市场调查方法(英文版)第九章

PowerPointPresentationbyCharlieCookTheUniversityofWestAlabamaWilliamG.ZikmundBarryJ.BabinPart2DesigningResearchStudies3rdEdition©2007Thomson/South-Western.Allrightsreserved.Chapter9ExperimentalResearch:AnOverview©2007Thomson/South-Western.Allrightsreserved.9–2LEARNINGOUTCOMES1.Createanexperimental,independentvariablethroughavalidexperimentalmanipulationofitsvalue2.Understandandminimizethesystematicexperimentalerror3.Knowwaysofminimizingexperimentaldemandcharacteristics4.Avoidunethicalexperimentalpractices5.Weighthetrade-offbetweeninternalandexternalvalidity6.RecognizetheappropriateusesoftestmarketingAfterstudyingthischapter,youshouldbeableto©2007Thomson/South-Western.Allrightsreserved.9–3TheNatureofExperiments•ExperimentDefinedAresearchinvestigationinwhichconditionsarecontrolled.•ExperimentalResearchAllowsaresearchertocontroltheresearchsituationsothatcausalrelationshipsamongvariablesmaybeevaluated.Independentvariablesareexpectedtodeterminetheoutcomesofinterest.Dependentvariablesaretheoutcomesofinteresttotheresearcherandthedecisionmakers.©2007Thomson/South-Western.Allrightsreserved.9–4BasicIssuesinExperimentalDesignManipulationoftheindependentvariableSelectionandassignmentofexperimentalsubjectsControloverextraneousvariablesSelectionandmeasurementofthedependentvariableExperimentalDesign©2007Thomson/South-Western.Allrightsreserved.9–5IssuesinExperimentalDesign•ManipulationoftheIndependentVariableIndependentvariable:avariablewithvaluesthatcanbemanipulated,oraltered,independentlyofanyothervariable.•ExperimentalTreatmentThetermreferringtothewayanexperimentalvariableismanipulated.Experimentalgroup:agroupofsubjectstowhomanexperimentaltreatmentisadministered.Controlgroup:agroupofsubjectstowhomnoexperimentaltreatmentisadministered.©2007Thomson/South-Western.Allrightsreserved.9–6ExperimentalDesign(cont’d)•SelectionandMeasurementoftheDependentVariableDependentvariable:thecriterionbywhichtheresultsofanexperimentarejudged;avariableexpectedtobedependentontheexperimenter’smanipulationoftheindependentvariable.Selectingdependentvariablesthatarerelevantandtrulyrepresentanoutcomeofinterestiscrucial.Choosingtherightdependentvariableispartoftheproblemdefinitionprocess—thoroughproblemdefinitionwillhelptheresearcherselectthemostimportantdependentvariable(s)whoseresultswillhelpmanagersindecisionmaking.©2007Thomson/South-Western.Allrightsreserved.9–7ExperimentalDesign(cont’d)•SelectionandAssignmentofTestUnitsTestunits:thesubjectsorentitieswhoseresponsestotreatmentaremeasuredorobserved.•SampleSelectionAndRandomSamplingErrorsSystematicornonsamplingerrorSubjectselection,experimentaldesign,andunrecognizedextraneousvariablesOvercomingsamplingerrorsRandomizationMatchingRepeatedmeasuresControloverextraneousvariables©2007Thomson/South-Western.Allrightsreserved.9–8ExperimentalDesign(cont’d)•SamplingErrors(cont’d)ExperimentalConfoundWhenthereisanalternativeexplanationbeyondtheexperimentalvariablesforanyobserveddifferencesinthedependentvariable.Onceapotentialconfoundisidentified,thevalidityoftheexperimentisseverelyquestioned.Sources:–Samplingerror–Systematicerror–Later-identifiedextraneousvariablesCarefulexperimentaldesigncanreducethelikelihoodofconfounds.©2007Thomson/South-Western.Allrightsreserved.9–9DemandCharacteristics•DemandCharacteristicAnexperimentaldesignelementorprocedurethatunintentionallyprovidessubjectswithhintsabouttheresearchhypothesis.•DemandEffectOccurswhendemandcharacteristicsactuallyaffectthedependentvariable.•ExperimenterBiasTheinfluenceofthepresence,actions,orcommentsofanexperimenteronsubjects’behavior.©2007Thomson/South-Western.Allrightsreserved.9–10DemandCharacteristics(cont’d)•ReducingDemandCharacteristics1.Useanexperimentaldisguise.2.Isolateexperimentalsubjects.3.Usea“blind”experimentaladministrator.4.Administeronlyonetreatmentleveltoeachsubject.©2007Thomson/South-Western.Allrightsreserved.9–11EXHIBIT9.1BySmilingorLookingSolemn,ExperimentersCanModifySubjects’Behavior©2007Thomson/South-Western.Allrightsreserved.9–12EstablishingControl•ConstancyofConditionsSubjectsinallexperimentalgroupsareexposedtoidenticalconditionsexceptforthedifferingexperimentaltreatments.•CounterbalancingAttemptstoeliminatetheconfoundingeffectsoforderofpresentationbyvaryingtheorderofpresentation(exposure)oftreatmentstosubjectgroups.©2007Thomson/South-Western.Allrightsreserved.9–13EthicalIssuesinExperimentation•DebriefingexperimentalsubjectsCommunicatingthepurposeoftheexperimentExplainingtheresearcher’shypothesesaboutthenatureofconsumerbehavior•Attemptstointerferewithacompetitor’stest-marketingeffortsSuchactsaschangingpricesorincreasingadvertisingtoinfluence(confound)competitors’test-marketingresultsareethicallyquestionable.©2007Thomson/South-Western.Allrightsreserved.9–14FundamentalQuestionsinExperimentation•LaboratoryExperimentAsituationinwhichtheresearcherhasmorecompletecontrolovertheresearchsettingandextraneousvariables.•FieldExperimentsResearchprojectsinvolvingexperimentalmanipulationsthatareimplementedinanatur