【复旦大学首批FIST项目传播学研究方法讲义】1-实证研究原理与设计(彭泰权)

1、实证研究原理与设计WinsonPeng彭泰权复旦大学2013年FIST课程·传播研究方法OutlineTheFoundationofSocialScienceSomeDialecticsofSocialResearchoDeductivevs.InductiveoIdiographicvs.NomotheticoQualitativevs.QuantitativeElementsinResearchDesignoResearchPurposesoUnitsofAnalysisoTheTimeDimension2LookingforRealityExperientialRealityVS.AgreementRealityKnowledgefromAgreementRealityoAssertionsmustbebothlogicalandempirical.oEpistemology–Thescienceofknowing.•Methodology–Asubfieldofepistemology.OrdinaryHumanInquiryoHumanswanttopredictthefuture.Why?•Werecognizethatthefutureiscausedinpartbythepresent.•Causeandeffectpatternsareprobabilisticinnature.oPredictionversusUnderstandingoTraditionoAuthorityErrorsinInquiryandSomeSolutionsoInaccurateObservations•Measurementdevicesofferaccuracy.oOvergeneralizations•Largeandrepresentativesamplesareasafeguardagainstovergeneralization.•Replication–Repeatingaresearchstudytotestandeitherconfirmorquestionthefindingsofanearlierstudy.oSelectiveObservations•Avoidlookingfor“deviant”cases.oIllogicalReasoning•“Gambler’sfallacy”What’sReallyReal?oThePremodernView–Thingsareastheyseem.oTheModernView–Acknowledgementofhumansubjectivity.oThePostmodernView–Thereisnoobjectivereality.TheFoundationsofSocialScienceThefoundationsofsocialsciencearelogicandobservation.Theory,NotPhilosophyorBeliefoSocialtheoryhastodowithwhatis,notwithwhatshouldbe.oTheory–Asystematicexplanationfortheobservationsthatrelatetoaparticularaspectoflife.oSocialsciencecanhelpusknowwhatisandwhy.SocialRegularitiesoExceptions?oSocialregularitiesrepresentprobabilisticpatterns.Aggregates,NotIndividualsoThecollectiveactionsandsituationsofmanyindividuals.oFocusofsocialscienceistoexplainwhyaggregatedpatternsofbehaviorareregularevenwhenindividualschangeovertime.ConceptsandVariablesoVariables–Logicalgroupingsofattributes.oAttributes–Characteristicsorqualitiesthatdescribeanobject.oIndependentVariable–Avariablewithvaluesthatarenotproblematicalinananalysisbutaretakenassimplygiven.oDependentVariable–Avariableassumedtodependonorbecausedbyanother.SomeDialecticsofSocialResearchInductiveandDeductiveTheoryoInduction(归纳)–Thelogicalmodelinwhichgeneralprinciplesaredevelopedfromspecificobservations.oDeduction(演绎)–Thelogicalmodelinwhichspecificexpectationsofhypothesesaredevelopedonthebasisofgeneralprinciples.DeductiveTheoryConstruction1.Specifythetopic.2.Specifytherangeofphenomenayourtheoryaddresses.3.Identifyandspecifyyourmajorconceptsandvariables.4.Findoutwhatisknownabouttherelationshipsamongthosevariables.5.Reasonlogicallyfromthosepropositionstothespecifictopicyouareexamining.InductiveTheoryConstructionObservingaspectsofsociallifeandseekingtodiscoverpatternsthatmaypointtorelativelyuniversalprinciples.oGroundedTheoryoFieldResearchTheLinksBetweenTheoryandResearchDeductiveModel–researchisusedtotesttheories.InductiveModel–theoriesaredevelopedfromanalysisofdata.QualitativeandQuantitativeDataoQualitativeData–non-numericaldataoQuantitativeData–numericaldataQuantitativeVS.QualitativeMethodsinCommunicationResearch(Trumbo,2004)58%57%35%41%7%2%0%20%40%60%80%100%1965-19891990-2000PercentTimeframeMixedQualitativeQuantitativeResearchMethodsinInternetStudies(Pengetal.,2013)IdiographicandNomotheticExplanationoIdiographic–Anapproachtoexplanationinwhichweseektoexhausttheidiosyncraticcausesofaparticularconditionofevent.oNomothetic–Anapproachtoexplanationinwhichweseektoidentifyafewcausalfactorsthatgenerallyimpactaclassofconditionsofevent.27TheLogicofNomotheticExplanationGoal:tofindafewfactorsthatcanaccountformanyofthevariationsinagivenphenomenon.Example:LegalizationofMarijuanaIdiographicApproachNomotheticApproachInformationfromparents,teachers,clergyPreviousexperiencesOthers?PoliticalorientationOthers?CriteriaforNomotheticCausality1.Thevariablesmustbecorrelated•Correlation–Anempiricalrelationshipbetweentwovariablessuchthatchangesinoneareassociatedwithchangesintheother,orparticularattributesinoneareassociatedwithparticularattributesintheother.2.Thecausetakesplacebeforetheeffect3.Thevariablesarenonspurious•SpuriousRelationship–AcoincidentalstatisticalcorrelationbetweentwovariablesshowntobecausedbysomethirdvariableNomotheticCausalAnalysisandHypothesisTestingoHypothesesarerequiredinnomotheticresearch.oTotestahypothesis:•Specifyvariablesyouthinkarerelated•Specifymeasurementofvariables•Hypothesizecorrelation,strengthofrelationship,statisticalsignificance•SpecifytestsforspuriousnessFalseCriteriaforNomotheticCausalityoCompleteCausationoExceptionalCasesoMajorityofCasesNecessaryandSufficientCausesAnecessarycauserepresentsaconditionthatmustbepresentfortheeffecttofollow.Asufficientcauserepresentsaconditionthat,ifpresent,guaranteestheeffectinquestion.Mostsatisfyingoutcomesinresearchincludebothnecessaryandsufficientcauses.NecessaryCauseSufficientCauseElementsinSocialResearch37ElementsofResearchDesignExamplesofResearchStrategiesoExploration,Description,orExplanation?oUnitofanalysis?oDimensionsoftimerelevant?38PurposesofSocialResearchExploratory:Mappingoutatopicthatmaywarrantfurtherstudylate