOFT-type training programshavebeendevelopedby severalbranchesof the military (often called Air CrewTraining,or ACT) (Prince& Salas1993).TheseACT programshavegener-ally similar resultsto CRM trainingand LOFT, andthe researchfindings fromoneareagenerallymirror thoseof theother.Forexample,thefinding thatthereis a high correlationbetweenCRM-type behaviorsandobjectiveandsubjec-tive measuresof the effectivenessof aircrews(Povenmireet al 1989)couldeasilyhavecomefrom eitherthecommercialor themili tary air crewresearchprograms.

Further,Prince& Salas(1993) note severalsimilarities betweenmili taryandcommercialresearchinto theoriginsof flight difficulties. Theseincludedproblemswith theexchangeof information in thecockpit,thedistribution andlevel ofpriority of tasks,and relationships within the crew.

It is importantto notethatCRM- andLOFT-typetraininghasnot yet fullytakenroot in the military’s flying culture,andthattheprogramsthathavebeendevelopedvary from oneservicebranchto anotherandfrom onecommandtoanother.This lack of consistencyacrosscommandsand servicesmay makefull-scaleadoption andacceptanceof suchprogramsmoredifficult to achievein themilitary than inthecommercialairlines.

OVERALL EFFECTIVENESSOF CRM TRAINING AND LOFT As notedabove,thereis agreatdealof researchontheeffectivenessof CRM trainingandLOFT, andthisbodyof work is exploredin muchgreaterdetailin Wieneretal (1993)thancan be covered here.

In summary, however,comparedwith no trainingof crewsin CRM, train-ing in CRM resultsin more crewsbeing ratedby crew evaluatorsas aboveaverage and fewer beingrated as below average(Helmreich etal 1990).Further,skills learnedin CRM trainingandLOFT areoftencitedby pilots asplaying akey rolein their handlingof crisis situations(e.g.National Transpor-tationSafety Board1990a,b).


Computer-AssistedGroups

The continuing spread of computerization hasbeen accompaniedby an expan-sionof researchon groupsthatusecomputersin their work. This researchhasin largepart focusedon comparingcomputer-mediatedgroupmeetingswithnon-computer-mediatedmeetingsand,wherework is doneby groups,on ideagenerationand choice making.

An interpretationandannotatedbibliographyof studies,especiallyexperi-ments,on computer-assistedgroups,is providedby Hollingshead& McGrath(1995).They identifiedfifty researchreportsover two decadesyielding about150 findings relevantto taskperformancein computer-mediatedgroups.Al -most all studies were done in laboratories with ad hoc groups. Overall,Hollingshead& McGrathfound that computer-mediatedgroupstendedto becharacterizedby lessinteractionand exchangethan face-to-facegroupsandtendto takelongerin their work. Whethercomputer-mediatedor face-to-facegroupsare superior in task performance(on dimensions other than speed)appearsto dependon thetask.Specifically,computer-mediatedgroupsappearsuperiorat generatingideasbut face-to-facegroupsappearsuperioron prob-lem-solving tasksand tasksrequiring the resolution of conflicts (of prefer-ences,for example).They also suggestthat a large partof the effect ofcomputertechnologyin groupsmaybedueto structuringof thetaskimposedby the useof computertechnologyratherthanotheraspectsof theelectronicmedium.

It is interesting to notethat increasedstructuring of the task—whetherbycomputersor by nontechnologicalmeans—seemsto enhancegroup processes.Consider,for example,the “stepladdertechnique,”in which a coregroupofperhapstwo membersmake a tentativedecision,and with eachsuccessive“step” a new memberis addedand a presentationis madeof the group’scurrent ideas,followedby areneweddiscussionof thepossibilities. Rogelberget al (1992) found that groupsusing this highly structuredprocessproducedhigherquality solutions(to a survivalproblem)thandid groupsusingconven-tional discussionmethods.Further,Hartell (1991)demonstratedthat teamsofundergraduates trained in and utilizing a system of Problem Identifica-tion/Verification dealtwith trouble-shooting tasks moreeffectively thanteamswho were nottrained.

CREATIVIT Y AND BRAINSTORMING Examplesof researchonbrainstormingcanbe foundin thework of Gallupe, Valacich, and colleagues. Dennis & Valacich(1993)reportedthatelectronicallyinteractinggroups(i.e. communicating viacomputers)producedmore ideasduringa brainstorming taskthan didnominalgroups(i.e. thosewhosemembersdid not interact).Gallupeet al (1991,1992,1994) compared face-to-face brainstorming with electronic brainstorminggroupsandfound the latter to be superioror the equalof interactinggroups.

320 GUZZO& DICKSON

Thesestudies suggestthat the electronicbrainstorming mediumreducestheextenttowhichtheproductionof newideasisblockedbysuchthingsaslisteningto othersor waitingfor a turn tospeak.

Sainfortetal (1990)comparedexperimentalgroupsusingacomputer-aideddecision system,a videotapetraining systemin conflict resolution, or nosupportsystem.They found that the computer-aidedgroupsgeneratedmorepotentialsolutionsto theproblemandperceivedthemselvesasmakinggreaterprogressthaneitherof the othergroups.Also, both technologygroups(com-puterandvideotape)weresignificantlymoreeffectivein solving theproblemthanthecontrol group.All of this researchcorrespondsto theconclusionsofHollingshead & McGrath(1995).

DECISIONMAKING McLeod’s (1992)meta-analysisof 13studiesexaminedtherelationshipbetweenvariouselectronicgroup decisionsupportsystemsandgroupprocessoutcomes.It wasshownthat theuse of electronicgroupsupportsystemsin groupdecisionmakingleadsto increasesin decisionquality, levelof focuson task,equalityof participation, andthe lengthof time requiredtoreacha decision.However,use of a group decisionsupportsystemled todecreasesin overall consensusand in satisfactionwith the processand thedecision.

Georgeet al (1992)examinedwhetherthe inclusion of a facilitator amonggroupsmakingdecisionsusingan electronicmeetingsystemwould haveaneffecton the group processor quality ofdecisions made.Theyfoundthattherewereno differencesin eithergroupprocessor outcomes(i.e. decisionquality)betweengroupsthatdeterminedtheir own groupprocessandthosefor whomthe group processwas determinedby a facilitator. Similarly, Archer (1990)found that if thephasesof a decisionprocessin a complexbusinesssituationwereorganizedand rational, therewas no differencein decisionquality be-tween computer-mediated and face-to-facedecisionmaking.

CONTEXTUAL ISSUES Contextual factorsother than the computerprogramsthemselvesalsoplay a role in computer-assistedgroups.Valacichet al (1994)foundsignificantly differentresultsbetweengroupsusingthesamecomputer-mediatedcommunicationsystemwhenall membersof thegroupwerein oneroomasopposedtowhenthemembersweredispersed.In thiscase,thedispersedgroupgeneratedmoreuniquesolutionsandsolutionsof higherquality thandidthe proximate group.

COMMUNICATION PATTERNS Several authors have reached similar conclu-sions about communication patternsin groupswho communicate solely orprimarily by computer.For example,Kiesler & Sproul (1992) found thatcommunication in such groupsis characterizedby greaterdirect advocacy,greaterequality of participation(evenwhen membersareof different status


levels),moreextremeor risky decisions,andmorehostile or extremecommu-nications(e.g.“flaming”) thanin face-to-facegroups.Dubrovskyet al (1991)alsofoundthatsocial-status inequalitieswerelesssalientin groupswho com-municatedandmadedecisionsby electronicmail thanin face-to-facegroups.However,theyalsofoundthatdifferencesin influencebasedon differencesinexpertisewerelesspronouncedin e-mailgroups.Theyreferto thesephenomenaas “the equalizationeffect.”

In some computer-mediated decision systems, communication amongmembersis anonymous. Jessupet al (1990) reportedthree experimentsinwhich theyshowedthatwhentherewasanonymity in thegroupdecision-mak-ing process,membersweremorecritical of ideasproposed,moreprobing intheir questioning,and morelikely to generate questionsandideas.

GROUP PROCESSES Sambamurthy et al (1993) found that experimentalgroups usinga computerizedgroupdecisionsupportsystemto makebudgetallocationdecisionshadbetterorganizeddecisionprocessesthandid groupsusing a paper-and-pencilversionof the decisionsupportsystemand than acontrolgroupto whichnodecisionsupportsystemwasprovided.However,thecomputerizedsystemalsoappearedtoreducethethoroughnessof thediscussionandledto alessintenselycritical decisionprocess.Likewise,Pooleetal (1993)foundthatuseof agroup-decisionsupportsystemimprovedtheorganizationofsubjects’ decision-makingprocessbutmayhaveledto lessthoroughandcriticaldiscussion.Keyset al (1988)usedundergraduatesin a studyof theeffectsofuseof a decision-support systemin a businessstrategygame,andfound thatstudentsin thecomputerconditiondid moreandbetterplanningthanthosein acontrolcondition. Aiken & Riggs(1993)examinedtheapplicability of agroupdecision-supportsystem,in which communication among group members wasalmostentirelyelectronic,to thequestionof groupcreativity.Theyfoundthatgroupsusingthegroupdecision-supportsystemweremoreproductiveandmoresatisfiedwith the processbecauseof suchthings as increasedparticipation,synergy,and enhanced structure.

SHORTFALLS OF COMPUTER-MEDIATED GROUP WORK Computer-mediatedgroupwork is not alwayssuperiorto face-to-faceinteraction,however.Straus& McGrath (1994) found that the productivity (in termsof quantitybut notquality) of face-to-facegroupson discussiontasksexceededthatof electroni-callymediatedgroupsandthatthisproductivity differencewasgreatestonthosetasksrequiring higher levels of coordination amonggroup members.Lea &Spears(1991)confirmedpreviousresearchthatgroupscommunicatingby wayof computersproducemorepolarizeddecisionsthando face-to-facegroups.Adrianson& Hjelmquist (1991)found lessconformityandopinionchangeingroupsusingcomputer-mediatedcommunicationthanin thoseusingface-to-

322 GUZZO& DICKSON

facecommunicationandfoundthatpersonalitycharacteristicsof groupmem-bers were onlyweakly related tothesecommunication patterns.

OTHER TECHNOLOGIES Computersare,of course,not the only technologicalinnovationusedfor groupcommunicationanddecisionmaking.Moresimplistictechnologysuchasteleconferencinghasalsobeenintroduced.Interestingly, thenegativeinterpersonalinteractionsfound in computer-basedcommunications(e.g.“flaming,” increasedtime to decision)appearto beabsentin teleconfer-encing,which is muchmoresimilar to face-to-facecommunications.Groupsmakingdecisionsviateleconferencingtendto takelesstimethandoface-to-facegroups,andmemberstendto perceivetheleader astakingon fewer leadershiproles (Rawlins1989).

SUMMARY Technologicalsystemsthatmorecloselymimic face-to-faceinter-action (e.g. videophonesand videoconferencing)are becomingmore widelyavailable,andtheseadvanceswill spurnew researchinto their useasgroupdecision-making tools. Simultaneously, useof systemsin which thereis noreal-timecommunicationis alsobecomingmoreandmorecommon(e.g.group-ware,list-servers). Thesecommunicationsystemsprovideample opportunitiesfor research.We believe that technology-basedgroup communication anddecision-making systemswill continueto thrive andthatresearchers willhaveto struggle to keepup with thepaceof programmeradvancesandpractitionerusage.

Defined Problem-SolvingGroups

Somegroupsare createdfor the specific purposeof generatingsolutionstoproblems.Quality circles and taskforces aretwo suchkindsof groups.

QUALITY CIRCLES Qualitycirclesweredevelopedasameanstogenerateideasthat, if implemented,would raisetheproductquality byreducingdefects,errorrates,andsoon.Qualitycircleswereaprecursorin theUnitedStatesto themorerecent“total quality movement”in which manymechanismsof quality (and,moregenerally,productivity) improvementare implemented to fostercontinu-ous improvementsin the quality of productsandof services.Quality circlestypicallyare6–12employeeswhoperformrelatedjobsandwhomeettodiscussproblems—andopportunities—toraisethequality or productivity of their partof anorganization.Theygeneratesolutionsthatmayormaynotbeimplementedby theorganization. The introductionof quality circles usually is accompaniedby training in group process(e.g. in structuredtechniquesfor diagnosingproblemsandbrainstorming) aswell astraining in aspectsof quality manage-ment,such asin workingwith statistical indicatorsof quality.


Although quality circleshavebeena popularform of groupsin organiza-tions, evidencesuggeststhat qualitycircles have relatively little enduringimpact on organizationaleffectiveness(Lawler et al 1992) and researchonthem hasdiminished. Steelet al(1990)studied quality circlesovera14-monthperiod ina United Statesfederalmint andfound no evidencethat they affectedimportantorganizationaloutcomes.Qualitycirclesmaysometimesbesuccess-ful at generatingso-calledbig hits early on (i.e. quality improvements thathavesubstantial economicvalueto a firm) but theevidencedoesnot indicatethatqualitycircles can maintain suchcontributions over time.

TASK FORCES Task forces are another kind of group created to solveproblems. Theyaretemporary,createdwith arelativelywell-boundedmandateto be fulfilled. Taskforceshavea morelimi ted time horizonthando qualitycircles; once the task is accomplished, the task force can disband.May &Schwoerer(1994) reported onthe creation oftask forcesto developandimplementwaysof reducingtheincidenceof cumulative traumadisorders(orCTDs)thatresultfromrepetitious,forcefulmovementsin ameat-packingplant.(Carpaltunnelsyndrome is onesuchdisorder.)Teamsweremadeup of 7–9volunteersrepresentingseveralfunctions(e.g.medical,management)andweretrainedin substantive issuesrelatedto CTDs.Theteamsappearedsuccessfulindecreasingtheincidenceandseverityof CTDs,thoughthenumberof productiondayslost to injurieswasunaffected.The authorsof the reportalsopresentedtheir views on the appropriatestructure,training, and supportof task forcessimilar to thosestudied.

AUTONOMOUSWORK GROUPS Weusethelabel“autonomouswork groups”asa synonymfor “self-managingteams”andfor “empoweredteams.”Theseareteamsof employeeswhotypically performhighlyrelatedorinterdependentjobs,who areidentifiedandidentifiableasa socialunit in anorganization,andwhoaregivensignificantauthority andresponsibility for manyaspectsof theirwork,suchasplanning,scheduling,assigningtaskstomembers,andmakingdecisionswith economicconsequences(usuallyup to a specificlimi tedvalue)(e.g.seeDobbelaere &Goeppinger 1993).

Theconceptof autonomouswork groups hasbeenin theliteraturefor half acentury.However,therewaslittle momentumfor their adoptionin US work-placesuntil thepastdecadeor soasfirms reducedlevelsof management,thusgiving overto lower-levelemployeesresponsibilities in thepastheldby man-agement,andasfirms soughtnew waysof increasingemployeeinvolvementand productivity. Autonomous work groupsare inherentin many recentat-temptsto radicallytransformorganizationalwork systems,atopicdiscussedinthe next sectionon teamsandorganizational change.This sectiondealswithresearchspecificallytargetedat autonomous workgroups.

324 GUZZO& DICKSON

Cohen& Ledford(1994)studieda largesampleof self-managingteamsatdifferent levelsand in varying functions in a service organization.Theseself-managingteamshad beenin existencefor two yearson average.Theyweresystematically matchedagainstcomparabletraditionally managedteams.Further,teamswerescreenedfrom the samplewhenthey did not unambigu-ously fulfill thedefinition of self-management.Criteria of teameffectivenessincludedratingson differentdimensionsof performance(e.g.quality, produc-tivity, safety)obtainedfrom differentsources(teammembersandhigherlev-els of management) aswell as indicatorsof effectivenessfrom companyre-cords,suchascustomercomplaints andmonetarylossesdueto absenteeism.Ratings indicatedthat self-managingteamswere more effective than theircomparisongroups.However, no significant differenceswere observedonmeasuresof effectivenessbasedon companyrecords.Work-relatedattitudes(e.g. satisfaction) were morefavorable amongmembersof self-managingteams.

Cordery et al (1991) reporteda study of autonomous work groupsat agreenfieldsite. A greenfieldsite is a new physical location of work. In thisstudyof mineralprocessingplantsin Australia,work groupsat thenewplantsitewerecomparedwith groupsin existingsites.An importantdifferentiatingfeatureof thenewsitewasthatanorganizationalstructureunlike thoseat anyexistingsiteswasimplemented.Thatorganizationalstructure“centeredon theoperationof autonomous work groupsin the processingarea”(Corderyet al1991,p. 465).Greenfieldteamsin this sitehad decision-making responsibilityfor such things as allocatingwork, attendingto administrativematters,andsettingpriorities,aswell ashavinginfluenceon hiring decisions.Their mem-bersalso acquiredmultiple skills and worked undera pay-for-skills rewardsystem. Traditional (nonautonomous)groups,againstwhichautonomousworkgroups were compared,also existed in parts of the new plant and in theestablishedsite.The primary intervention wasthusa changein the natureofgroupwork, in the competenciesof members(throughmultiskilling), and ingroups’ supportingorganizationalcontext(rewardsystem,authority system,informationavailability). This intervention secondarilyinfluencedindividualinputsthroughits creationof multiskilledgroupmembers.

TheCorderyetal (1991)dataindicatedthatautonomouswork groupswereassociatedwith morefavorableemployeeattitudesthanweretraditionalworkgroups, though this differenceabatedovertime (measurementsweremadeat8and 20 months after the greenfieldstart-up).However, both turnover andabsenteeism werehigher among membersof autonomous work groups incomparisonwith traditional groups.

TheCorderyetal (1991)studywas muchlike an earlier studyby Wall etal(1986)thatcontrastedautonomouswork groupsin greenfieldandestablishedsitesengagedin food production. Theearlierstudyalsofoundhigherturnover


amongemployeesin the greenfieldsite. However,the findings of thesetwostudiescontradictthe reportby Weisman et al (1993),who found thathigherretention(i.e. lower turnover)amongnurseswasassociatedwith self-manage-ment practices.A previousreview of researchby Beekun(1989) concludedthat theuseof autonomous work teamsis associatedwith decreasesin absen-teeism and turnover. Otherresults that differedfrom Corderyetal (1991)werereportedby Wall et al (1986),who found lessevidenceof positive attitudinalconsequencesof autonomous work groupsthandid the latter study.Barker’s(1993) casestudy report noted that membersof self-managingteamshadlower levelsof absenteeismandtardinessbecausethe membersof the teamsenforced attendanceandon-timenormsmuchmorestrictly thanmanagershadenforced those policies prior tothe implementationof teams.

Overall there issubstantial variancein researchfindings regardingtheconsequencesof autonomouswork groupson suchmeasuresasproductivity,turnover,andattitudes.This variancemay indicatethat theeffectsof autono-mouswork groupsarehighly situationally dependent.That is, the effectsofautonomouswork-group practicesmay depend onfactors suchas thenature ofthe work force (e.g. its dominantvalues)and the natureof the organization(e.g.informationandrewardsystems).Smith& Comer(1994)did addresstheproposition thatthesuccessenjoyedby self-organizingteams(self-organizingteamsaresimilar to autonomous work groups)may dependon the situation.Througha laboratoryexperiment,Smith& Comerdemonstratedthatself-man-aginggroupscanbeexpectedto bemoresuccessfulin turbulentenvironments.This study is unique in its attempt toprovide direct answersto complexquestionsaboutthe “fit” of autonomous(andrelatedforms of) work groups.Considerablymore researchwill be required,given the numberof possiblefactorsthatcouldmoderatethe impactof autonomouswork groupsin organi-zations.

TEAMS AND CHANGEIN ORGANIZATIONAL SYSTEMS

Groupsarealmostalwaysembeddedin largersocialsystems(e.g.communi-ties, schools,businessorganizations).Thesesocial systemsthat surroundteamsdefineamajorpartof thecontextin which teamperformanceoccurs.AsLevine& Moreland(1990)havepointedout, toomuchpastresearch ongroupperformanceeffectivenesshas beendevoid of attention to the linkagesbe-tweengroupperformanceandaspectsof the socialsystemsin which groupsarelocated.For theoristssuchasMcGrath(1991),a fundamentalassumptionaboutthenatureof groupsis that theyarepartially nestedwithin, andlooselycoupledto, a surroundingsocial system.“Partially nested”refersto the factthat individualsoften aremembersof more than one group and that groupsmaybepartsof morethanonesocialsystem.“Loosely coupled”refersto the

326 GUZZO& DICKSON

fact that there are few clear, mechanistic-like connectionseither betweengroupsandsurroundingsystems or within groups,a point similar to Guzzo&Shea’s (1992)metaphor of groups beingsystems morelike cloudsthanclocks.Another of McGrath’s (1991) fundamentalassertionsabout the nature ofgroupsis thatin such systemsthey performmultiple tasksconcurrently.

There are severalconsequencesof taking seriously the conceptof theembeddednessof teamsin organizations.Oneis that teamperformanceeffec-tiveness and the factors that bring it about are tied to the nature andeffectivenessof the entire organization.Changesin teameffectivenesscanthus haveconsequencesfor changein the larger system,suchas when im-proved performance by a team or set of teams is thought to yield greaterprofits for a business. Perhapswe usuallythink of team-organizationlinkagesin just this way: that teamperformancecontributesto organizationalperform-ance.

The regularity and strengthof suchlinkagesbetweenthe performanceofcomponents(individuals, teams,departments)and overall organizational ef-fectivenessis explored in Harris (1994). That work mostlyaddressestheapparentparadoxthat investmentsin computertechnologymay bring aboutimprovements in performanceat the componentlevel but do not necessarilytranslateinto larger systemimprovements. It also raiseswidely applicableissuesaboutmeasurement,thenatureof socialsystems,andcross-levelinflu-ences.In light of theseconsiderations, it could be quite wrong to maketheeasyassumption that improvements in teamperformanceyield gainsfor thewhole organization.

Team-organization linkagesalso imply that changesin the larger socialsystemcanbring aboutchangesin the teamssituatedin it. That is, oneneednot directly interveneinto teamsto changetheir performance:Interventionsinto the surrounding organizational systemmay bring aboutimproved(or, ifthe intervention isa poor one,reduced) team performance.

The teams-in-organizational-contextperspectiveis complex. It obscurescause-and-effectrelationsso perceptiblefrom experimentalstudiesof groupsstripped of context. I t impl ies that the effects of interventions made atone level (individual, group, organization) may reside at another level.And it implies that multiple simultaneous influences on and of teams maybe takingplacein thesesocialsystems. Complicatedthoughit is, it is impera-tive to examineresearchevidenceon teamsand changein organizationalsystems.

Researchevidenceon teamsand organizational changetendsto be of auniquecharacter.Understandablytherearefewer controlled,experiment-likemethodsandfar morecasestudiesandsurveys.This is an embodimentof aclassicaltrade-off of rigor for relevancein research.However,thereare bynow quite largenumbersof less-rigorousbut highly relevantresearchreports.


It is likely that weaknessesof researchdesign in someare at least partlycompensatedby strengthsin theresearchdesignsof other reports.

An indication of just how many suchreportsexist is given by Macy &Izumi (1993).Theypresentedtheresultsof ameta-analysisof 131field studies(yielding 506 effect-sizeestimates)of organizationalchangethat appearedover a 30-yearperiod.Interestingly,theyencountered1800studies,only 131of which provided sufficient quantitative data for their meta-analysis.(Ofthese131 studies,88.5%werepublishedin refereedjournals.) We focusfirston their findingswith regardto broadorganizationalchangeandthenaddressthosefindingsmostspecificto teams inorganizations.

In regardto overallorganizational change,Macy& Izumi (1993)foundthatindicatorsof financial performanceshow the greatestimprovementswhenmultiple changesaresimultaneouslymadein aspectsof organizationalstruc-ture, humanresourcemanagementpractices,andtechnology.Macy & Izumireport a +0.37correlation betweenthe number of changesmade (“actionlevers” in their terminology) and indicatorsof financial performance.Othercriteria of change(e.g.employeeattitudes)showedno suchrelationship.Butof the many action levers that can be pulled in large-scaleorganizationalchange efforts, whichspecific oneshave thegreatest impact?

With effect-sizemeasuresof financialperformanceasdependentvariables,theactionleverswith thegreatestimpactincludedthecreationof autonomouswork groupsand teamdevelopmentinterventions. Group-orientedinterven-tionsalsoshowedevidenceof improving behavioralmeasuresof performancesuchas turnoverand absenteeism.Other interventions showingappreciablerelationshipsto financial indicatorsof organizationalperformanceincludedjob redesign,increasedemployeeinvolvement,changes(mostly flattening) oforganizationalhierarchies,and changesin workflow. (Macy & Izumi 1993suggestviewing thesefindings with cautionowing to the sometimes smallnumberof caseson which they arebased.)Employeeattitudesshowedlitt lesystematicimprovement withtheseinterventions.

In summary,accordingto Macy & Izumi (1993): Multi faceted,system-wide organizational interventions show the most reliable positive impacton organizational effectiveness,team-orientedinterventionsareoneof a fewsubsetsof interventionsthathavethemostnotableeffects,andteam-orientedinterventionsaffectbothfinancialandbehavioral measuresof performance.

A nonquantitative, comprehensivereview of researchevidenceon teams,organizationalsystems,andeffectivenesswasprovidedby Applebaum& Batt(1994). Applebaum& Blatt describedalternativeorganizationalsystems inwhich teamsareof greateror lessersignificanceaswell asattemptsto trans-form organizationsto moreteam-basedsocialsystems.Historically, accordingto theseauthors,teamsaresignificantelementsin SwedishsociotechnicalandJapaneselean-productionmodels of work organization. Incontrast,teams

328 GUZZO& DICKSON

havenot beenemphasizedin Germanor traditionalAmericanhumanresourcemodelsof organization.

With existing modelsof work organizationsuchas theseas a backdrop,Applebaum& Blatt (1994)examinedexperimentsin workplaceinnovation inAmericanorganizations.Applebaum& Blatt draw on two lines of evidenceabout the use of innovative work practicesand their impact. One line ofevidenceconsistsof 12 largesurveysreportedbetween1982and1993.Theother consistsof 185casestudies.

With regardto teams,Applebaum& Blatt (1994) relatedthat in recentyearsmanyUS organizationshavebeenexperimenting with team-basedworkarrangements.More specifically, it wasestimatedin 1990 that 47% of largeUS companiesmadeuseof self-directed,autonomous work teamsand thattherewasa stronggrowth trendin the useof suchteamsfrom 1987to 1990(Lawler et al 1992). Quality circles were the most frequently implementedtype of team,estimatedto be presentin 66% of the largestcompaniesin theUnitedStates(Lawler et al 1992).Anotherestimateof thepopularityof teamsin organizations wasprovidedby Gordon(1992).Gordonreportedthat80%oforganizationswith 100 or moreemployeesusedteamsin someway andthat50%of employeesin theseorganizationsaremembersof at leastoneteamatwork.

There are, however,many variationsin team-basedorganizationalprac-tices. In someorganizationsthe introduction or renewedemphasison teamsrepresentsonly a small marginal changeto standardoperatingprocedureswhile in otherstheadoptionof teamsis apartof a large-scaleattemptatradicalorganizationaltransformation.Further,in somebut not all organizationstheimplementation of team-basedwork arrangementsmay be accompaniedbychangesin hiring, compensation,decisionmaking,technology,andotherproc-esses.As Applebaum& Blatt (1994)aptlynoted,in practice“teams”is oneofseveral“commonly abusedterms” (p. 72). Given this variation, the path tounambiguous conclusions abouttheconnectionsbetweenteamsandorganiza-tional effectivenessis oftenquitehardto find. The following conclusionsareofferedcognizantof thecaveatsandqualificationsrequiredby thestateof theresearchevidence.

Applebaum& Blatt (1994), largely on the basisof their review of casestudies,concludedthat thereis clearevidencethat team-basedwork arrange-mentsbring aboutimprovedorganizationalperformance,especiallyin meas-uresof efficiency (e.g. reducedcycle times in production)and quality (e.g.fewer defectsin products).Someresearchreportsrun counterto this conclu-sion (e.g.Robertsonet al 1992).However,Applebaum& Blatt’s (1994)con-clusionsaresupportedby thework of Levine& D’AndreaTyson(1990),whoexaminedthe effects of employeeparticipation on productivity. Levine &D’Andrea Tyson identified threeforms of participation: consultative, repre-


sentative,andsubstantive, the latter form constituting the greatestdegreeofparticipation.Consultative participation, for example,may comethroughthecreationof quality circles,representativeparticipation throughlabor-manage-ment committees, andsubstantive participationthrough autonomous workgroups.Cotton (1993) also largely concurred,identifying autonomous workgroupsandself-determining teamsasstructuresthatprovidefar morepartici-pation than quality circles or various forms of representativeparticipation.Levine & D’Andrea Tyson(1990)reviewedempiricalevidencefrom diversesources(e.g.organizationalpsychology,economics,industrial relations)andconcludedthat “participationusually leadsto small, short-runimprovementsin performance and sometimes leads to significant, long-lasting improve-ments in performance” (p. 203, emphasis in original) and that “ there isusually a positive,often small effect of participation on productivity, some-times a zero or statistically insignificant effect, andalmostnevera negativeeffect” (pp. 203–4).Substantiveparticipation,accordingto Levine & D’An-dreaTyson, is the form most likely to result in significant, long-lasting in-creasesin productivity, and work teamsare the primary meansby whichsubstantive participation is attained.Cotton (1993), too, found self-directedwork teamsto be “an effective way to improve employeeproductivity andattitudes”(p. 199)andfoundlittle evidencethatconsultativeor representativeparticipationhas the sameconsequences.

A national surveyof 727USwork establishmentsconductedin 1991alsoisa sourceof evidenceon theimpactof team-basedorganizationalarrangements(seeSpaeth& O’Rourke1994for a descriptionof thesurveyprocedures).Anestablishment is a location of employment. Small businessenterprisesaremore likely to have a single establishmentwhereaslarge enterpriseshavemany.The relationship betweenperformanceandthe team-basedwork prac-ticeswasanalyzedby Kalleberg& Moody (1994).They found thatorganiza-tionsadoptingsetsof practicesthat includedteamsasanimportantelementoforganizationdesigntendedto excelon severalperformancedimensions (e.g.employeerelations,productquality) thoughnot on thedimensionof customerservice.Note that in this surveyperformancewasassessedby ratings(ratherthan,say,by measuresof output) madeby an establishment’s representative,the samerepresentativewho provided other information about their estab-lishment. Thus, in this survey,the potential existsthat somepart of the ob-served relationships are attributableto a response-responsebias.

In summary,ampleevidenceindicatesthatteam-basedformsof organizingoften bring abouthigherlevelsof organizationaleffectivenessin comparisonwith traditional, bureaucraticforms. This evidence,however,is confoundedbecausemore than one change(e.g. more than just the creationof teams)typically is implementedin studiesof organizational change,andmeasuresofeffectivenessreflectmore than just those contributions uniquely attributableto

330 GUZZO& DICKSON

teams.The question“What makesteamseffective?”is directly addressedbyresearchon groupcomposition, leadership,goal setting,andthe like. In con-trast,researcherson teamsandorganizationalchangeask“To whatextentdoteamsaselements inlargersocialsystemscontribute to systemeffectiveness?”For manygroupresearchersandtheoriststhis is a rathernontraditionalques-tion. And it is a vexing questionfor all, althoughthere is consistent, andsometimes quite powerful evidencethat teamscontribute to organizationaleffectiveness.

DISCUSSION

This reviewhas sampled awide-ranging collectionof researchstudieson teameffectiveness,focusingon work teamsin organizationalsystems.Studiesem-phasizedin the review are thosecentrally concernedwith someaspectofeffectivenessasadependentvariableandwith changesandinterventionsmadeto influencetheeffectivenesswith which teamsperform.Ratherthanrestatingthe findings in summaryform, this final sectionconsidersselectedissuesraisedby the researchreview. We first highlight threeopen issues(out ofmany)in teameffectivenessresearch.Then,newerwavesin teamresearch areidentified and briefly considered,including thosemost directly related toissuesdiscussedin this review. Finally, we discuss“points of leverage”forinterveningto affectteamperformance.Thoughtson futureresearch andtheo-rizing areoffered throughout.

OpenQuestions

What is diversity? How doesit affect team performance?Thesetwo openquestionsabout teamcomposition and effectivenessprovide fertile soil forfurther researchandtheorizing.

DIVERSITY Diversity refersto dissimilarity amongmembersin termsof gen-der,ethnicity,race,personality, culture,andfunctionalexperience,amongotherthings. There is evidencethat team effectivenessis well-servedby diversememberswhenteamsperformcognitive,creativity-demandingtasks.Thisisnotto saythatdiversemembershipmight not payoff in enhancedeffectivenessinothertaskdomains; rather,too little is now known to draw firm conclusions.Also, it isnotknownwhetherall formsof diversitycontributein similarportionsor in similar waysto teamperformanceon intellective tasks.In fact, thereis areal need todevelop theory and dataon the waysin which dissimilarity amongmemberscontributesto taskperformance.Justasresearchon goal andteamperformancehasbegunto emphasizethemediatingprocessesconnectinggoalsand team effectiveness,researchon diversity in teamsshould increasinglyemphasizethe processes thatmediateits effects.


FAMILI ARITY When does familiarity help and hurt team effectiveness?Researchon familiarity amongcoal-miningcrews,cockpit crews,and otherwork groupsshowsa benefitto familiarity. That is, thegreaterthefamiliarityamongmembersof a group, the greatertheir performance.However,otherresearchindicatesthattoo-familiarcockpitcrewsmay,in fact,bemoreinclinedto makeerrors.Perhapsthevalueof familiarity is time-dependent.Thatis,highfamiliarity among members(or high interpositional knowledge, asdiscussedby Cannon-Bowers et al 1995) may have the greatest uti l i ty early in ateam’ s existence, perhapsby fostering therapid appearanceof coordinationand integration of team members’ efforts. High familiarity may have valueat other times, too, such as in times of stressor high demand.However,familiarity mayeventuallybecomealiabili ty asthelackof membershipchange(andthus the lack of any unfamiliar membersbeing introducedinto a team)contributesto stultificationandentropyin teams.Thevenerablework by Katz(1982) suggestedthat communicationwithin andbetweenteamsdeclinesasteamsage, thuscommunication maybe animportantmediator of the effectsoffamiliarity.

TEAM BOUNDARIES Whereareteamboundaries?Theboundariesof teamsareimaginarylinesof demarcationseparatingmemberfrom outsider.Boundariesare essentialto the definition of teams(Sundstromet al 1990) and to thepsychologyof being a memberof the in-group vs the out-group.In manyinstancesteamboundariesarereinforcedby suchthingsasuniformsandtheuseof spaceor turf. However,theboundariesof teamsmayat other times be quitedifficult to discern.“Virtual teams”—teamswhosemembersare connectedthrougha networkof computers—areexamplesof teamswhoseboundariesofinclusionandexclusionmay be quite difficult to establish,especiallyif indi-vidualsmayselectivelyjoin anelectronicconversationfor somebut not all ofthe team’s existence.But problemsof establishing teamboundariesare notlimitedto electronicgroups.Vandermark(1991)andLichtenberget al (1990)suggestedthat therearebenefitsto including as teammemberspersonswhomight traditionally havebeenconsideredontheperiphery.Vandermark(1991)raisedtheissuewith regardtotheinclusionof cabincrewsin thecockpitresourcemanagementtrainingof flight crews;Lichtenberget al (1990)raisedtheissuewith regardto psychiatricaidesandtheir role in teamsof health-careprofes-sionals.Further,viewingteamsasentitiesembeddedin largersystemspopulatedby individualswhoaremembersof morethanoneteamalsocancomplicatetheidentification of teamboundaries.We believethatfutureresearchis neededtoclarify issuesof inclusionandexclusionbyvirtueof teamboundaries(for furtherdiscussion,seeGuzzo1996),how boundariesrelateto effectiveness,andhowthe natureof boundariesmight shapethe effectsof interventionsintendedtoraise teamperformance.

332 GUZZO& DICKSON

NewWaves,NewDirections

We briefly considerthreeareasof researchin which therehavebeenrecentsurgesof interest:electronicallymediatedteams,interventionsfor enhancingteam effectiveness,and teamsin the contextof socialsystems.

ELECTRONICALLY MEDIATED TEAMS Although the first studiesof electronicallymediatedteamsweredonenearlytwo decadesago,thepaceof researchonsuchteamshasaccelerated in recent years. Nodoubt thisisattributabletomany factors,not the leastof which is thedecreasingexpenseof the technologyneededforsuchresearch.And newtechnologies(e.g.videoconferencing,communication,andsupportsoftwarefor groups)continuallycreateopportunities to conductnewresearch.Thereis nodoubtthatelectronicallymediatedteamswill becomeanincreasinglycommon featureof theorganizationallandscape.We thereforesuggestthat researchon electronicallymediatedgroupsbreakfree from thetradition of comparingthosegroupsto face-to-facegroups.Instead,futureresearchshouldaccept such groupson theirown terms.It should focus insteadon contrastingtechnologiesandon teameffectivenessunderdifferentwaysofutilizing availabletechnologies.Froma practicalpoint of view we needmoreresearchon how to maximize teameffectivenesswith newtechnologies.Fromatheoreticalpointof viewweneedbetterinsightsandexplanationsof thedriversof thedynamicsof teamperformanceandeffectivenessundersuchtechnologies.

INTERVENTIONS New waysof interveningto improveteameffectivenessarein theworks.Many of thesearetied to a foundation of researchon teamworkandeffectivenessin military teams.Salasetal (1995)pointedoutthat,althoughtherehavebeenfewdirecttestsof team-traininginterventionsin recentresearchon mili tary teams,knowledgehasprogressedto a point wheresuchtraininginterventionsare now possible,groundedin workableconceptualizationsofcompetenciesandtaskrequirementsin teams.Newwaysof interveningarealsoon the horizon dueto new methodologiesof teamresearchand newtheoreticalmodelsof teamperformance (e.g. see Guzzo &Salas 1995).

TEAMSIN CONTEXT A thirdnotableareaof expandingresearchinterestis teamsin context.Theoft-citedrecognition that,historically, thebulk of psychologicalresearchhasexaminedteamsin theabsenceof consideration of their contextsis giving way tomore frequentstudies ofteams in naturalistic settings, such asorganizations.We expect this shift to be accompaniedby new theoreticalemphasesandinsights,especiallyasthey relateto the influenceof aspectsofthe teams’ environments. In organizations,suchenvironmental factorscouldinclude intraorganizational factorssuchas rewardpracticesand informationsystems,aswell asextraorganizationalfactorssuchasthecustomer demandsand businessenvironments.


Points of Leverage

Threeprimarypointsof leverageexist for interveningto enhanceteameffec-tiveness. Oneis the design ofthe group. Designincludes suchthings asspecificationof membership,of memberrolesandmethodsof their coordina-tion, andof goals.Severalstudieswe havereviewedconcerndesignasa pointof leveragefor raisingteameffectiveness.Diversityof membershipandsizeofgroup, for example,have been found to be related to team effectiveness,althoughtherelationshipsarenotcompletelyconsistentacrossall studiesor allgroup tasks.The effect of goalson group performancehasbeenmore uni-formly found to bepositive, althoughevenherewe foundonestudythatwasanexceptionto thepatternof evidence.Whatwe arecalling “design” is verymuch like whattraditionalmodelsof group performancereferto as“inputs” inthe input-process-outputdescriptionof groupperformance.

The “process” elementin the traditional input-process-output model in-cludesboth socialprocessesin groups(e.g.cohesiveness)andtaskprocesses(e.g.rulesof taskperformance).Groupprocessis thusa secondleverage pointat which interventions can be made to improve team effectiveness.Someevidencein the literaturereviewedfound, for example,that groupcohesive-nesscan contributeto performance,and other studiesfound that structuredtask processes—suchas the stepladdertechniquefor group problem solv-ing—cancontributepositively toperformance.

The traditional input-process-output model would be too confining if itsinterpretationwere restrictedto the idea that inputs (i.e. membercharac-teristics,goals)fully determinegroupprocess.Inputsinfluencegroupprocessbut maynot stronglyconstrainit. Onefactorthatcanstronglyconstraingroupprocessis the technologywith which a groupworks,suchascomputers.Ourreviewof computer-assistedgroupsindeedshowstheir processto bedifferent(e.g.more equalbut lessoverallmemberparticipation) from non-computer-as-sistedgroupsand that thesedifferencesmay or may not result in enhancedeffectiveness, dependingon factors suchas the task.

A third point of leveragefor enhancingteameffectivenessis the context.That is, teamperformancecanberaisedby changingtheconditionsin whichteamsperform.Severallinesof evidencewe havereviewedpoint to thepowerof the contextas a driver of teameffectiveness.Organizationalleaders,forexample,area partof thecontextin which work groupsperform,andleadershavebeenshownto influenceteameffectiveness.Cockpit resourcemanage-mentandits variationsappearto havepositive effectson flight crewsbecausesuchinterventions changethe organizational context (values,culture)in whichcrewsareformedandcarryout their work. Further,large-scaleorganizationalchangeefforts that changethe socialsystemof which teamsarea part havebeen shown to enhanceeffectiveness. Thepoint of leveragewith themost consis-

334 GUZZO& DICKSON

tentresearchsupportfor affectingteamperformanceis thecontext.In fact,it isprobablymostjustifiable to concludethat thegreatestchangesin teameffec-tivenessaremostlikely to be realizedwhenchangesin teams’ organizationalcontextaresupportedby theappropriate teamdesignand process.

Any Annual Reviewchapter, aswell asany arti clecited in an Annual Reviewchapter,may bepurchased fromthe Annual ReviewsPreprints and Reprints service.

1-800-347-8007; 415-259-5017; email: [email protected]

LiteratureCited

Adrianson L, Hjelmquist E. 1991. Group proc-essesin face-to-face and computer medi-ated communication. Behav. Inf. Tech.10(4):281–96

Aiken MW, Riggs M. 1993. Using a groupdecision support system for creativity. J.Creat. Behav. 27(1):28–35

AlderferCP.1977. Group andintergroup rela-tions. In Improving the Quali ty of WorkLife, ed. JR Hackman, JL Suttle, pp.227–96. Pallisades,CA: Goodyear

ApplebaumE, Blatt R. 1994. The NewAmeri-canWorkplace.Ithaca, NY:ILR

Archer NP. 1990. A comparison of computerconferenceswith face-to-facemeetings forsmallgroup businessdecisions.Behav. Inf.Tech.9(4):307–17

Bantel KA, JacksonSE. 1989. Top manage-ment and innovations in banking: Doescomposition of the top teamsmakeadiffer-ence?Strateg. Manage.J. 10:107–24 (Spe-cial issue)

Barker JR. 1993. Tightening the iron cage:concertive control in self-managing teams.Adm.Sci. Q. 38:408–37

Beekun RI. 1989. Assessing the effectivenessof socio-technical interventions:antidoteorfad?Hum.Relat. 47:877–97

Bil lings CE. 1991. Human-centered aircraftautomation: a concept and guidelines.Tech.Memo.103885. Moffett Field, CA:NASA-Ames Res.Cent.

Bowers CA, Braun CC, HolmesBE, MorganBB Jr. 1993a. The development of aircrewcoordination behaviors. In Proc. SeventhInt. Symp. Aviat. Psychol., pp. 573–77. Co-lumbus, OH

Bowers CA, Deaton J, Oser RL, Prince C,Kolb M. 1993b. The impact of automationon crew communication andperformance.In Proc. Seventh Int. Symp. Aviation Psy-chol., pp. 573–77. Columbus, OH

Brawley LR, Carron AV, Widmeyer WN.1992. The nature of group goals in sportsteams:a phenomenological analysis. SportPsychol. 6:323–33

Butler RE. 1993. LOFT: Full-mission simula-

tion asCrewResourceManagement Train-ing. SeeWieneretal 1993, pp. 231–59

Campion MA, MedskerGJ, Higgs AC. 1993.Relations between work group charac-teristics andeffectiveness:implications fordesigning effective work groups. Pers.Psychol. 46:823–50

Cannon-BowersJA, TannenbaumSI, SalasE,Volpe CE. 1995.Defining competenciesand establishing team training require-ments. See Guzzo & Salas 1995, pp.333–80

Cohen SG, Ledford GE Jr. 1994. The effec-tivenessof self-managingteams:a field ex-periment. Hum.Relat. 47:13–43

CorderyJL, Mueller WS,Smith LM. 1991. At-titudinal andbehavioral effectsof autono-mous group working: a longitudinal fieldstudy. Acad. Manage.J. 34:464–76

Costley J, Johnson D, Lawson D. 1989.Acomparison of cockpit communicationB737–B757. In Proc. Fifth Int. Symp.Aviat. Psychol., pp. 413–18. Columbus,OH

Cotton JL. 1993. EmployeeInvolvement. New-bury Park, CA: Sage

deArmasParedesM, Riera-MilianMA. 1987.Analisisdela communicacionenunequipodeportivo y su influencia en los resultadosde este.[Analysis of communication in asports teamand its influenceon perform-ance.]Bol. Psicol. Cuba 10:37–48(Abstr.)

Dennis AR, Valacich JS. 1993. Computerbrainstorms: more heads arebetter thanone.J. Appl. Psychol. 78:531–37

Diehl A. 1991. The effectivenessof trainingprograms forpreventing aircrew“error.” InProc. Sixth Int. Symp. Aviat. Psychol., pp.640–55. Columbus:Ohio State Univ.

DobbelaereAG, Goeppinger KH. 1993. Theright way and the wrong way to set up aself-directed work team. Hum. Resour.Prof. 5:31–35

Dubnicki C, Limburg WJ. 1991. How dohealthcareteams measure up?Healthc.Fo-rum34(5):10–11

Dubrovsky VJ, Kiesler S, Sethna BN. 1991.


The equalization phenomenon: status ef-fects in computer-mediated and face-to-face decision-making groups. Hum.-Com-put. Interact.6(2):119–46

Earley PC. 1994. Self or group? Cultural ef-fects of training on self-efficacy and per-formance.Adm.Sci. Q. 39:89–117

Eden D. 1990a. Pygmalion without interper-sonal contrast effects: whole groups gainfrom raisingmanagerexpectations.J. Appl.Psychol. 75:394–98

Eden D. 1990b. Pygmalion in Management:Productivity as a Self-Fulfill ing Prophecy.Lexington, MA: Lexington Books

Evans CR, Dion KL. 1991. Group cohesionand performance: a meta-analysis. SmallGroupRes.22:175–86

Fandt PM, RichardsonWD, ConnerHM. 1990.Theimpactof goal-setting on teamsimula-tion experience. Simul. Gaming 21(4):411–22

FousheeHC, Lauber JK, Baetge MM, AcombDB. 1986. Crewperformanceasa functionof exposure to high-density, short-haulduty cycles. NASA Tech. Memo. 88322.Moffett Field, CA: NASA-Ames Res.Cent.

Gallupe RB, Bastianutti L, Cooper WH. 1991.Brainstorming electronically. J. Appl. Psy-chol. 76:137–42

Gal l upe RB, Cooper WH, Grise’ M-L ,Bastianutti LM. 1994. Blocking elec-tronic brainstorms. J. Appl. Psychol. 79:77–86

Gal l upe RB, Dennis AR, Cooper WH,Valacich JS, Bastianutti L, NunamakerJ.1992. Electronic brainstorming and groupsize.Acad. Manage.J. 35:350–69

George JF, Dennis AR, NunamakerJF. 1992.An experimental investigation of facilita-tion in an EMS decision room. Group De-cis.Negot. 1(1):57–70

George JM, Bettenhausen K. 1990. Under-standing prosocial behavior, salesperform-ance,and turnover: a group-level analysisin a service context. J. Appl. Psychol. 75:698–709

Ginnett RC. 1993. Crewsasgroups: their for-mation andtheir leadership. SeeWieneretal 1993, pp. 71–98

Goodman PS, Leyden DP. 1991. Familiarityand group productivity. J. Appl. Psychol.76:578–86

Gordon J. 1992. Work teams—How far havetheycome?Training29:59–65

GuzzoRA, SalasE, eds.1995. TeamEffective-ness and Decision Making in Organiza-tions.SanFrancisco:Jossey-Bass

Guzzo RA, SheaGP. 1992. Group perform-anceand intergroup relations in organiza-tions. In Handbook of Industrial and Or-ganizational Psychology, ed. MD Dun-nette, LM Hough, 3:269–313. Palo Alto,CA: Consult. Psychol. Press.2nd ed.

GuzzoRA, Yost PR, Campbell RJ, Shea GP.1993. Potencyin groups:articulatingacon-struct. Br. J. Soc. Psychol. 32(1):87–106

GuzzoRA. 1996. Fundamental considerationsabout workgroups. In In Handbook ofWork Group Psychology, ed. M West.Chichester:Wiley. In press

Hackman JR. 1987. The design of workteams.In Handbook of Organizational Behavior,ed. JW Lorsch, pp. 315–42. EnglewoodCliff s, NJ:Prentice-Hall

HackmanJR,ed.1990. Groups That Work andThoseThat Don’t. SanFrancisco: Jossey-Bass

HackmanJR. 1993. Teams,leaders, and or-ganizations: new directions for crew-ori-entedfl ight training. SeeWieneretal 1993,pp. 47–70

Haleblian J, Finkelstein S. 1993. Top manage-mentteamsize,CEOdominance,andfirmperformance: the moderating rolesof envi-ronmental turbulenceanddiscretion. Acad.Manag. J. 36:844–63

HarrisDH, ed.1994. Organizational Linkages:Understanding the Productivity Paradox.Washington, DC: Natl. Acad.Press

HartelCEJ.1991. Improving team-assisteddi-agnostic decision making: sometrainingpropositions and an empirical test. PhDthesis.Colo. StateUniv., Fort Coll ins

Helmreich RL, FousheeHC. 1993. Why crewresource management? Empirical andtheoretical bases of human factors train-ing in aviation. SeeWieneret al 1993, pp.3–45

Helmreich RL, Wilhelm JA. 1991. Outcomesof crewresourcemanagementtraining. Int.J. Aviat. Psychol. 14:287–300

Helmreich RL, Wilhelm JA, Gregorich SE,Chidester TR. 1990. Preliminary resultsfrom the evaluation of cockpit resourcemanagementtraining: performance ratingsof fli ghtcrews.Aviat. SpaceEnviron. Med.576–79

HollingsheadAB, McGrath JE. 1995. Com-puter-assistedgroups: a critical review ofthe empirical research.SeeGuzzo& Salas1995, pp. 46–78

JacksonSE, Brett JF, SessaVI, Cooper DM,Julin JA, Peyronnin K. 1991. Somediffer-ences make a difference: individual dis-similarity and group heterogeneity as cor-relatesof recruitment, promotion, andturn-over.J. Appl. Psychol. 76:675–89

JacksonSE, May KE, Whitney K. 1995. Un-derstanding the dynamics of diversity indecision-making teams.SeeGuzzo& Salas1995, pp. 204–61

JacobsD, Singell L. 1993. Leadership andor-ganizational performance: isolating linksbetweenmanagers and collective success.Soc. Sci. Res.22:165–89

Jessup LM, Connolly T, Tansik DA. 1990. To-ward a theory of automated group work:

336 GUZZO& DICKSON

the deindividuating effects of anonymity.Small Group Res.21(3):333–48

KallebergAL, Moody JW. 1994. Human re-source management and organizationalperformance.Am.Behav. Sci. 37:948–62

Katz RL. 1982. Theeffects of group longevityon project communication and perform-ance.Adm.Sci.Q. 27:81–104

Katzenbach JR, Smith DK. 1993. The disci-pline of teams.Harv. Bus.Rev.71:111–20

Keys B, Burns O, CaseT, Wells RA. 1988.Decision support package in a businessgame:performanceandattitudinal affects.Simul. Games.19(4):440–52

Kiesler S, Sproul L. 1992. Group decisionmaking and communication technology.Organizational Behav. Hum. Decis.Proc-ess.52(1):96–123

LauberJK. 1984. Resourcemanagementin thecockpit. Air LinePilot 53:20–23

Lauber JK. 1993. Foreword. SeeWieneret al1993, pp. xv–xviii

Lawler EE, Mohrman SA, Ledford G. 1992.EmployeeInvolvement and TQM: Practiceand Results in Fortune 5000 Companies.SanFrancisco:Jossey-Bass

Lea M, SpearsR. 1991. Computer-mediatedcommunication, de-individuation andgroup decision-making. Int. J. Man-Mach.Stud. 34(2):283–301

Lee C. 1989. The relationship betweengoal-setting, self-eff icacy, and female fieldhockey team performance. Int. J. SportPsychol. 20(2):147–61

Leedom DK, Simon R. 1995. Improving teamcoordination: a case for behavior-basedtraining. Mil. Psychol. 7(2):109–22

Levine DI, D’Andrea TysonL. 1990. Partici-pation, productivity, and the firm’s envi-ronment. In Paying For Productivity, ed.AS Blinder, pp. 183–237. Washington,DC: BrookingsInst.

Levine JM, Moreland RL. 1990. Progressinsmall group research.Annu. Rev.Psychol.41:585–634

Lichtenberg PA, Strzepek DM, Zeiss AM.1990. Bringing psychiatric aides into thetreatmentteam:anapplication of theVeter-ansAdministration’s ITTG model. Geron-tol. Geriatri. Educ. 10(4):63–73

Locke EA, Latham GP. 1990.A Theory ofGoal -Setting and Task Per formance.EnglewoodCliff s, NJ:PrenticeHall

Macy BA, Izumi H. 1993. Organizationalchange, design, and work innovation: ameta-analysis of 131 North Americanfieldstudies—1961–1991. In Research in Or-ganizational Changeand Development, ed.W Passmore, R Woodman, 7:235–313.Greenwich,CT: JAI

Magjuka RJ, Baldwin TT. 1991. Team-basedemployeeinvolvementprograms:effectsofdesign and administration. Person. Psy-chol. 44:793–812

May DR, Schwoerer CE. 1994. Employeehealth by design: using employeeinvolve-ment teams in ergonomic job redesign.Person. Psychol. 47:861–76

McGrathJE.1991. Time, interaction, andper-formance:a theory of groups.Small GroupRes.22:147–74

McLeod PL. 1992. An assessmentof the ex-perimental li terature on electronic supportof group work: results of a meta-analysis.Hum.-Comput. Interact. 7(3):257–80

McLeodPL, Liker JK, Lobel SA. 1992. Proc-essfeedback in taskgroups:anapplicationof goal setting. J. Appl. Behav. Sci. 28:15–41

Mitchell TR, Silver WS. 1990. Individual andgroupgoalswhenworkersare interdepend-ent: effectson taskstrategiesandperform-ance.J. Appl. Psychol. 75:185–93

National Transportation SafetyBoard. 1990a.Aircraft Accident Rep.: United Air li nesFl ight 811, Boeing 747-122, N4713U.Honolulu, HI , Feb. 24, 1989. (NTSB/AAR/90/01). Washi ngton, DC: Natl .Transp.Saf.Board

National Transportation SafetyBoard. 1990b.Aircraft Accident Rep.: United Air li nesFlight 232, McDonnell -Douglas DC-10-10. Sioux GatewayAirport, Sioux City, IA,July 19, 1989. (NTSB/AAR/90/06). Wash-ington, DC: Natl. Transp.Saf.Board

Orasanu JM. 1993. Decision-making in thecockpit. See Wieneretal 1993, pp. 137–72

Orlady HW, FousheeHC, eds.1987. Proc. ofthe NASA/MAC Workshop on Cockpit Re-source Manage. (NASA Conf. Publ. 2455).Moffett Field, CA: NASA-Ames Res.Cent.

PearsonCAL. 1991. An assessmentof extrin-sic feedback on participation, role percep-tions, motivation, and job satisfaction in aself-managedsystemfor monitoring groupachievement. Hum.Relat. 44:517–37

Poole MS, HolmesM, Watson R,DeSanctisG.1993. Group decisionsupport systemsandgroup communication: acomparisonof de-cision-making in computer-supported andnon-supported groups. Commun. Res.20(2):176–213

Povenmire HK, RockwayM, BuneckeJL, Pat-ton MW. 1989. Cockpit resource manage-ment skills enhancecombat mission per-formancein B-52 simulator. In Proc. FifthInt. Symp.Aviat. Psychol., pp. 310–25. Co-lumbus,OH

PrinceC, SalasE. 1993. Training andresearchfor teamwork in the military aircrew. SeeWieneretal 1993, pp. 337–66

RawlinsC. 1989. The impactof teleconferenc-ing on the leadership of small decision-making groups. J. Organ. Behav. Manage.10(2):37–52

Robertson D, Rinehart J, Huxley C, and theCAW Research Group onCAMI . 1992.


Team concept and Kaizen: Japanesepro-duction managementin a unionizedCana-dian auto plant. Stud. Polit. Econ. 39:77–107

Rogelberg SG, Barnes-Farrell JL, Lowe CA.1992. The stepladder technique: analterna-tive group structure facilitating effectivegroup decision making. J. Appl. Psychol.77:730–37

Sainfort FC, Gustafson DH, Bosworth K,Hawkins RP. 1990. Decision support sys-tem effectiveness: conceptual frameworkand empirical evaluation. Organ. Behav.Hum.Decis.Process.45(2):232–52

Salas E, Bowers CA, Cannon-Bowers JA.1995. Mi li tary teamresearch:10 yearsofprogress.Mil. Psychol. 7:55–75

Sambamurthy V, Poole MS, Kelly J.1993. Theeffectsof variations in GDSScapabili tieson decision-making processes in groups.Small Group Res.24(4):523–46

Shamir B. 1990. Calculations, values, andidentities:the sources ofcollectivistic workmotivation. Hum.Relat. 43:313–32

Sheppard JA. 1993. Productivity loss in per-formance groups: a motivational analysis.Psychol. Bull . 113:67–81

Smith C, Comer D. 1994. Self-organization insmall groups: a study of group effective-ness within non-equil ibrium conditions.Hum.Relat. 47:553–81

Smith KA, Smith KG, Olian JD, Sims HP,O’Bannon DP, Scully J. 1994. Top man-agement team demography and process:the role of social integration andcommuni-cation. Adm.Sci. Q. 39:412–38

Spaeth JL, O’Rourke DP.1994. Designing andimplementing the national organizationsstudy. Am.Behav. Sci. 37:872–90

Steel RP, Jennings KR, Lindsey JT. 1990.Quality circle problem solving and com-mon cents:evaluation study findings froma United States federalmint. J. Appl. Be-hav. Sci. 26:365–81

Stout RJ, SalasE, CarsonR. 1994. Individualtaskproficiency andteamprocess:What’simportant for teamfunctioning? Mil. Psy-chol. 6(3):177–92

StrausSG, McGrath JE. 1994. Does the me-dium matter? The interaction of task typeandtechnology on group performance andmembersreactions. J. Appl. Psychol. 79:87–97

Sundstrom E, De MeuseKP, Futrell D. 1990.Work teams: applications and effective-ness.Am.Psychol. 45:120–33

Urban JM, BowersCA, Monday SD, MorganBB Jr.1995. Workload, teamstructure,andcommunication in teamperformance. Mil.

Psychol. 7(2):123–39ValacichJS,George JF,NunamakerJF,Vogel

DR. 1994. Physical proximity effects oncomputer-mediated group idea generation.Small GroupRes.25(1):83–104

Vandermark MJ. 1991. Should flight atten-dants be includedin CRM training? A dis-cussion of a major air carrier’s approachtototal crew training. Int. J. Aviat. Psychol.1(1):87–94

Vroom VH. 1964. Work and Motivation. NewYork: Wiley

Wall TD, Kemp NJ, JacksonPR, Clegg CW.1986. Outcomes of autonomous workgroups:a field experiment. Acad. Manage.J. 29:280–304

WatsonWE, KumarK, MichaelsenLK. 1993.Cultural diversity’s impact on interactionprocessand performance: comparing ho-mogeneous anddiversetaskgroups. Acad.Manage.J. 36:590–602

WatsonWE, MichaelsenLK, SharpW. 1991.Member competence, group interaction,andgroup decision making: a longitudinalstudy. J. Appl. Psychol. 76:803–9

Weingart LR. 1992. Impact of group goals,task component complexity, effort, andplanning on group performance.J. Appl.Psychol. 77:682–93

Weingart LR, Weldon E. 1991. Processesthatmediate the relationship between a groupgoal and group member performance.Hum.Perform. 4:33–54

WeismanCS, Gordon DL, CassardSD, Ber-gner M.1993. The effectsof unit self-man-agement on hospital nurses’ work process,work satisfaction, and retention. Med.Care. 31(5):381–93

Weldon E, JehnKM, Pradhan P. 1991. Proc-essesthat mediatethe relationship betweena group goal andimprovedgroup perform-ance.J. Pers. Soc. Psychol. 61:555–69

Weldon E, Weingart LR. 1993. Group goalsand group performance.Br. J. Soc. Psy-chol. 32:307–34

Wiener EL, Kanki BG, Helmreich RL, eds.1993. Cockpit Resource Management.SanFrancisco:Academic

WiersemaMF, Bird A. 1993. Organizationaldemography in Japanesefirms: group het-erogeneity, individual dissimilarity, andtopmanagement team turnover.Acad. Man-age.J. 36:996–1025

Wilpert B. 1995. Organizational behavior.Annu. Rev.Psychol. 46:59–90

Zaccaro SJ, Gualtieri J, Minionis D. 1995.Taskcohesion asa facilitatorof teamdeci-sion making under temporal urgency. Mil.Psychol. 7(2):77–93

338 GUZZO& DICKSON

teams in organizations: recent research on performance...

Documents