社会研究方法双语.pptx

社会研究方法双语教学课件Chapter 1 introduction lWhatissocialresearch?WayofknowingsocialrealitybydirectexperiencelDefinitionlThedifferencesbetween“socialresearch”and“社会调查研究“Functions or purposesof social researchlDescription(描述)Amajorpurposeofmanysocialscientificstudiesistodescribesituationsandevents.Theresearcherobservesandthendescribeswhatwasobserved.Sincescientificobservationsiscarefulanddeliberate,however,scientificdescriptionsaretypicallymoreaccurateandprecisethancasualdescriptions.(P73)Examplesl北京汽车市场调查农村居民收入差距Explanation(解释)lThesecondgeneralpurposesofsocialscientificresearchistoexplainthings.Reportingthevotingintentionsofanelectorateisadescriptiveactivity,butreportingwhysomepeopleplantovoteforcandidateAandothersforcandidateBisanexplanatoryactivity.Reportingwhysomecitieshavehighercrimeratesthanothersisacaseofexplanation,butsimplyreportingthedifferentcrimeratesisacaseofdescription.Prediction(预测)lForexample,thegoalofregressionanalysisisfindouttherelationshipbetweentwoormorevariables.2.Types of research methodslObjetctivedimensionA.Census(普查).lAnenumeration(列举)ofthecharacteristicsofsomepopulation(总体).Acensusisoftensimilartoasurvey,withthedifferencethatthecensuscollectsdatafromallmembersofthepopulationwhilethesurveyislimitedtoasample.B.B.Sampling survey(抽样调查)lCarefulprobabilitysamplingprovidesagroupofrespondentswhosecharacteristicsmaybetakentoreflectthoseofthelargerpopulation,andcarefullyconstructedstandardizedquestionnairesprovidedatainthesameformfromallrespondents.C.Case study(个案研究)lTakeonlyseveralmembersfromthepopulationandstudythemindetail.Purposive Dimensiondescriptivestudies(描述性研究)explanatorystudies(解释性研究)time dimension.lCross-sectionalStudy(横剖研究).Astudythatisbasedonobservationsrepresentingasinglepointintime.lLongitudinalStudy（纵贯研究）.Astudydesigninvolvingthecollectionofdataatdifferentpointsintime,ascontrastedwithacross-sectionalstudy.lLongitudinalstudiesaredesignedtopermitobservationsoveranextendedperiod.Threetypesoflongitudinalstudiesshouldbenotedhere.lTrend studies(趋势研究)arethosethatstudychangeswithinsomegeneralpopulationovertime.ExampleswouldbeacomparisonofU.S.Censusovertime,showinggrowthinthenationalpopulation,oraseriesofGallupPollsduringthecourseofanelectioncampaign,showingtrendsintherelativestrengthsandstandingofdifferentcandidates.lCohort Studies(同期群研究)examinemorespecificsubpopulations(cohorts)astheychangeovertime.Typically,acohortisanagegroup,suchasthosepeoplebornduringthe1920s,butitcanalsobebasedonsomeothertimegrouping,suchaspeopleattendingcollegeduringtheVietnamWar,peoplewhogotmarriedin1964,andsoforth.lAnexampleofcohortstudywouldbeaseriesofnationalsurveys,conductedperhapsevery10years,tostudytheeconomicattitudesofthecohortbornduringthedepressionoftheearly1930s.asampleofpersons20-25yearsofagemightbesurveyedin1960,andanothersampleofthose40-45yearsofagein1970.Althoughthespecificsetofpeoplestudiedineachofthosesurveyswouldbedifferent,eachsamplewouldrepresentthesurvivorsofthecohortbornbetween1930and1935.lPanel Studies(定组研究定组研究,追踪研究追踪研究)aresimilartotrendandcohortstudiesexceptthatthesamesetofpeopleisstudiedeachtime.Oneexamplewouldbeavotingstudyinwhichthesamesampleofvoterswasinterviewedeverymonthduringanelectioncampaignandaskedforwhomtheywouldintendedtovote.Suchastudywouldmakeitpossibletoanalyzeoveralltrendsinvoterpreferencesfordifferentcandidates,butitwouldhavetheaddedadvantageofshowingtheprecisepatternsofpersistenceandchangeinintentions.lForexample,atrendstudythatshowedthatCandidatesAandBeachhadexactlyhalfofthevotersonSeptemberfirstandonOctoberfirstaswellcouldindicatethatnoneoftheelectoratehadchangedvotingplans,thatallofthevotershadchangedtheirintentions,orsomethingbetween.ApanelstudywouldeliminatethisconfusionbyshowingwhatkindsofvotersswitchedfromAtoBandwhatkindsswitchedfromBtoA,aswellasotherfacts.Procedures of social researchlPreparatorystage(准备阶段)lDatacollectionstage(收集资料阶段)lAnalysisstage(分析阶段)lSummarystage(总结阶段)Chapter 2 research designChapter 2 research designl1.Choosearesearchprojecta)Howtochoosearesearchprojectlb)Factorsrelatingwithresearchprojectchoicec)Principlesregardingresearchprojectchoicel2.PreliminaryExplorationla)Literaturereviewlb)Filedobservationl3.ResearchProjectDesigna)Researchhypothesisb)Researchplan2.1 Literature Reviewl1.PurposesofLiteratureReviewToavoidredundantresearchandtrytomakenewcontributionsToprovidebasesforhypothesisTotakeotherresearchesasreferencesforyourresearchplanl2.HowtoReviewLiteratureSnowballmethod:accordingtothereferencesandnotesoftheexistingliteraturetolookformorerelatedliteratureElectronicresources2.2 Field ObservationlMethods:colloquia(座谈会）,interview,refertoliteraturelPurpose1:forquestionnairedesignlExample:howtomeasurepeasantfamilyincomeintothreelevels:“high”,“medium”and“low”lPurpose2:forhypothesislExample:EconomicdevelopmentImplementationofelectoralsystemVillagersparticipation3.Research Project Designl3.1ResearchHypothesisHypothesis:Anexpectationaboutthenatureofthingsderivedfromatheory.Functionsofhypothesis:lToguildaresearchlTorelatetheoreticalconceptswithempiricaldatalToexplorenewtheoreticalknowledgePrinciplesformakinghypothesislConsistentwithexistingtheorieslConsistentwithconfirmedfactslCanbeverifiedbyexperience3.2 Research Project DesignlPurposeslPopulationandobjectslSamplingmethodslMethodsofdatacollectionanddataanalysislOrganizationlBudgetandfacilitieslWages,travellingexpenses,expenseforcopyingandprintinglFacilities:camera,taperecorder,computerlTimetableChapter 3 Samplingl3.1IntroductiontoSamplingl1.Thehistoryofsamplingl2.Samplingconceptsandterminologyl3.2ProbabilitySampling(随机抽样）l1.Simplerandomsampling(SRS)简单随机抽样l2.Systematicsampling系统抽样l3.Stratifiedsampling分层抽样l4.Clustersampling整群抽样l5.Multi-stagesampling多段抽样l3.3Non-ProbabilitySampling（非随机抽样）l1.Purposiveorjudgmentsampling立意抽样l2.Quotasampling配额抽样l3.Snowballsampling滚雪球抽样3.1 Introduction to Sampling3.1 Introduction to Samplingl1.ThehistoryofsamplingPoliticalpollingbyLiteracy Digest In1920,Digesteditorsmailedpostcardstopeopleinsixstates,askingthemwhotheywereplanningtovoteforinthepresidentialcampaignbetweenWarrenHardingandJamesCox.Nameswereselectedforthepollfromtelephone directories and automobile registration lists.Basedonthepostcardssentback,theDigestcorrectlypredictedthatHardingwouldbeelected.Inelectionsthatfollowed,themagazineexpandedthesizeofitspoll,andmadecorrectpredictionsin1924,1928,and1932.In1936,basedontwomillionrespondentsanswers,theDigestpredictedthatRepublicancandidateAlfLandonwouldget57%ballotsandincumbentPresidentFranklinRooseveltwouldgetonly43%.Twoweekslater,votersgaveRooseveltathirdterminofficebythelargestlandslideinhistory,with61percentofthevote.Theproblemlayinthesamplingframeused:telephonesubscribersandautomobileowners.Suchasamplingdesignselectedadisproportionatelywealthypeople,especiallycomingonthetailendoftheworsteconomicdepressioninthenationhistory.3.1 Introduction to Samplingl(continued)lIncontrasttotheLiteracy Digest,GeorgeGallupcorrectlypredictedthatRooseveltwouldbeatLandon.Gallupssuccessin1936hingedonhisuseofquotasampling.lQuotasamplingisbasedonaknowledgeofthecharacteristicsofthepopulationbeingsampled:whatproportionaremen,whatproportionwomen,whatproportionsareofvariousincomes,ages,etc.Peopleareselectedtomatchthepopulationcharacteristics.3.1 Introduction to Sampling2.Sampling Concepts and Terminology 2.Sampling Concepts and Terminology(1)(1)i.1.Element（研究单位）.An element is that unitaboutwhichinformationiscollectedandwhichprovidesthebasisofanalysis.Typically,in survey research,elements are people orcertaintypesofpeople.However,otherkindsofunitscanconstitutetheelementsforsocialresearch;families,socialclubs,orcorporationsmightbetheelementsofastudy.(Note:Elementsandunitsofanalysisareoftenthesameinagivenstudy,thoughtheformerreferstosampleselectionwhilethelatterreferstodataanalysis.)2.Sampling Concepts and Terminology(2)(2)l1.Population(总体).A population is the theoretically specifiedaggregationofstudyelements.For example,specifying the term“college students”would include aconsiderationoffull-timeandpart-timestudents,degreecandidatesandnon-degreecandidates,undergraduateandgraduatestudents,andsimilarissues.l2.Study Population（研究总体）.A study population is thataggregationofelementsfromwhichthesampleisactuallyselected.Asapracticalmatter,youareseldominapositiontoguaranteethateveryelement meeting the theoretical definitions laid down actually has achanceofbeingselectedinthesample.Evenwherelistsofelementsexistforsamplingpurposes,thelistsareusuallysomewhatincomplete.Somestudents are always omitted,inadvertently,from student roster.Sometelephonesubscribersrequestthattheirnamesandnumbersbeunlisted.Thestudypopulation,then,istheaggregationofelementsfromwhichthesampleisselected.2.Sampling Concepts and Terminology(3)(3)l3.SamplingUnit（抽样单位）.Asamplingunitisthatelementorsetofelementsconsideredforselectioninsomestageofsampling.In a simple,single-stage sample,the sampling units are the same as theelements.In more complex samples,however,different levels of samplingunitsmaybeemployed.Forexample,youmightselectasampleofcensusblocksinacity,thenselectasampleofhouseholdsontheselectedblocks,andfinallyselectasampleofadultsfromselectedhouseholds.l4.SamplingFrame（抽样框）.Asamplingframeistheactuallistofsamplingunitsfromwhichthesample,orsomestageofthesample,isselected.l5.ObservationUnit（观察单位）.Anobservationunit,orunitofdatacollection,isanelementsfromwhichinformationiscollected.Again,theunitofanalysisandunitofobservationareoftenthesametheindividual personbut that need not be the case.Thus the researcher mayinterview heads of households(the observation unit)to collect informationaboutallfamilymembersofthehouseholds(theunitsofanalysis).2.Sampling Concepts and Terminology(4)(4)l6.Variable（变量）.A variable is a set of mutually exclusiveattributes:sex,age,employmentstatus,andsoforth.l7.Parameter（参数值）.Aparameteristhesummarydescriptionofagivenvariableinapopulation.l8.Statistic（统计值）.Astatisticisthesummarydescriptionofagiven variable in a sample.Sample statistics are used to makeestimatesofpopulationparameters.l9.SamplingError（抽样误差）.Probabilitysamplingmethodsseldom,ifever,providestatisticsexactlyequaltotheparametersthattheyareusedtoestimate.Probabilitytheory,however,permitsustoestimatethedegreeoferrortobeexpectedforagivensampledesign.2.Sampling Concepts and Terminology(5)(5)l10.ConfidenceLevelsandConfidenceIntervals（显著性水平与置信区间）.We express the accuracy of our sample statistics interms of a level of confidence that the statistics fallwithinaspecifiedintervalfromtheparameter.Forexample,wemaysayweare95percentconfidentthatoursamplestatisticsarewithinplusorminus5percentagepointsofthepopulationparameter.3.2 Probability Sampling(1)(1)lSimple Random Sampling(简单随机抽样).Atype of probability sample in which the unitscomposingapopulationareassignednumbers,asetofrandomnumbersisthengenerated,andtheunits having those numbers are included in thesample.Although probability theory and thecalculationsitprovidesassumethisbasicsamplingmethod,itisseldomusedforpracticalreasons.3.2 Probability Sampling(2)(2)lSystematicSampling(系统抽样).Atypeofprobabilitysample in which every kth unit in a list is selected forinclusion in the sample:e.g.,every 25th student in thecollegedirectoryofstudents.Kiscomputedbydividingthesizeofthepopulationbythedesiredsamplesizeandiscalled the sampling interval.Within certain constraints,systematicsamplingisafunctionalequivalentofsimplerandomsamplingandusuallyeasiertodo.lSamplinginterval=populationsize/samplesizelsamplingratio=samplesize/populationsize 3.2 Probability Sampling(3)(3)Stratified sampling(分分层层抽抽样样):toorganizethepopulationintohomogeneoussubsets(withheterogeneity between subsets.)and to selecttheappropriatenumberofelementsfromeach.3.2 Probability Sampling(4)lClusterSampling(整群抽样).Amultistagesampleinwhichnaturalgroups(clusters)aresampledinitially,withthemembersofeachselectedgroupbeingsubsampledafterward.Forexample,youmightselectasampleofU.S.collegesanduniversitiesfromadirectory,getlistsofthestudentsatalltheselectedschools,thendrawsamplesofstudentsfromeach.3.3 Non-Probability Sampling(1)(1)lPurposiveorjudgmentalsampling(立意抽样).Atypeofnonprobabilitysamplinginwhichyouselecttheunitstobeobservedonthebasisofyourownjudgmentaboutwhichoneswillbethemostusefulorreprsentative.3.3 Non-Probability Sampling(2)(2)lQuotasampling(配额抽样).Atypeofnon-probability sampling in which units areselected into the sample on the basis ofprespecifiedcharacteristics,sothatthetotalsample will have the same distribution ofcharacteristicsasareassumedtoexistinthepopulationbeingstudied.3.3 Non-Probability Sampling(3)(3)lSnowballsampling(滚雪球抽样).Anon-probabilitysamplingmethodoftenemployed in filed research.Each personinterviewed may be asked to suggestadditionalpeopleforinterviewing.3.4 Factors influencing sample sizelA.populationsize样本规模lB.populationheterogeneity样本异质性lvariance(方差)lC.permitedsamplingerror允许抽样误差Chapter 4 Social Measurementl4.1OperationalizationandSocialMeasurementlA.OperationalizationofResearchProject（研究课题的操作化）lB.SocialMeasurement（社会测量）4.2LevelsofSocialMeasurementA.NominalMeasure（定类测量）B.OrdinalMeasure（定序测量）C.IntervalMeasure（定距测量）D.Ratiomeasure（定比测量）l4.3.ReliabilityandValiditylA.Reliability（信度）lB.Validity（效度）lC.Relationsbetweenreliabilityandvalidity4.1 Operationalization and Social MeasurementlA.OperationalizationofResearchProjecta.OperationaldefinitionofconceptlOperationaldefinitionadefinitionthatspellsoutpreciselyhowtheconceptwillbemeasured.Strictlyspeaking,anoperationaldefinitionisadescriptionofthe“operations”thatwillbeundertakeninmeasuringaconcept.b.ChoiceofindexeslExample:Economicdevelopment-annualincomepercapita;collectiveincomeIntelligence-Couplerelation-c.OperationalizationofhypothesislConcept:Industrialization-HumanrelationlIndex:industrialoutput-timesvisitingeachotherlphonesubscribersB.Social MeasurementlConceptualizationllNominaldefinitionllOperationaldefinitionlmeasurementsintherealworldlDefinition:in order to understand the nature,characteristics andconditionsoftheobjects,weallocatesomenumbersorsymbolstotheobjectsaccordingtosomeregulations.Thisprocessiscalledsocialmeasurement.lThreeelementsofsocialmeasurementObjectsNumberorsymbolsregulations4.2 Levels of Social MeasuremntlA.NominalMeasurelVariableswhoseattributeshaveonlythecharacteristicsofexhaustivenessandmutualexclusivenessarenominalvariables.Examplesofthesewouldbesex,religiousaffiliation,politicalpartyaffiliation,birthplace,collegemajor,andhaircolor.lB.OrdinalMeasurelVariableswhoseattributesmaybelogicallyrank-orderedareordinalmeasures.Thedifferentattributesrepresentrelativelymoreorlessofthevariable.Variables of this type are social class,conservatism,alienation,prejudice,andthelike.c.Interval MeasurelFor the attributes composing some variables,theactualdistanceseparatingthoseattributesdoeshavemeaning.Suchvariablesareintervalmeasures.Forthese,thelogicaldistancebetweenattributescanbeexpressedinmeaningfulstandardintervals.A physical science example would be the Fahrenheit or Celsiustemperaturescale.Thedifference,ordistance,between80degreesand90 degrees in the same that between 40 degrees and 50 degrees.However,8