1976年综合社会调查(General Social Survey, 1976)_数据挖掘_科研数据

1976年综合社会调查(GeneralSocialSurvey,1976)数据介绍：TheGeneralSocialSurveys(GSS)havebeenconductedbytheNationalOpinionResearchCenterannuallysince1972,exceptfortheyears1979,1981,and1992(asupplementwasaddedin1992),andbienniallybeginningin1994.TheGSSaredesignedaspartofaprogramofsocialindicatorresearch,replicatingquestionnaireitemsandwordinginordertofacilitatetime-trendstudies.Itemsonreligioninthe1976GSSincludereligiouspreference,churchattendance,beliefsaboutlifeafterdeath,andattitudestowardorganizedreligion.关键词：综合社会调查,每年,全国民意调查,社会指标研究,宗教,genralsocialsurvey,NationalOpinionResearch,annually,socialindicatorresearch,religion,数据格式：TEXT数据详细介绍：GeneralSocialSurvey,1976TheGeneralSocialSurveys(GSS)havebeenconductedbytheNationalOpinionResearchCenterannuallysince1972,exceptfortheyears1979,1981,and1992(asupplementwasaddedin1992),andbienniallybeginningin1994.TheGSSaredesignedaspartofaprogramofsocialindicatorresearch,replicatingquestionnaireitemsandwordinginordertofacilitatetime-trendstudies.Itemsonreligioninthe1976GSSincludereligiouspreference,churchattendance,beliefsaboutlifeafterdeath,andattitudestowardorganizedreligion.DataFileCases:1,499Variables:369WeightVariable:DataCollectionDateCollected:February,March,April,andMayof1976FundedByNationalScienceFoundationCollectionProceduresSurveydataweretakenutilizingpersonalinterviews.SamplingProceduresIntheoriginalNationalScienceFoundationgrant,supportwasgivenforamodifiedprobabilitysample.Samplesforthe1972through1974surveysfollowedthisdesign.Thismodifiedprobabilitydesign,describedbelow,introducesthequotaelementattheblocklevel.TheNSFrenewalgrant,awardedforthe1975--1977surveys,providedfundsforafullprobabilitysampledesign,adesignwhichisacknowledgedtobesuperior.Thus,havingthewherewithaltoshifttoafullprobabilitysamplewithpredesignatedrespondents,the1975and1976studieswereconductedwithatransitionalsampledesign,viz.,one-halffullprobabilityandone-halfblockquota.Thesamplewasdividedintotwopartsforseveralreasons:1)toprovidedataforpossiblyinterestingmethodologicalcomparisons;and2)onthechancethattherearesomedifferencesovertime,thatitwouldbepossibletoassignthesedifferencestoeithershiftsinsampledesigns,orchangesinresponsepatterns.Forexample,ifthepercentageofrespondentswhoindicatedthattheywere'veryhappy'increasedby10percentbetween1974and1976,itwouldbepossibletodeterminewhetheritwasduetochangesinsampledesign,oranactualincreaseinhappiness.Thereisconsiderablecontroversyandambiguityaboutthemeritsofthesetwosamples.Textbooktestsofsignificanceassumefullratherthanmodifiedprobabilitysamples,andsimplerandomratherthanclusteredrandomsamples.Ingeneral,thequestionofwhattodowithamixtureofsamplesisnoeasiersolvedthanthequestionofwhattodowiththe'pure'types.InvestigatorswhohaveappliedstatisticalteststopreviousGeneralSocialSurveydatashouldcontinuetoapplythosetests.Investigatorswhohaverefrainedfromapplyingsuchtestsmaynowwanttoperformanalysesontheprobabilitysubsample.Thiswould,ofcourse,reducethenumberofcasesbyonehalf.Whateverchoiceinvestigatorsmake,itshouldberememberedthatthetwosubsamplesrepresentthesameuniverse.Havingallowedfortheappearanceofallitemsinthetransitionalsampledesign,theGeneralSocialSurveythenswitchedtoafullprobabilitysampleforthe1977,1978,1980,and1982--2008surveys.ThevariableSAMPLEcanbeusedtoseparatetheblockquotaandfullprobabilitysamplesonthe1975and1976surveys(2188)(Source:Davis,JamesAllan;Smith,TomW.;andMarsden,PeterV.GeneralSocialSurveys,1972--2008:CumulativeCodebook/PrincipalInvestigator,JamesA.Davis;DirectorandCo-PrincipalInvestigator,TomW.Smith.Chicago:NationalOpinionResearchCenter,2009.2,656pp.,28cm.--(NationalDataProgramfortheSocialSciencesSeries,no.18).)PrincipalInvestigatorsJamesA.DavisTomW.SmithPeterV.MarsdenGSSSampling(FullProbabilityPortion)Thefollowingappliestothefullprobabilitysamplethatcomprisesapproximatelyhalfofthecasesinthisdataset:Thenationalsampleusesavariationofthestratifiedprobabilityproportionaltosize(PPS)method.First,thenationisdividedintoclusters,calledprimarysamplingunits(PSUs).PSUsarebasedoncounties;thetotalnumberofhouseholdsineachcounty,asreportedbythemostrecentcensus…EachmetropolitancountyisregardedasaPSU,asaremanynonmetropolitancounties.Sparselypopulatedruralcountiesaremergedintoadjacentcountyunitstoobtainaminimumof2,000housingunits.SinceNORCsamplesarebasedon100PSUs,thetotalnumberofhousingunitsforayearisdividedby100toyieldasamplinginterval.(SomePSUshavetotalnumbersofhousingunitsexceedingthesamplinginterval,makingitimpossibleforthemnottobeincludedinasample.SeeTourangeau,Johnson,Qian,andShin,1993forfurtherdiscussiononhowthesePSUsaretreated.)EachPSUisassignedanumberrangebasedonitstotalnumberofhousingunits.Thenarandomnumberisselected,andthefirstPSUselectedistheonewithinwhoserangetherandomnumberfalls.AdditionalPSUswereselectedbyaddingthesampleintervaltotheinitialrandomnumberonacumulativebasis.Oncethe100PSUsareselected,blocksegmentsconsistingoflinkedgroupsofoneormoregeographicallycontiguouscensusblocsarealsoselected.SegmentselectionsweremadeusingPPSmethodsonthebasisofthetotalnumberofhousinguni