市場調(diào)查方法第九章課件

2024/5/13市場調(diào)查方法第九章LEARNINGOUTCOMESCreateanexperimental,independentvariablethroughavalidexperimentalmanipulationofitsvalueUnderstandandminimizethesystematicexperimentalerrorKnowwaysofminimizingexperimentaldemandcharacteristicsAvoidunethicalexperimentalpracticesWeighthetrade-offbetweeninternalandexternalvalidityRecognizetheappropriateusesoftestmarketingAfterstudyingthischapter,youshouldbeableto?2007Thomson/South-Western.Allrightsreserved.TheNatureofExperimentsExperimentDefinedAresearchinvestigationinwhichconditionsarecontrolled.ExperimentalResearchAllowsaresearchertocontroltheresearchsituationsothatcausalrelationshipsamongvariablesmaybeevaluated.Independentvariablesareexpectedtodeterminetheoutcomesofinterest.Dependentvariablesaretheoutcomesofinteresttotheresearcherandthedecisionmakers.?2007Thomson/South-Western.Allrightsreserved.BasicIssuesinExperimentalDesignManipulationoftheindependentvariableSelectionandassignmentofexperimentalsubjectsControloverextraneousvariablesSelectionandmeasurementofthedependentvariableExperimentalDesign?2007Thomson/South-Western.Allrightsreserved.IssuesinExperimentalDesignManipulationoftheIndependentVariableIndependentvariable:avariablewithvaluesthatcanbemanipulated,oraltered,independentlyofanyothervariable.ExperimentalTreatmentThetermreferringtothewayanexperimentalvariableismanipulated.Experimentalgroup:agroupofsubjectstowhomanexperimentaltreatmentisadministered.Controlgroup:agroupofsubjectstowhomnoexperimentaltreatmentisadministered.?2007Thomson/South-Western.Allrightsreserved.ExperimentalDesign(cont’d)SelectionandMeasurementoftheDependentVariableDependentvariable:thecriterionbywhichtheresultsofanexperimentarejudged;avariableexpectedtobedependentontheexperimenter’smanipulationoftheindependentvariable.Selectingdependentvariablesthatarerelevantandtrulyrepresentanoutcomeofinterestiscrucial.Choosingtherightdependentvariableispartoftheproblemdefinitionprocess—thoroughproblemdefinitionwillhelptheresearcherselectthemostimportantdependentvariable(s)whoseresultswillhelpmanagersindecisionmaking.?2007Thomson/South-Western.Allrightsreserved.ExperimentalDesign(cont’d)SelectionandAssignmentofTestUnitsTestunits:thesubjectsorentitieswhoseresponsestotreatmentaremeasuredorobserved.SampleSelectionAndRandomSamplingErrorsSystematicornonsamplingerrorSubjectselection,experimentaldesign,andunrecognizedextraneousvariablesOvercomingsamplingerrorsRandomizationMatchingRepeatedmeasuresControloverextraneousvariables?2007Thomson/South-Western.Allrightsreserved.ExperimentalDesign(cont’d)SamplingErrors(cont’d)ExperimentalConfoundWhenthereisanalternativeexplanationbeyondtheexperimentalvariablesforanyobserveddifferencesinthedependentvariable.Onceapotentialconfoundisidentified,thevalidityoftheexperimentisseverelyquestioned.Sources:SamplingerrorSystematicerrorLater-identifiedextraneousvariablesCarefulexperimentaldesigncanreducethelikelihoodofconfounds.?2007Thomson/South-Western.Allrightsreserved.DemandCharacteristicsDemandCharacteristicAnexperimentaldesignelementorprocedurethatunintentionallyprovidessubjectswithhintsabouttheresearchhypothesis.DemandEffectOccurswhendemandcharacteristicsactuallyaffectthedependentvariable.ExperimenterBiasTheinfluenceofthepresence,actions,orcommentsofanexperimenteronsubjects’behavior.?2007Thomson/South-Western.Allrightsreserved.DemandCharacteristics(cont’d)ReducingDemandCharacteristicsUseanexperimentaldisguise.Isolateexperimentalsubjects.Usea“blind”experimentaladministrator.Administeronlyonetreatmentleveltoeachsubject.?2007Thomson/South-Western.Allrightsreserved.EXHIBIT9.1

BySmilingorLookingSolemn,ExperimentersCanModifySubjects’Behavior?2007Thomson/South-Western.Allrightsreserved.EstablishingControlConstancyofConditionsSubjectsinallexperimentalgroupsareexposedtoidenticalconditionsexceptforthedifferingexperimentaltreatments.CounterbalancingAttemptstoeliminatetheconfoundingeffectsoforderofpresentationbyvaryingtheorderofpresentation(exposure)oftreatmentstosubjectgroups.?2007Thomson/South-Western.Allrightsreserved.EthicalIssuesinExperimentationDebriefingexperimentalsubjectsCommunicatingthepurposeoftheexperimentExplainingtheresearcher’shypothesesaboutthenatureofconsumerbehaviorAttemptstointerferewithacompetitor’stest-marketingeffortsSuchactsaschangingpricesorincreasingadvertisingtoinfluence(confound)competitors’test-marketingresultsareethicallyquestionable.?2007Thomson/South-Western.Allrightsreserved.FundamentalQuestionsinExperimentationLaboratoryExperimentAsituationinwhichtheresearcherhasmorecompletecontrolovertheresearchsettingandextraneousvariables.FieldExperimentsResearchprojectsinvolvingexperimentalmanipulationsthatareimplementedinanaturalenvironment.?2007Thomson/South-Western.Allrightsreserved.EXHIBIT9.5

TheArtificialityofLaboratoryversusFieldExperiments?2007Thomson/South-Western.Allrightsreserved.IssuesofExperimentalValidityInternalValidityTheextentthatanexperimentalvariableistrulyresponsibleforanyvarianceinthedependentvariable.Didtheexperimentanswerthequestionofcausaleffect?Didthemanipulationdowhatitwassupposed(predicted)todo?ManipulationChecksAvaliditytestofanexperimentalmanipulationtomakesurethatthemanipulationdoesproducedifferencesintheindependentvariable.?2007Thomson/South-Western.Allrightsreserved.ExtraneousVariablesAffectingInternalValidityInternalValidityMaturationTestingInstrumentationSelectionMortalityHistory?2007Thomson/South-Western.Allrightsreserved.EffectsofExtraneousVariablesonValidityHistoryEffectOccurswhensomechangeotherthantheexperimentaltreatmentoccursduringthecourseofanexperimentthataffectsthedependentvariable.CohortEffectAchangeinthedependentvariablethatoccursbecausemembersofoneexperimentalgroupexperienceddifferenthistoricalsituationsthanmembersofotherexperimentalgroups.?2007Thomson/South-Western.Allrightsreserved.EffectsofExtraneousVariables…(cont’d)MaturationEffectsEffectsthatareafunctionoftimeandthenaturallyoccurringeventsthatcoincidewithgrowthandexperience.TestingEffectsAnuisanceeffectoccurringwhentheinitialmeasurementortestalertsorprimessubjectsinawaythataffectstheirresponsetotheexperimentaltreatments.?2007Thomson/South-Western.Allrightsreserved.EffectsofExtraneousVariables…(cont’d)InstrumentationEffectAchangeinthewordingofquestions,achangeininterviewers,orachangeinotherprocedurescausesachangeinthedependentvariable.SelectionEffectSamplebiasthatresultsfromdifferentialselectionofrespondentsforthecomparisongroups,orasampleselectionerror.MortalityEffect(SampleAttrition)Occurswhensomesubjectswithdrawfromtheexperimentbeforeitiscompleted.?2007Thomson/South-Western.Allrightsreserved.IssuesofExperimentalValidity(cont’d)ExternalValidityTheaccuracywithwhichexperimentalresultscanbegeneralizedbeyondtheexperimentalsubjects.Studentsurrogates:Atypical?Trade-OffsBetweenInternalandExternalValidityArtificiallaboratoryexperimentsusuallyarehighininternalvalidity,whilenaturalisticfieldexperimentsgenerallyhavelessinternalvalidity,butgreaterexternalvalidity.?2007Thomson/South-Western.Allrightsreserved.ClassificationofExperimentalDesignsBasicExperimentalDesignAnexperimentaldesigninwhichonlyonevariableismanipulated.DiagrammingExperimentalDesigns:Symbols?2007Thomson/South-Western.Allrightsreserved.ExamplesofQuasi-ExperimentalDesignsQuasi-experimentalDesignsExperimentaldesignsthatdonotinvolverandomallocationofsubjectstotreatmentcombinations.OneShotDesign(AfterOnly): X O1OneGroupPretest–Posttest: O1 X O2StaticGroupDesign: Experimental X O1

Control O2?2007Thomson/South-Western.Allrightsreserved.AlternativeExperimentalDesignsPretest–PosttestControlGroupDesign

(Before–AfterwithControl)Experimental RO1XO2Control RO3XO4

PosttestOnlyControlGroup

(After-OnlywithControl)Experimental RXO1Control RO2CompromiseDesigns?2007Thomson/South-Western.Allrightsreserved.EXHIBIT9.3

ProductPreferenceMeasureinanExperiment?2007Thomson/South-Western.Allrightsreserved.Test-MarketingTestMarketingAnexperimentalprocedurethatprovidesanopportunitytotestanewproductoranewmarketingplanunderrealisticmarketconditionstomeasuresalesorprofitpotential.ControlledExperimentationNotjusttryingsomethingou