Emotional intelligence and job satisfaction, Testing the mediatory role of positive and negative aff

www.elsevier.com/locate/paid

Emotionalintelligenceandjobsatisfaction:Testingthemediatoryroleofpositiveandnegativeaffectatwork

KonstantinosKafetsios

aba,*,LeonidasA.Zampetakis

bDepartmentofPsychology,UniversityofCrete,Rethimnon74100,Crete,Greece

DepartmentofProductionEngineeringandManagement,TechnicalUniversityofCrete,Crete73100,Greece

Received25June2007;receivedinrevisedform21September2007;accepted3October2007

Availableonline14November2007

Abstract

Thestudytestedtheextenttowhichpositiveandnegativeaffectatworkmediatepersonalityeffects(Emo-tionalIntelligence)onjobsatisfaction.Participantswere523educatorswhocompletedtheWongLawEmo-tionalIntelligenceScale,aversionoftheJobAffectScaleandtheGeneralIndexofJobSatisfaction.ResultsusingstructuralequationmodellingindicatedthatpositiveandnegativeaffectatworksubstantiallymediatetherelationshipbetweenEIandjobsatisfactionwithpositiveaffectexertingastrongerinfluence.Inmales,affectatworkfullymediatedtheEIeffectonjobsatisfaction.AmongthefourEIdimensions,useofemotionandemotionregulationweresignificantindependentpredictorsofaffectatwork.Theresultsconfirmexpec-tationsderivingfromAffectiveEventsTheoryregardingtheroleofworkaffectivityasaninterfacebetweenpersonalityandworkattitudesandextendtheliteratureonEIeffectsinorganizationalsettings.Ó2007ElsevierLtd.Allrightsreserved.

Keywords:Emotionalintelligence;Affectatwork;Jobsatisfaction

1.Introduction

Inthecontextoftheemerging‘affectiverevolution’insocialandorganizationalpsychology(Barsade&Gibson,2007)Emotionalintelligence(EI)isproposedasanimportantpredictorof

*Correspondingauthor.Tel./fax:+302831077534.

E-mailaddress:k.kafetsios@psy.soc.uoc.gr(K.Kafetsios).

0191-8869/$-seefrontmatterÓ2007ElsevierLtd.Allrightsreserved.doi:10.1016/j.paid.2007.10.004

K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722713

keyorganizationaloutcomesincludingjobsatisfaction(Daus&Ashkanasy,2005;VanRooy&Viswesvaran,2004).ThereisaccumulatingevidencethatEIabilitiesandtraitsinfluencejobsat-isfaction(e.g.,Carmeli,2003;Sy,Tram,&O’Hara,2006)buttoourknowledge,thereisnostudytestingtheextenttowhichaffectivecomponentsoftheworkexperiencemaymediatesuchaconnection.Theoriesofemotioninorganizations(i.e.,AffectiveEventsTheory,Weiss&Cro-panzano,1996)suggestthataffectivestatesatworkarekeyvehiclesofpersonalityandorganiza-tionalinfluencesonjobsatisfactionandperformance.Moreover,ithasbeenpointedoutthatthereislittleresearchexaminingdispositionalfactorsofjobsatisfaction(especiallyaffect-related;Connolly&Viswesvaran,2000).Emotionalintelligencethuscanconstitutealinkbetweentraitlevelaffectivity,workaffectandjobsatisfactionandthepresentstudytestedsuchamodelusingastructuralequationsmodelingapproach.1.1.DefiningEI

AtatheoreticallevelEIreflectstheextenttowhichapersonattendsto,processes,andactsuponinformationofanemotionalnatureintra-personallyandinter-personally.However,thereareensuingdebatesattheoperationallevelmainly,thathaveledtotwodistinctapproaches:theabilityandtraitEmotionalIntelligence.TheabilityapproachusesmaximumperformancemeasuressuchastheMayer,Salovey&CarusoEmotionalIntelligenceTest(Mayer,Salovey,&Caruso,2002)toassessindividualdifferencesintheinterfaceofemotionwithcognitivepro-cesses(Mayer&Salovey,1997).TraitEIontheotherhandincorporatesEIabilitieswithinamoregeneralframeworkofindividualself-perceivedemotionalityandemotionefficacy(e.g.,Petrides&Furnham,2000).Nevertheless,sometraitEIapproachesandcorrespondingmeasures(e.g.,EQi,Bar-On,1997)divergefromtheconceptualizationofEIdescribedinthebeginningofthissectioninthattheyfocusonindividualdifferencesinadaptationtoenvironmentaldemands.1.2.EIandjobsatisfaction

Thereareseveralreasonswhyworkers’EImayinfluencejobsatisfaction.Interpersonally,emo-tionawarenessandregulatoryprocessesassociatedwithEIareexpectedtobenefitpeoples’socialrelationshipshenceaffectingtheexperienceofemotionandstressatwork.Intrapersonally,useofemotionandbeingawareofone’sownemotionscanleadtoregulatingstressandnegativeemo-tionsothatonecanperformbetteratwork.

ResearchthathasempiricallyexaminedrelationshipsbetweenEIandjobsatisfactionhasre-turnedmixedfindings.AnumberofstudieshaveobservedweaktomodestrelationshipsbetweentraitEImeasures(i.e.,EQi,Carmeli,2003;Kafetsios&Loumakou,2007;aGreektraitEIscale,Vacola,Tsaousis,&Nikolaou,2003)andjobsatisfaction.Arecentstudyoffoodserviceworkersandtheirmanagers(Syetal.,2006)observedapositiveassociationbetweenanabilitybasedEIscale(WongLawEmotionalIntelligenceScale,WLEIS;Wong&Law,2002)andjobsatisfactioninemployeesandtheirmanagers.Finally,astudyofagroupofmanagersLopes,Grewal,Kadis,Gall,andSalovey(2006)foundlinksbetweenEIabilitiesandaffectiveproxiesofjobsatisfactionmeasuredviaselfandsupervisorreports.

LikeinSyetal.’s(2006)studyweusedaself-reportmeasureofEI(WLEIS)thatconceptuallysubscribestotheabilityapproachandhasshowntohavegooddiscriminantandpredictive

714K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722

validityinorganizationalsettings.ResearchhasshownthatthescaleisdistinctfromtheBigFivepersonalityfactorsandhasconvergentvaliditywithotherEImeasuressuchastheTraitMeta-moodscale(Law,Wong,&Song,2004).TheWLEIShasproventobeapredictorofjobsatis-factionindifferentorganizationallevels(Lawetal.,2004;Syetal.,2006).1.3.Theroleofpositiveandnegativeaffect

Workaffectisanimportantaspectoftheworkexperience.AffectiveEventsTheory(Weiss&Cropanzano,1996)proposesthatcumulativeaffectiveexperiencesintheworkenvironmentto-getherwithotherfactors(includingpersonality)shapeworkers’jobrelatedattitudes.Affectatworkcannotbeequatedtojobsatisfactionsinceworkattitudesalsoreflectcognitiveevaluationsandbeliefs(Weiss,2002).Affectisasubjectivefeelingstatewithapositiveandnegativehedonictoneandshouldbedistinguishedfromdiscreteemotionsandmoodsinthatithasaspecificcon-textualelementbutnotaparticulartargetorelicitor(Frijda,1986).Therearedebatesoverwhetherthedimensionsthatunderliepositiveandnegativeaffectivityarebestdescribedintermsofactivation(e.g.,Watson,Clark,&Tellegen,1988)orvalence(Barrett&Russell,1998).Basedonevidencethathedonictoneandnotactivationisasignificantcorrelateofjobsatisfaction(Weiss,Nicholas,&Daus,1999)thepresentstudydistinguishedprimarilybetweenpositiveandnegativeaffectivity.Positiveandnegativeaffectasdifferentbutcorrelateddimensions(Russell&Carroll,1999).

1.4.Overviewofthestudy’saimsandhypotheses

Thestudyaimedfirstlytodeterminewhether,andtheextenttowhich,EIisassociatedwithaffectatworkandjobsatisfaction.Basedonrecentfindings(Lopesetal.,2006;Syetal.,2006)weexpectedthatself-reportedEIabilitieswillbeassociatedwithpositiveaffectandjobsatisfac-tionandinverselyrelatedtonegativeaffectatwork.Moreover,wewantedtoexaminewhichEIdimensionsmayberelatedtopositiveandnegativeaffectandjobsatisfaction.Secondly,weaimedtotestwhether,andtheextenttowhich,positiveandnegativeaffectatworkmediateEIeffectsonjobsatisfaction.BasedonAffectiveEventsTheory,weexpectedthatworkaffectwouldatleastpartiallymediateEIeffectsonjobsatisfaction;wedidnothaveanyhypotheseswithregardstotherelativestrengthofpositiveandnegativeaffectinthisrelationship.WeappliedaStructuralEquationsModellingapproachinordertoclearlydistinguishthemediatorypowerofpositiveandnegativeaffectwhilecontrollingforcommonvariancebetweenpositiveandnegativeaffect.2.Method2.1.Participants

Participantswere523teachers,workinginprimaryandsecondaryeducationfromvariousre-gionsinGreece.Thesampleincluded155malesand368femalesaged25–59years(M=38.48years,SD=8.02).Theaveragelengthofserviceasaneducatorwas12years(SD=8.48).Halfoftheparticipants(51%)wereinvolvedinteachinginprimaryeducation.Administrationof

K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722715

thequestionnaireswascarriedoutbytrainedpost-graduatestudentswhoactedasresearchassis-tantsandnomonetaryincentivewasprovided.2.2.Measures

AllscalesweretranslatedintoGreek(bythefirstauthor),blindlyback-translatedbyaGreekgraduatestudentwithsomeitemsmodifiedtoenhancethenaturalismofthetranslations(VandeVijver&Leung,1997).

2.2.1.Jobsatisfaction

WeadaptedintoGreektheGeneralIndexofJobSatisfaction(Brayfield&Rothe,1951).Thescalecomprisesof18items(Cronbach’sa=.92).Basedontheresultsofanexploratoryfactoranalysiswecreatedfourparcelsofthisconstruct.Weassigneditemstoindicatorsontherelativesizeoftheirfactorloadingsinordertoevenlydistributeitemsacrossindicators.Coefficientalphasforthefourscaleswere:Job1–.81;Job2–.81;Job3–.80;Job4–.79.

2.2.2.Emotionalintelligence

Weusedtheself-reportWongLawEmotionalIntelligenceScale(WLEIS,Wong&Law,2002).ThescaleconsistsoffourdimensionsthatareconsistentwithMayerandSalovey’s(1997)defini-tionofEI.TheSelf-EmotionAppraisal(SEA)dimensionassessesanindividual’sself-perceivedabilitytounderstandtheiremotions.TheOthers’EmotionAppraisal(OEA)dimensionassessesaperson’stendencytobeabletoperceiveotherpeoples’emotions.TheUseofEmotion(UOE)dimensionconcernstheself-perceivedtendencytomotivateoneselftoenhanceperformance.TheRegulationofEmotion(ROE)dimensionconcernsindividuals’perceivedabilitytoregulatetheirownemotions.Coefficientsalphasforthefourelementswere:SEA:.83;OEA:.77;ROE:.83;UOE:.79.

Weusedconfirmatoryfactoranalysis(AMOS7.0;Arbuckle,2006)toevaluatethefactorialstructureoftheGreekversionoftheWLEIS.Wecompared2alternativemodels:Model1spec-ifiedasinglefactorbehindallthe16items,whileModel2specifiedthefourcorrelateddimensionsfromtheirrespectiveitemsandthenasecond-orderfactorbehindthefourEIdimensions.Model2fittedthedatabetter[v2(100,N=523)=423.43,p=0.000;RMSEA=0.079(90%CI:0.071–0.084);GFI=0.915;CFI=0.932;andAIC=495.43]thanModel1[v2(104,N=523)=1084.16,p=0.000;RMSEA=0.134(90%CI:0.127–0.142);GFI=0.770;CFI=0.744andAIC=1148.84].Theseresultsareinlinewithpreviousstudies(Lawetal.,2004;Wong&Law,2002)andindicatethatitemsforEImeasurementcanserveasareasonableestimateoftheirdimensions,andthatthedimensionsinturncanrepresentanunderlyingmultidimensionalEIconstruct.Theinternalconsistencyforall16itemswasa=.90.

2.2.3.Positiveandnegativeaffectatwork

Toassessaffectatworkweused12itemsoftheJobAffectScale(Brief,Burke,George,Robinson,&Webster,1988)thatassessesparticipants’experienceofpositiveandnegativeaf-fectatworkduringthepreviousweekona5-pointscale.Thesixpositiveitems(JAS-PA)were:enthusiastic,elated,active,strong,happy,andexcited.Thesixnegativeaffectivestates(JAS-NA)were:hostile,scornfull,fearful,sleepy,placid,andsad.Thepositiveandnegative

716K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722

affectpartsofthescalehadgoodinternalconsistency(Alphas.and.78,respectively).Wefollowedthesameprocedureaswiththejobsatisfactionconstructandwecreatedtwoindi-catorsforbothpositiveandnegativeparts.Alphasforthescaleswere:PA1–.82;PA2–.81;NA1–.68;NA2–.61.2.3.Analyticstrategy

InthepresentstudywetestedwhetherpositiveandnegativeaffectivityatworkmediatedfullyorpartiallytherelationshipbetweentraitEIandjobsatisfactioninthecontextofStruc-turalEquationModellingusingmaximumlikelihoodestimationmethod.Priortotheanalysisdatascreeningwasperformedanddataweretestedfordeviationfromnormality.FollowingHall,Snell,andFoust(1999)weformeditemparcelsonthebasisoffactoranalysisinordertocontrolforinflatedmeasurementerrorsandimprovethepsychometricpropertiesofthevariables.Weusedatwo-stageanalyticprocedure:instage1thefour-factormodelwasfittedtothedataandthenameasurementmodelspecifyingperfectcorrelationamongallfourlatentvariableswasassessedtotestoveralldiscriminability.Theone-factormodelalsoprovidesatestforcommonmethodbias.Weusedthesequentialv2differentialtest(SCDT)toassessnestedmodelcomparisons.Weemployedseveralmodelfitstatistics(RMSEA:RootMeanSquareErrorApproximation;CFI:ComparativeFitIndex;GFI:GoodnessofFitIndex;RMR:RootMeanSquareResidual;AIC:AkaikeInformationCriterion;Shook,Ketchen,Hult,&Kacmar,2004).InordertoselectamongthecompetingstructuralmodelsweappliedmodelselectionforSEM(Raftery,1993).Weemployedastepwisestrategyinmodelselection,whichincludedforwardselectionandbackwardeliminationfeatures.WeusedSteiger’sPowerAnalysis(StatSoft,2001)toestimateSEMmodel-levelpower.Finally,weusedbootstrappingprocedures(resampled1000timesandusedthepercentilemethodtocreate95%confidenceintervals).3.Results

Table1presentsmeans,standarddeviationsandvariableintercorrelations.TraitEIwassignif-icantlyrelatedtojobsatisfaction,positiveaffectandnegativeaffect.Table1alsorevealedthatjobsatisfactionwaspositivelyandsignificantlyrelatedtopositiveaffectandnegativelyrelatedtoneg-ativeaffect.

3.1.Assessmentofmeasurementmodel

Table2displaysthefitstatisticsforthemeasurementmodel.Overall,thehypothesizedmea-surementmodelfitthedataquitewellwhenevaluatedintermsoftherecommendedcutoffsorthecombinationcutoffapproach(Shooketal.,2004).Thehypothesizedmeasurementmodelfitthedatabetterthanasinglefactormodel,bothintermsofthefitstatisticsandwhendirectlycontrastedwithachangeinv2testandAIC.Insummary,theresultssuggestthattheproposedfactorstructurepresentsastatisticallyadequateandsufficientfittothedata,indicatingtheab-senceofseverecommonmethodvariance.

K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722

Table1

Descriptivestatisticsandintercorrelationsfortotalsample

M

1.2.3.4.5.6.7.

GenderaAge

YearsinserviceTraitEI

JobsatisfactionPositiveaffectNegativeaffect

38.4811.905.283.883.461.41

SD8.028.470.810.650.800.55

1–

À0.13**0.030.040.01*0.06À0.03

2–

3

4

5

6

7

717

0.66**À0.000.060.03À0.14**–

À0.030.020.01À0.13**(0.90)0.43**0.42**À0.27**(0.92)0.**À0.44**(0.)À0.36**(0.78)

Note:N=523.Internalreliabilitiesareinparenthesis.aGenderiscoded1=male2=female.*p<0.05.**p<0.01.

Table2

MeasurementmodelsfitstatisticsModel

HypothesizedfourfactormeasurementmodelOnefactor

measurementmodel

v2230.16**706.83**df48

5.47**Dv2RMSEA

0.085(90%CI:0.074–0.096)

GFI0.926

CFI0.948

RMR0.032

AIC290.16

520.115(90%CI:0.145–0.166)0.8120.8140.171758.83

Note:v2:chi-squarestatistic.**p<0.001.

3.2.Assessmentofstructuralmodels

Thenextstepwastoconsidercomparativemodelsspecifyingtotaleffects(directandindirect),completemediationandpartialmediation.Resultsofthemodelselectionprocedureindicatedthattherewas97.9%probability(intermsofAkaikeweights)thatthebestmodelistheonepresentedinFig.1.Weconsideredtheerrortermsofpositiveandnegativeaffectcorrelatedinlinewithpre-viousresearch(Barrett&Russell,1998).

Thismodelrevealedagoodfittothedata:v2(48,N=523)=230.16,p=0.000;RMSEA=0.085(90%CI:0.074–0.096);GFI=0.926;CFI=0.948;RMR=0.032andAIC=290.16.ThemodelpostulatedthattheeffectsoftraitEIonjobsatisfactionwerepartiallymediatedbypositiveandnegativeaffect(seeTable3).Poweranalysis(e1=0.08,a=0.05,N=523,df=48)suggestedaverylowprobabilitytorejectthemodel(.5%).ThestandardizeddirecteffectofEIonjobsatisfactionwas0.14(p<0.01).EIhaddirecteffectsonpositiveaffect(0.49,p<0.01)andnegativeaffect(À0.36,p=0.002)alongwithindirecteffectsonjobsatisfac-tion(0.24,p<0.01).Insum,thestandardizedtotaleffectofEIonjobsatisfactionwas0.49(95%percentileconfidenceinterval:0.40–0.58,p<0.01).

Positiveaffecthadastatisticallysignificantdirecteffectonjobsatisfaction(0.56,95%percentileconfidenceinterval(CI):0.43–0.66,p<0.01)Negativeaffecthadadirectnegativeeffectonjob

718K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722

Fig.1.Standardizedresultsofthestructuralmodelassessment.

Table3

Standardizeddirectandindirecteffectsandtheassociated95%confidenceintervalsPredictor

OutcomePositiveaffectDirect

TraitEIPositiveaffectNegativeaffect

0.49*[0.39–0.58]––

Indirect–––

NegativeaffectDirect

À0.36*[(À0.47)–(À0.26)]––

Indirect–––

JobsatisfactionDirect

0.14*(0.04–0.25)0.56*(0.43–0.66)

À0.19*[(À0.30)–(À0.075)]

Indirect0.24*(0.18–0.32)––

Note:Theupperandlowerboundsofthe95%confidenceinterval(showninparentheses)werebasedonthefindingsfromabootstrappinganalysisusingthepercentilemethod.*p<0.01.

satisfaction(À0.19,95%CI:(À0.30)–(À0.075),p<0.01).Thecorrelationbetweentheerrortermsofpositiveandnegativeeffectwassignificant(À0.36,p<0.01,95%CI:(À0.50)–(À0.22),p<0.01).Theproportionofvarianceinjobsatisfactionexplainedbythecollectivesetofpredic-torswas57%.

K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722719

3.3.Genderdifferences

WefoundnostatisticallysignificantgenderdifferencesinEIandworkaffectbutwomenscoredhigherthanmenonjobsatisfactionatastatisticallysignificantlevel.Interestingly,jobsatisfactiondevelopmentfolloweddifferentpatternsbetweengenders.FormaleteachersthemostprobablemodelintermsofAkaikeweights(92%probability)wastheonewherepositiveandnegativeaffectfullymediatetraitEIeffectsonjobsatisfaction.ThestandardizeddirecteffectoftraitEIonpo-sitiveaffectwas0.62(95%CI:0.43–0.77,p<0.01),andonnegativeeffectwas(À0.40),(95%CI:(À0.56)–(À0.21),p<0.01).Positiveaffecthadastrongdirecteffectonjobsatisfaction0.77,(95%CI:0.56–0.98,p<0.01),whilethedirecteffectofnegativeaffectonjobsatisfactionwasnotsig-nificantlydifferentfromzero0.17,(95%CI:(À0.39)–0.02,p=0.09).Thecorrelationbetweentheerrortermsofpositiveandnegativeaffectwerenotsignificantlydifferentfromzeroatthe0.05level:0.24,(95%CI:(À0.5)–0.002,p=0.053).Theaforementionedrelationshipsamongvari-ablesexplained71%ofthevarianceinjobsatisfaction.

Ontheotherhand,forfemalesthemostlikelymodel(99%)wasonewheretraitEIhadadirectandindirecteffectonjobsatisfaction,thatistosay,therelationshipbetweenEIandjobsatisfactionispartiallymediatedbypositiveandnegativeaffect.ThestandardizeddirecteffectoftraitEIonjobsatisfactionwas0.13,(95%CI:0.06–0.23,p=0.002);onpositiveaffectwas0.35(95%CI:0.24–0.46,p=0.002),andonnegativeaffectwas(À0.19),(95%CI:(À0.32)–(À0.09),p=0.002).Therewasastrongdirecteffectfrompositiveaffecttojobsatisfaction0.40,(95%CI:0.27–0.55,p=0.002),whilethedirecteffectofnegativeaffectonjobsatisfac-tionisÀ0.24,[95%CI:(À0.39)–(À0.03),p=0.007].ThetotalstandardizedeffectoftraitEIonjobsatisfactionwas0.47,(95%CI:0.35–0.57,p=0.002).Thecorrelationbetweentheerrortermsofpositiveandnegativeaffectwassignificantlydifferentfromzeroatthe0.05level:0.42,(95%CI:(À0.58)–(À0.23),p=0.02).Theaforementionedrelationshipsamongvariablesexplained51%ofthevarianceinjobsatisfaction.3.4.SeparatingtheeffectsoftheEIdimensions

FinallyweperformedthreemultipleregressionanalysesinordertodeterminetherelativeeffectsofthefourEIbranchesonaffectatworkandjobsatisfaction.ThefourEIbranchesexplained21%inthepositiveemotionvarianceandthismodelwassignificant(F(4,518)=33.96,p<.001).AmongthefourEIbranchesonlyuseofemotionandregulationofemotionweresig-nificantpredictorsofpositiveaffectatwork(b=.14,p<.01andb=.36,p<.001,respectively).ThefourEIbranchesexplained8%inthenegativeemotionvariance(F(4,518)=17.7,p<.001)andagain,useofemotionandemotionregulationweresignificantindependentpredictors(b=À.12,p<.01)and(b=À.21,p<.001,respectively).Finally,threeofthefourEIbranchesweresignificantindependentpredictorsofjobsatisfaction(otherappraisalofemotion,b=.13,p<.05,useofemotion,b=.16,p<.01,regulationofemotion,b=.28,p<.001,R2=20%,F(4,518)=32.03,p<.001).Enteringpositiveandnegativeaffectinasecondsteprenderedthepartialcorrelationsofemotionandregulationofemotionwithjobsatisfactionnon-significant.Understandingofemotionhowever,retainedasignificantuniquepredictiononjobsatisfaction(b=.09,p<.05),suggestingthatdirecteffectsfromtraitEIonjobsatisfactionmayowethistotheEIdimension.

720K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722

4.Discussion

Thepresentstudyextendsanemergingbodyofresearchonaffectivityintheworkplacebytest-ingforlinksbetweentraitlevelemotionality(EI),affectatworkandjobsatisfaction.Theresultsunderlinetheimportantroleofpositiveandnegativeaffectatworkinthisrelationship.

Inkeepingwithrecentstudies(Lopesetal.,2006;Syetal.,2006)theresultsdemonstratedcon-vincinglythatEIisanimportantpersonality-levelpredictorofworkaffectivityandjobsatisfac-tion.SubsequentregressionanalysesindicatedthatuseofemotionandemotionregulationaretwoEIdimensionspredictiveofpositiveandnegativeaffectatworkandperceivingothers’emo-tionswasuniquelyassociatedwithjobsatisfactionafindingthatprimarilyconcernedwomen.Notably,thesefindingsareatoddswithstudiesshowingweakrelationshipsbetweensometraitEImeasures(i.e.,Kafetsios&Loumakou,2007)andjobsatisfaction.Giventhatsomeofthisre-searchevidenceconcernedthesamepopulationasinthisstudyonemayexcludeorganizationallevelvariablesaspossiblemoderatorsandfocusonthemeasurementinstrumentsbeingused(Brackett&Mayer,2003).Inthepresentstudy,weusedaself-reportmeasureofEI(WLEIS)thatadherestotheabilitymodelofEIandconfirmeditspsychometricpropertiesandutilityforanon-Englishspeakingculture.

TheresultsfromtheSEMsuggestedthatinabsolutetermsthedirecteffectofEIisstrongerforpositiveaffectcomparedtonegativeaffect.Thisisanimportantfindingsupportingaccumulatingevidencefortheprimacyofpositiveovernegativeaffectasapredictorofworkoutcomes(Tho-resen,Kaplan,Barsky,&deChermont,2003).Positiveaffectisasourceofhumanstrength(Isen,2003)andpositiveaffectpredisposespeopletocognitions,feelingsandactionsthatpromotethebuildingofpersonalandsocialresources(Fredrickson,2001;Lyubomirsky,King,&Diener,2005).PersonshighonEIseemtobebestsuitedtofollowthese‘broadeningandbuild’strategiesalsointheworkenvironment.

Thereweresomegenderdifferencesinthepatternofresults.Maleworkers’positiveandnegativeaffectatworkfullymediatedEIeffectsonjobsatisfactionwhereasinfemaleworkerssomedirectef-fectswereobservedatalevelcomparablewiththefinalmodelpresentedinFig.1.Theresultsfromsubsequentregressionanalysesindicatedthatwhencontrollingforpositiveandnegativeaffectatworkperceivingothers’emotionwasuniquelyassociatedwithjobsatisfactionafindingthatmayre-flectfemaleworkrelatedgenderrole-characteristics(Petrides&Furnham,2006).

Importantly,theresultsfromthecurrentconfirmatorymodelsupportkeyassumptionsofAffectiveEventsTheory(Weiss&Cropanzano,1996)regardingtheroleofpositiveandnegativeaffectatwork.Thetheorypromotesamultifacetedviewofjobsatisfactionincorporatingaffective,cognitiveandpersonalityelements(Weiss,2002).Clearly,positiveandnegativeaffectivityatworkexplainedasignificantpartinjobsatisfactionvarianceasexpectedbyAETsuggestingthateventstakingplaceatworkinfluenceemployees’attitudestowardstheirwork.Theresultspointtothepersonalityantecedentstoaffectatworkandfurtherresearchcouldstrivetodistinguishthoseef-fectsfromorganizationallevelvariables(e.g.,organizationalclimate,managerialstructure,etc.).4.1.Limitationsandfurtherresearch

OneofthelimitationsofthecurrentstudywastheadherencetoadimensionalmodelofworkemotionalityandfutureresearchcouldtestwhetherdifferentfacetsofEIareassociatedwith

K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722721

discreteemotionsatwork.Furthermore,thestudyusedaretrospectivemeasureofworkaffectiv-ityandfutureresearchshouldexpandthestudyofaffectivephenomenaatworkusingmeasuresof‘on-line’affectsucheventsamplingmethodologiesorobservationalmethods.Finally,inlinewithAET,futureresearchcouldidentifytheworkeventsthatgiverisetopositiveandnegativeemo-tionatworkandforwhichEIactsasamoderator.ItwouldalsobeinterestingtoexaminetheextenttowhichEIandaffectatworkinterfacetoinfluenceworkattitudesinoccupationswithjobcharacteristicsdifferentthaneducators’.References

Arbuckle,J.L.(2006).AMOS7.0.Userguide.Chicago:SmallWatersCorporation.

Bar-On,R.(1997).Bar-onemotionalquotientinventorytechnicalmanual.Toronto:Multi-HealthSystems.

Barrett,L.,&Russell,J.A.(1998).Independenceandbipolarityinthestructureofcurrentaffect.JournalofPersonalityandSocialPsychology,74,967–984.

Barsade,S.G.,&Gibson,D.E.(2007).Whydoesaffectmatterinorganizations?AcademyofManagementPerspectives,21(1),36–57.

Brackett,M.A.,&Mayer,J.D.(2003).Convergent,discriminant,andincrementalvalidityofcompetingmeasuresofemotionalintelligence.PersonalityandSocialPsychologyBulletin,29,1147–1158.

Brayfield,A.,&Rothe,H.(1951).Anindexofjobsatisfaction.JournalofAppliedPsychology,35,301–311.

Brief,A.P.,Burke,M.J.,George,J.M.,Robinson,B.S.,&Webster,J.(1988).Shouldnegativeaffectivityremainanunmeasuredvariableinthestudyofjobstress?JournalofAppliedPsychology,73,193–198.

Carmeli,A.(2003).Therelationshipbetweenemotionalintelligenceandworkattitudes,behaviorandoutcomes.JournalofManagerialPsychology,18,788–813.

Connolly,J.J.,&Viswesvaran,C.(2000).Theroleofaffectivityinjobsatisfaction:Ameta-analysis.PersonalityandIndividualDifferences,29,265–281.

Daus,C.S.,&Ashkanasy,N.M.(2005).Thecasefortheability-basedmodelofemotionalintelligenceinorganizationalbehaviour.JournalofOrganizationalBehavior,26,453–466.

Fredrickson,B.L.(2001).Theroleofpositiveemotionsinpositivepsychology:Thebroaden-and-buildtheoryofpositiveemotions.AmericanPsychologist,56,218–226.

Frijda,N.(1986).Theemotions.Cambridge:CambridgeUniversityPress.

Hall,R.J.,Snell,A.F.,&Foust,M.S.(1999).ItemparcelingstrategiesinSEM:Investigatingthesubtleeffectsofunmodeledsecondaryconstructs.OrganizationalResearchMethods,2(3),233–256.

Isen,A.(2003).Positiveaffectasasourceofhumanstrength.InL.Aspinwall&U.Staudinger(Eds.),Apsychologyofhumanstrength(pp.179–195).Washington,DC:APA.

Kafetsios,K.,&Loumakou,M.(2007).Acomparativeevaluationoftheeffectsoftraitemotionalintelligenceandemotionregulationonaffectatworkandjobsatisfaction.InternationalJournalofWorkOrganizationandEmotion,2(1),71–87.

Law,K.S.,Wong,C.,&Song,L.J.(2004).Theconstructandcriterionvalidityofemotionalintelligenceanditspotentialutilityformanagementstudies.JournalofAppliedPsychology,,483–496.

Lopes,P.N.,Grewal,D.,Kadis,J.,Gall,M.,&Salovey,P.(2006).Evidencethatemotionalintelligenceisrelatedtojobperformanceandaffectandattitudesatwork.Psicothema,18(1),132–138.

Lyubomirsky,S.,King,L.,&Diener,E.(2005).Thebenefitsoffrequentpositiveaffect:Doeshappinessleadtosuccess?PsychologicalBulletin,131(6),803–855.

Mayer,J.D.,&Salovey,P.(1997).Whatisanemotionalintelligence?InP.Salovey&D.Sluyter(Eds.),Emotionaldevelopmentandemotionalintelligence:Implicationsforeducators.NewYork:BasicBooks.

Mayer,J.D.,Salovey,P.,&Caruso,D.R.(2002).MSCEITuser’smanual.Toronto,Ontario:Multi-HealthSystemsInc.

Petrides,K.V.,&Furnham,A.(2000).Onthedimensionalstructureofemotionalintelligence.PersonalityandIndividualDifferences,29,313–320.

722K.Kafetsios,L.A.Zampetakis/PersonalityandIndividualDifferences44(2008)712–722

Petrides,K.V.,&Furnham,A.(2006).Theroleoftraitemotionalintelligenceinagender-specificmodeloforganizationalvariables.JournalofAppliedSocialPsychology,36(2),552–569.

Raftery,A.E.(1993).Bayesianmodelselectioninstructuralequationmodels.InK.A.Bollen&J.S.Long(Eds.),Testingstructuralequationmodels(pp.163–180).NewburyPark,CA:SagePublications.

Russell,J.A.,&Carroll,J.M.(1999).Onthebipolarityofpositiveandnegativeaffect.PsychologicalBulletin,125,3–30.

Shook,C.L.,Ketchen,D.J.J.,Hult,G.T.M.,&Kacmar,K.M.(2004).Anassessmentoftheuseofstructuralequationmodelsinstrategicmanagementresearch.StrategicManagementJournal,25,397–404.StatSoft,(2001).StatisticaforWindowsv.6.0:Dataanalysissoftwaresystem.Tulsa,OK:USA.

Sy,T.,Tram,S.,&O’Hara,L.A.(2006).Relationofemployeeandmanageremotionalintelligencetojobsatisfactionandperformance.JournalofVocationalBehavior,68,461–473.

Thoresen,C.J.,Kaplan,S.A.,Barsky,A.P.,&deChermont,K.(2003).Theaffectiveunderpinningasofjobperceptionsandattitudes:Ameta-analyticreviewandintegration.PsychologicalBulletin,129(6),914–945.

Vacola,M.,Tsaousis,I.,&Nikolaou,I.(2003).Theroleofemotionalintelligenceandpersonalityvariablesonattitudestowardorganizationalchange.JournalofManagerialPsychology,19(2),88–110.

VandeVijver,V.,&Leung,K.(1997).Methodsanddataanalysisforcross-culturalresearch.ThousandOaks,CA:Sage.VanRooy,D.L.,&Viswesvaran,C.(2004).Emotionalintelligence:Ameta-analyticinvestigationofpredictivevalidityandnomologicalnet.JournalofVocationalBehavior,65(1),71–95.

Watson,D.,Clark,L.A.,&Tellegen,A.(1988).Developmentandvalidationofbriefmeasuresofpositiveandnegativeaffect:ThePANASscales.JournalofPersonalityandSocialPsychology,,1063–1070.

Weiss,H.M.(2002).Deconstructingjobsatisfaction:Separatingevaluations,beliefsandaffectiveexperiences.HumanResourceManagementReview,12,173–194.

Weiss,H.M.,&Cropanzano,R.(1996).Affectiveeventstheory:Atheoreticaldiscussionofthestructure,causesandconsequencesofaffectiveexperiencesatwork.InB.M.Staw&L.L.Cummings(Eds.).Researchinorganizationalbehavior(Vol.19,pp.1–74).Greenwich,CT:JAI.

Weiss,H.M.,Nicholas,J.P.,&Daus,C.S.(1999).Anexaminationofthejointeffectsofaffectiveexperiencesandjobbeliefsonjobsatisfactionandvariationsinaffectiveexperiencesovertime.OrganizationalBehaviorandHumanDecisionProcesses,78(1),24.

Wong,C.S.,&Law,K.S.(2002).Theeffectsofleaderandfolloweremotionalintelligenceonperformanceandattitude:Anexploratorystudy.TheLeadershipQuarterly,13,243–274.