清华MBA系列课件消费行为学Perceptual Map of Beers Example.docx

清华MBA系列课件消费行为学Perceptual Map of Beers Example.docx

《清华MBA系列课件消费行为学Perceptual Map of Beers Example.docx》由会员分享，可在线阅读，更多相关《清华MBA系列课件消费行为学Perceptual Map of Beers Example.docx（12页珍藏版）》请在冰点文库上搜索。

清华MBA系列课件消费行为学PerceptualMapofBeersExample

PerceptualMapofForeignBeersinthePRC:

AnExample

Background

BecauseofitspotentialtosurpasstheUSasthelargestbeermarketintheworld（intermsofvolume）bytheyear2000,thePeople’sRepublicofChina（PRC）isaveryattractivemarketforeverymultinationalbeercompany.SeveralmultinationalbeercompaniesarecurrentlyoperationinChina.Somecompanieshavebeenquitesuccessfulwhilesomearestillstrugglingtogainfootholdinthismarket.Theexampledemonstratedinthispaperisanactualresearchthatwasconductedforamultinationalbeercompany.TheobjectiveofthisstudyistoinvestigatethebeerconsumptionbehaviorofChineseconsumersaswellastoinvestigatetherelativepositionsofseveralleadingforeignbeersintheseconsumers’mind.Onlytheprocedure,findings,andmanagerialimplicationsoftheperceptualmappingoftheresearchwillbedemonstratedanddiscussedhere.

Researchprocedure

Step1:

ExploratoryResearch

TheobjectiveoftheexploratoryresearchwastoidentifyproductattributesthatareimportanttoChinesebeerdrinkers.Theproductattributesidentifiedwerethenusedtodesignthequestionnaireforthemainstudy.Intheexploratoryresearch,afocusgroupinterviewofeightChineseconsumerswasconductedtoidentifyattributesorcharacteristicsofbeerthatareimportanttothem.Altogether,tenproductattributeswereidentified:

packagedesign,taste,color,levelofalcohol,varietyofsizes,price,countryoforigin,availability,advertisingandreputation.

Step2:

QuestionnaireDesign

Basedontheproductattributesidentifiedinthefocusgroupinterview,thequestionnaireforthemainstudywasdesigned.（Seethequestionnairedistributedearlier.）Althoughfivebrandsofforeignbeerswereincludedinthestudy:

Carlsberg,Foster,Guinness,Heineken,andSanMiguel.Notethatthefirsttenquestionsforeachbrandisthequestionformeasuringtheoverallattitudetowardthebrandandtheintentiontochoosethebrand.Theattituderatingquestionisneededfortheidentificationofthe“idealline”onwhichtheidealpositionislocatedbyusingtheattitudetowardthebrandasthedependentvariableandhefactorscoresastheindependentvariables.Thebetacoefficientsoftheindependentvariables（whichrepresenttherelativeimportantofeachfactorscoreininfluencingtheattitude）willbeusedtocalculatetheanglefortheidealline.Theintentionratingisneededtomeasuretherelationshipbetweentheattitudetowardthebrandandtheintentiontobuybyusingtheintentionasthedependentvariableandtheattitudeastheindependentvariable.Thisregressionanalysisisneededtoseewhetherornotattitudetowardthebrandisagoodpredictorofintentiontobuythebrand.

Step3:

TheMainStudy

ResearchDesign.Theresearchdesignusedinthemainstudywasasamplesurveybyface-to-faceinterview.

SampleandSamplingProcedure.ThedatawerecollectedfromChinesebeerdrinkersinGuangzhou,Beijing,andShanghaibyconvenientsamplingatrestaurantsandbars/pubs.Altogether,67completeanduseablequestionnaireswereobtained.79%ofthesamplesweremale.Formartialstatus,56.7%weremarried.Intermsofeducation,88.1%hadcollegeeducation.Foroccupation,44.8%wereprofessional,43.3%werewhitecollars.

DataCollection.Thedatawerecollectedbypersonalinterviewatrestaurantandbars/pubs.Theinterviewersreadthequestionsinthequestionnairetotherespondentandrecordedtheanswersobtainedinthequestionnaire.

Step4:

DataAnalysis

DataCodingandDataEntry.Afterthedatacollectionwascompleted,thequestionnaireswerecodedandenteredintoSPSSforWindowsVersion12.0.1.Sinceeachrespondentevaluatefivebrandsofbeers,thedatafromeachrespondentarerepresentedbyfive“cards”orrowsofdatainput.Exhibit1showsthesampleofthedatainputfile.

StatisticalAnalysis.TocreateperceptualmapsfortheforeignbeerinChina,thecoordinatesforeachbrandofbeersareneeded.Forperceptualmappingbyfactoranalysis,thecoordinatesforthebrandsareobtainedbyaveragingthefactorscoreofeachbrandacrosstherespondents.Thefactorscorescanbeobtainedbythefollowingprocedure:

a.Fromthemainmenu,select“Analyze”thenselect“DataReduction”thenselect“FactorAnalysis”（seeExhibit2）.Thefactoranalysiswindowwillbeopenatthispoint.Thefactoranalysiswindowconsistsofaboxshowingallvariablesinthedatasetandthe“Variables”box.Therearealsofivebottomsforoptionsincluding“Descriptive”,“Extraction”,“Rotation”,“Scores”,and“Options”.

b.Selectvariablesthatrepresentproductattributes（i.e.,X1toX10）fromtheboxshowingthelistofvariablesintothe“Variables”box（seeExhibit3）.Thisstepidentifiesvariablestobeanalyzedbyfactoranalysis.

c.Open“Rotation”boxandselect“Varimax”rotation（seeExhibit4）,andthenclick“Continue”.Thisstepspecifiesthetyperotationtobeused.

d.Open“Scores”boxandselect“Saveasvariables”andfor“Method”select“regression”（seeExhibits5）,andthenclick“Continue”.Thisstepwillcreatefactorscoresbyregressionmethodandaddthemtothedatasetasnewvariables.

e.For“Extraction”usedefaultsfor“Method”（i.e.,Principalcomponent）,“Analyze”（i.e.,Correlationmatrix）,“Display”（i.e.,Unrotatedfactorsolution）and“Extract”（i.e.,Eigenvaluesover1）,andthenclick“Continue”.Thisstepwillcommandtheprogramtocomputecorrelationmatrixforthevariablesfromtherawdata.Thiscorrelationmatrixwillbe,inturn,analyzedtoobtainafactormatrix.Principalcomponentanalysisisspecifiedasthemethodofchoiceinextractingthefactors.Forthenumberoffactorstobeextracted,thelatentrootoreigenvaluecriterion（greaterthan1）willbeused.The（unrotated）componentmatrixwillalsobegivenasanoutput.

f.For“Descriptive,”usedefaultfor“Statistics”（i.e.,Initialsolution）andthenclick“Continue”.

g.For“Options”usedefaultfor“MissingValues”（i.e.,Excludedcaseslistwise）.For“CoefficientDisplayFormat”,select“Sortbysize”（seeExhibit6）,andthenclick“Continue”.Thisstepwillcreatefactormatricesofwhichvariablesaresortedbysizeindescendingorder.

h.Atthispoint,wearebackatthefactoranalysiswindow.Click“OK”toruntheprogram.Theoutputwillbeshownintheoutputwindow.

ResultsandInterpretation

Basedontheabovefactoranalysisprocedure,thefollowingoutputareobtained:

Communalities,TotalVarianceExplained,ComponentMatrix（orFactorLoadingsTable）,RotatedComponentMatrix（orRotatedFactorLoadingTable）,andComponentTransformationMatrix.TherelevantoutputsforperceptualmappingaretheTotalVarianceExplainedandtheRotatedComponent（orFactor）Matrix.Seetheoutputdistributedearlierfordetail.

Step1:

DeterminationoftheNumberofFactorstoRetain

Thefirststepintheanalysisofresultsistodeterminethenumberoffactorstoberetainedforfurtheranalysis.TheTotalVarianceExplainedintheoutputcontainstheinformationregardingthetenpossiblefactorsandtheirrelativeexplanatorypowerasexpressedbytheireigenvalues.Inadditiontoassessingtheimportanceofeachfactor,eigenvaluesarealsousedtodeterminethenumberoffactorstoberetained.Accordingtothelatentrootoreigenvaluecriterion,twofactorswillberetainedbecauseonlycomponent1（factor1）andcomponent2（factor2）havelatentrootsoreigenvaluesgreaterthan1（seeExhibit7）.

Step2:

NamingtheFactors

Thenamingprocessinvolvessubstantiveinterpretationofthepatternoffactorloadingsforthevariables,includingtheirsigns,inanefforttonameeachofthefactors.Beforeinterpretation,aminimumacceptablelevelofsignificanceforfactorloadingmustbeselected.Allsignificantfactorloadingstypicallyareusedintheinterpretationprocess.Butvariableswithhigherloadingsinfluencetoagreaterextentthenameorlabelselectedtorepresentafactor.

TheVarimaxrotatedcomponentanalysisfactormatrixwhichcontainsfactorloadingsisshownintheRotatedFactorMatrix（orfactorloadingtable）intheoutput（Exhibit8）.Forthesamplesizeof67,factorloadingshavetobe±0.65orabovetobeconsideredsignificant.Fromthematrix,it’sclearthattaste,color,packagedesign,reputation,andcountryoforiginloadsignificantlyonFactor1,andprice,availability,levelofalcohol,levelofadvertising,andvarietyofsizesloadsignificantlyonFactor2.Noneofthevariablesloadssignificantlyonmorethanonefactor.

ForFactor1,allthefivevariables（taste,color,packagedesign,reputation,andcountryoforigin）representdifferentdimensionsoftheproduct.ItisreasonabletonameFactor1asproductquality.ForFactor2,thefivevariables（price,availability,levelofalcohol,levelofadvertising,andvarietyofsizes）representtheotherthreecomponentsofthemarketingmix.Itisreasonabletonamethefactormarketingactivities.

Notethatthenamingprocessisbasedonthesubjectiveopinionoftheresearcher.Forthisreason,theprocessofnamingfactorsissubjecttoconsiderablecriticism.Butifalogicalnamecanbeassignedthatrepresentstheunderlyingnatureofthefactors,itusuallyfacilitatesthepresentationandunderstandingofthefactorsolutionandthereforeisajustifiableprocedure.

Step3:

IdentificationofPositionsoftheBrands

Sincetwofactorsareobtainedinthefactoranalysisprocess,twofactorscoresaregeneratedandaddedtothedataset.Factorscoresarecompositemeasureofeachfactorcomputedforeachsubjectforeachbrand.Inthisstudy,twofactorscorescanbeusedtoreplacetheoriginalsetoftenvariable