SSCI期刊投稿與研究方法.ppt

1 SSCI期刊投稿與研究方法 政大企管系蔡維奇彰化師大人管所2006年12月29日 2 Agenda TheconceptualmodelandmethodologyfortwoofmypublishedpapersReviewers commentstoeachpaperandmyreplytoreviewerconcerns sample researchdesign procedure etc LessonslearnedFinalthoughts 3 TheroleofMethodology QualityofMethodsandAnalysis hasbeenviewedbyjournaleditorsasoneofthemostimportantcharacteristicsforpublishablejournalarticles 17of60items SeeDesrositersetal 2002 Writingresearcharticles Updateonthearticlereviewchecklist InS G Rogelberg Ed Handbookofresearchmethodsinindustrialandorganizationalpsychology p 460 Malden MA Blackwell 4 PaperReviewed 1 Tsai W C 2001 Determinantsandconsequencesofemployeedisplayedpositiveemotions JournalofManagement 27 497 512 JOBJOB 1strevision JAPJOM 1strevision JOM 2ndrevision JOM accepted 5 TheoreticalModel Psychologicalclimateforservicefriendliness Employeedisplayedpositiveemotions Purchasedecision Willingnesstoreturn Willingnesstorecommend ControlvariablesSexvariablesclerkexperiencedemotions Controlvariables Sex agevariablesclerkprofessionalability 6 7 Methods Measures 8 Methods Measures continued 9 CommentsandReply 1 Sample samplerepresentativeness Ifcustomerswhohadanegativeexperiencewithaclerkdeclinedparticipationmorefrequently couldyouhaveoversampledpositivecustomerreactions Wouldn tcustomersandsalesclerksinaworsemoodbelesslikelytoparticipate Ithinkthismightonlyrestricttherangeofdatainyoursample 10 CommentsandReply 2 ResearchDesign controlvariableYoushouldalsopointoutthatthereisaverygreatdealofvarianceincustomerreactions notexplainedbyyourmodel JOBEditor 11 CommentsandReply 2 ResearchDesign JAPversion thereisstillagreatdealofvarianceincustomerreactionswhichisnotexplainedbythemodelinthisstudy Otherpotentialpredictorsmightbestoreatmosphere e g musicinthestore insidedecoration facilities e g spaceforrest spaceforleisure andthesalesclerk sbehaviorsotherthandisplayedemotions Yoo Park MacInnis 1998 12 CommentsandReply 2 ResearchDesign severalkeyvariablesthatprovidealternativeexplanationsforyourresultsareomitted Forinstance clerks emotionaldisplaysmightcovarywithpriceorcustomercharacteristics andnoattempttoeliminatethesealternatives commentsfromJAPEditor 13 CommentsandReply 3 ResearchDesign directionofcausationGiventheabsenceofanexperimentalorquasi experimentaldesign nocausalinferencescanbedrawn Customerreactionsmayhavecauseddisplayedemotionorviceversa commentsfromJAPEditor OKwithJOMreviewers 14 CommentsandReply 3 ResearchDesign JOMversion becausethedataarecross sectional thedirectionofcausalitycannotbeunambiguouslydetermined Nonetheless somethoughtscometomindthatmaypartiallymitigatethisconcern BecausecustomersindicatedtheirreactionstowardthestoreAFTERtheyhadencounteredtheemployees emotionaldisplays itislessplausibletoarguethatcustomerreactionisinfactacauseofemployeedisplayedemotions Afterall causeshavetobeantecedenttotheireffects 15 CommentsandReply 4 ProcedureMoredescriptionisneededof a howobserversweretrained and b theactualprocessofobservingthetransactions Wereemployeesawareofthepresenceoftheobservers forexample 16 CommentsandReply 5 Measures constructvalidityIamnotsurethatyourmeasureof professionalability Iencourageyoutoestablishsomeargumentforitsconstructvalidity 17 CommentsandReply 5 Measures Alternatively thesalesclerk sprofessionalabilitycouldberatedbycustomers However thiswouldleadtotheproblemofcommonmethodbias Iamunabletoprovideevidence giventhekindofdataIhave 18 CommentsandReply 6 Measures socialdesirabilityDoyouthinksocialdesirabilityisdrivingtheresponsesofthesalesclerksforthePANASscales 19 CommentsandReply 6 MeasuresIagreetherewasapossibilitythatsalesclerksoverstatedtheirlevelsofpositiveaffect However ifsocialdesirabilitywasthemajorforce thevarianceofthisvariableshouldbecomerelativelysmall IcheckedthestandarddeviationvalueofPAreportedinWatson etal 1988 andfoundthatitwasatasimilarlevel SD 7 2 5 pointscale tothevaluefoundinthisstudy SD 6 4 4 pointscale Thus itseffectshouldnotbetooserious 20 CommentsandReply 7 Measures JOBinitialversion lowalpha alphaforthedisplayedemotionsmeasurewas 51 suggestingthatthesixcriteriawerenotsufficientlyhomogeneous However thecreationofthisvariablewasstilldeemedmeaningfulsinceeachofthesesixcriteriareflectedpartofthedisplayedemotionsconstructand AsarguedbyNunnallyandBernstein 1994 the heterogeneitywouldbealegitimatepartofthetestifitwerepartofthedomainofcontentimpliedbytheconstruct p 312 21 CommentsandReply 7 Measures Reviewercomments Idon tthinkyourdiscussionsurroundingtheissueoflowalphafordisplayedemotionsprovideanadequatejustification Ithinkmuchmoreneedstobedonetotryandcleanupthatvariable Iunderstandyourpointthattheconstructisheterogeneous Iseetwosolutions 1 selectasubsetoftheitems 2 treatthesixitemsasseparateindicatorsofdisplayedemotions 22 CommentsandReply 7 Measures JOMinitialversion Iaddedaparagraphtothemethodsectionasfollows Analternativewaytodealwiththelowalphaistotreatthesixcriteriaasseparateindicatorsofdisplayedpositiveemotionsratherthantocreateacomposite Ihaveruntheregressionsinthisway Iaddedaparagraphtothediscussion Consequently NOconcernsonthisissuehavebeenraisedbyJOMreviewers 23 CommentsandReply 8 Statisticalanalysis Itseemsitwouldbebettertotestalltherelationshipsatonetimeusingstructuralequationmodelingratherthantestingtherelationshipsseparately Iagreewiththereviewersthathierarchicalregressionismoreappropriate Evenbetterwouldbecausalmodeling Editoropinion 24 CommentsandReply 9 Cross culturalmethodology generalizationGiventhatyoursamplewasTaiwanese andthatthejournalisaUSbasedjournal youneedtoconvinceyouraudienceofgeneralizabilityofyourfindings Iwouldarguethatsalesclerks expressionsofeveryindicatorchoseninthisstudywouldmakecustomersfeeltheyarewarmlywelcomedinbothTaiwanandUS althoughitislikelythat emotionalindicators maynotbeidentical 25 CommentsandReply 9 Cross culturalmethodology foreignscalesYouusedAmericanscalesinaforeigncountry Thatisahugeproblem Withoutaccountingforthespecificculturaldifferencesyourfindingsareclosetomeaningless ConcernsaboutusingscalesoriginallywritteninEnglishinaforeigncountry maybepartiallymitigatedforthefollowingreasons SimilarfactorstructuresofthescalesSomecongruentresearchfindings 26 PaperReviewed 2 Tsai W C Chen C C Chiu S F 2005 Exploringboundariesoftheeffectsofapplicantimpressionmanagementtacticsinjobinterviews JournalofManagement 31 108 125 JOM 1strevision JOM 2ndrevision JOM accepted 27 TheoreticalModel Customer contactRequirement InterviewStructure InterviewerEvaluation ApplicantIMself focusednon verbal InterviewLength 28 Methods Participants Procedure 141applicantsofnon managerialpositions 33interviewersof24firmsinTaiwan 3industries Onlyoneinterviewerinvolvedineachinterview Surveycompletedbyinterviewer Surveycompletedbybothinterviewerandapplicant 29 Methods Measures 30 Methods Measures continued 31 Methods 9Controlvariables ApplicantPhysicalAttractiveness Sex Age andConscientiousnessInterviewerExperience Training Sex Age andPre interviewImpression 61 32 CommentsandReply 1 SampleTheapplicantsparticipatedinanaverageof6 09interviewsinpastyears Doesthismeanthattheyinterviewed6timesannually Ordoesthismean 33 CommentsandReply 2 ResearchDesign controlvariableIncludingninecontrolvariablescertainlyhelpseliminatealternativeexplanations Atthesametime itcutsintotheabilityofyouranalysestodetecteffectswithsmallsamplesizes Youmaywanttoconsidercollectingadditionaldata Toyourcredit youusedalotofcontrolvariableshere However insomesense Ithinkyouusedtoomany Pleasedrop Editoropinion 34 CommentsandReply 2 ResearchDesignWouldyouexpectindustryrelateddifferencestoinfluenceyourresults Ifso thenyoushouldcontrolforindustry Similardifferencesmaybefoundacrossorganizations althoughcontrollingforsuchdifferencesbecomesmoretroublesome 35 CommentsandReply 2 ResearchDesignInthisstudywecreatedtwodummyvariables withserviceindustryasthereferentcategory NotethatweDIDNOTinclude firm asacontrolvariable Giventhesmallsamplesizeinourstudy adding24firmvariablestotheregressionwouldgreatlyreducethestatisticalpowerofourstatisticalanalyses seeCohen Cohen 1983 thusitmaybeinappropriatetodoso 36 CommentsandReply 2 ResearchDesign 2ndcomments Iagreethatitwilltakeuptoomanydegreesoffreedomtoincludeadummyvariableforeachorganization However youmightacknowledgeinthelimitationsthatyouhavemultipledatapointsthatarelinkedtothesameorganizationandthesameinterviewer anddiscusshowthatmightinfluenceyouranalysis 37 CommentsandReply 2 ResearchDesign 2ndreply multipledatapointswerelinkedtothesameinterviewerandthesameorganization westillcouldnotcontrolforallinterviewereffectsand ofcourse theorganizationeffects Thismightviolatethestatisticalassumptionofindependentobservations Kenny LaVoie 1985 Weencouragefutureresearchtoutilizelargersamplesizesandtakeappropriateactions e g standardizeinterviewerratingsbeforepoolingacrossinterviewers todealwithsuchissues 38 CommentsandReply 3 Measures lowalphaAmajorissuewiththisstudyisthelackofreliablemeasures Severalalphasarebelow 7 andevenbelowthemarginallyacceptable 65 Myfirstpreference togobackandtrydifferenttypesofrotationsanddatareductiontechniques Ifthatdoesnotwork thenIwouldpreferyoutosimplyusesomesoloitems Editoropinion 39 CommentsandReply 3 MeasuresWefollowedyoursuggestionbyfirsttryingdifferenttypesofrotationsanddatareductiontechniques butcouldnotincreasealphas Thus assuggestedbyyou weusedasingleitemtomeasurenon verbaltactics 40 CommentsandReply 4 Measures factoranalysis Thereaderneedsmoreinformationabouthowthesefactoranalyseswereconducted e g methodofextraction methodofrotation etc 41 CommentsandReply 5 Measures descriptionofselecteditemsDidyouusethefull scalesfromKristofandhercolleaguestomeasureself focusedandnonverbalIMtactics Ifyouonlyusedselecteditems whatwerethedecisioncriteria 42 CommentsandReply 6 Statisticalanalysis Iftherearehighcorrelationsyoumightcenterthevariablesbeforeconstructingtheinteractionterms Jaccard Turrisi Van 1990 ifyouhavenotdonesoalready weaddedthefollowingexplanationtotherevision Tocounterproblemsofmulticollinearityintestsofinteractionterms wecenteredallindependentandmoderatingvariablesbeforecreatingtheinteractionterms 43 CommentsandReply 7 Cross culturalmethodology scaletranslationThestudywasconductedinTaiwan WerethesurveysinEnglish Andifso howfluentweretheparticipantsinEnglish 44 CommentsandReply 7 Cross culturalmethodology Infact allquestionnairesinthisstudywereinChinese asitistheofficiallanguageinTaiwan VandeVijverandLeung 1997 indicatedtheimportanceofevaluatingthetranslatabilityofaninstrumentinamultilingualcontext itemsshowculturalidiosyncrasies e g localidiomandmetaphors shouldbereplaced Itseemsthatscalesusedinourstudydidnotcontainsuchpoorlytranslatableitems 45 LessonsLearned 1 SampleHasadequatelydetaileddemographics Hasacceptablereturnratesandattritionrates e g addressestheinfluenceofnon respondentsanddrop outs 46 LessonsLearned 2 ResearchDesignDesignisconstructedsoastoruleoutalternativeexplanation Includesneededcontrolvariables Useslogicaltimingofmeasurement especiallyregardinglongitudinaldesigns 47 LessonsLearned 3 ProcedureExplainsproceduresinadequatedetail enoughtoallowareplication yetisreasonablysuccinct 48 LessonsLearned 4 MeasurementExplainsfullywhenexisting acceptedmeasuresarenotused Usesmeasuresthatarefreefrombias e g socialdesirability Presentsevidenceofconstructvalidity e g convergentanddiscriminantvalidity factoranalysis asneeded 49 LessonsLearned 5 Measurement conti Hasadequatetypesandlevelsofreliability e g internalconsistency inter rater 50 LessonsLearned 6 StatisticalAnalysisProperuseofdataanalyticapproachesandtechniques Useofminimallysufficientstatistics notundulyfancyandcomplex 51 LessonsLearned 7 Cross culturalMethodologyBefamiliarwithcross cultureresearchmethodology e g theuseofresearchinstruments SemanticequivalenceConceptualequivalenceScalingequivalence 52 AdditionalIssues Avoidscommonmethodvarianceorexplainwhyitis