搜索
Linkingleaderbehaviorandleadershipconsensustoteamperformance:IntegratingdirectconsensusanddispersionmodelsofgroupcompositionMichaelS.Colea,⁎,ArthurG.Bedeianb,HeikeBruchcaTexasChristianUniversity,M.J.NeeleySchoolofBusiness,DepartmentofManagement,FortWorth,TX76109,UnitedStatesbLouisianaStateUniversity,LA,UnitedStatescUniversityofSt.Gallen,SwitzerlandarticleinfoabstractAvailableonline15March2011Wesoughttoestablishwhethermeanlevelwithin-teamassessmentsofaleader'stransformationalbehaviorandtheextentofperceivedvariability(i.e.,consensus)amongteammembers’ratingsaroundthismeanlevelareseparateyetrelatedindicatorsofleaderquality.Tothisend,usingdatafrom108workteamsinamultinationalfieldsetting,weexploredtherelationshipbetweenmanagers’transformationalleadershipbehaviorandteamperformanceasmoderatedbytheextentofvariabilityamongteammembers’leadershipbehaviorratings.Recognizingthattherelationshipbetweenleaderbehaviorandteamperformanceisindirect,wefurtherexaminedwhetherteamempowermentservedasamediatingmechanismthroughwhichtransformationalleadershipisevoked.Studyresults,basedonordinaryleastsquares(OLS)regressionanalysesandbootstrappedestimates,wereconsistentwiththehypothesizedconceptualschemeofmoderatedmediation,inthatthejointeffectsoftransformationalleadershipbehaviorandconsensusabouttransformationalleadershipwerefoundtohaveanindirecteffectonteamperformancethroughteamempowerment.Thesefindingsbroadenthefocusoftransformationalleadershiptheorybyillustratingthat,withinateamcontext,thetransformationalleadership–performancelinkageismorenuancedthanpreviouslybelieved.©2011ElsevierInc.Allrightsreserved.Keywords:Transformationalandcontingent-rewardleadershipDispersionmodelofleadershipConsensusconstructsTeamperformanceTeamempowermentInanefforttounderstandtheantecedentsofjobperformance,leadershipresearchhastraditionallyemphasizedtherelationshipbetweenmanagersandsubordinatesasindividuals(Day,Gronn,&Salas2004).Withthewidespreadshifttoteam-basedorganizations,however,thereismountingevidencethatamanager'sleadershipbehaviorinsupervisingsubordinatesasagroupinfluencestheperformanceinandofteams(Kozlowski&Ilgen2006;Mathieu,Maynard,Rapp,&Gilson2008).Recognizingthisshiftinemphasis,Chen,Kirkman,Kanfer,Allen,andRosen(2007),amongothers,haveobservedthatafocusongroupmethodsofleadershipraisesaseriesofyetunansweredquestions.Forexample,whatcompromisesmayoccurwhenmanagingbothindividualsandteamsasawhole?Doeffortstoleadandmotivateteamsaffectindividualleader–followerinteractions,andwhatarethelikelyconsequencesforone-on-onerelationshipsbetweenteammembers?Arethebehaviorsnecessaryforbuildingandleadinghighperformingteamsthesameasformotivatingperformanceattheindividuallevel?Suchquestionshavepromptedleadershipresearchers(e.g.,Chen&Kanfer2006)toexploretheinterplaybetweensimultaneouslymanagingteammembersasindividualsandasagroupintoto.Illustrativeoftheseresearchers’effortsarestudiesinvestigatingleadershipbehaviorasagroup-levelphenomenon.Bass,Avolio,Jung,andBerson(2003),forinstance,wereinterestedinexamininghowleaderbehaviorsdirectedatteamsasawholerelatetounitperformance.Indoingso,theyreliedonadirectconsensusmodelofcomposition(Chan1998).Consistentwiththedirectconsensusmodel,individualandteameffectsarepartitionedusinggroupmeanscorestomeasuretheabsolutelevelofaTheLeadershipQuarterly22(2011)383–398⁎Correspondingauthor.Tel.:+18172576796.E-mailaddress:m.s.cole@tcu.edu(M.S.Cole).1048-9843/$–seefrontmatter©2011ElsevierInc.Allrightsreserved.doi:10.1016/j.leaqua.2011.02.012ContentslistsavailableatScienceDirectTheLeadershipQuarterlyjournalhomepage:(e.g.,rwg;James,Demaree,&Wolf1984)aretypicallythencomputedtodeterminewhethercreatingaggregatescoresfromindividualleveldataisempiricallyjustifiable;unexplainedvarianceisregardedasnoiseormeasurementerror(Kozlowski&Klein2000).Incontrasttothisview,ithasbeenincreasinglyarguedthatunexplainedvarianceamongteammembers’assessmentsmayconveyrelevantunit-levelinformationratherthansimplyreflecterrorvariance(Chan1998;Harrison&Klein2007).Exploringthispossibility,leadershipresearchershaverecentlyconsideredwhataretermeddispersionmodelsofgroupcomposition.Inadispersioncompositionmodel,within-groupvariance(or,alternatively,within-groupagreementorconsensus)istreatedasameaningfulhigherlevelconstructratherthanastatisticalprerequisiteforaggregation(Chan1998).Withintheleadershipdomain,theextenttowhichteammembersagreeaboutthequalityoftheirmanagers’leadershipbehaviorhasbeendubbedleadershipconsensus(Feinberg,Ostroff,&Burke2005;seealsoLindell&Brandt2000).Inadispersioncompositionmodel,leadershipconsensusisconsideredhighwhenteammembers’perceptionsofaleader'sbehaviorarehomogenous.Conversely,leadershipconsensusisconsideredlowwhenteammembers’perceptionsofaleader'sbehaviorareheterogonousordissimilar.Ineitherinstance,thedegreeofconsensusis,bydefinition,agroup-levelmeasureoftheamountofvariabilityinteammembers’perceptionsofthequalityoftheirmanagers’leadershipbehavior.Asdiscussedbelow,evidencefromahandfulof