中文版五条目杜克大学宗教指数量表的内部一致性和重测信度研究.pdf

·30O · ShanghaiArchivesofPsychiatry,2014，Vo1．26，No．5 — — -Researchmethodsinpsychiatry= Internalconsistencyandtest-r·etestreliabilityoftheChinese versionofthe5一item DukeUniversityReligionIndex HanhuiCHEN ，ZhizhongWANG ，MichaelR．PHILLIPS，YanliSUN ，HuiG．CHENG Background：TheDukeUniversityReligionIndex(DUREL)isawidely—used5-item scaleassessingreligiosity． Aim ：Assesstheinternalconsistency,reliability, andfactorstructureoftherevisedChineseversionof DUREL． Methods：Usingprobabilityproportionatetosize(PPS)methodswerandomlyidentified3981households witheligibleoccupantsin20primarysamplingsitesinNingxiaHuiAutonomousRegion，aprovince in northwestChinainwhich34％ofthepopulationareMusIimsoftheHuiethnicgrouP．In3054householdsa screeninginterview wascompletedandanadultfam iIvmemberwasrandom lyselected；2425respondents completedthesurvey(includingtheDUREL)and188randomlyselectedindividualsrepeatedthesurveyan averageof2．5dayslater． Results：TheinternaIconsistency(Cronbachs0)【ofthe5itemsinthefuIIsamplewas0．90；itrangedfrom 0．70to0．90invarioussubgroupsofsubjectsstratifiedbyethnicity,urbanversusruraIresidence，and aboveversusbelow medianeducation．Thetest-retestreliabitity(intraclasscorrelationcoefficient)forthe totaIscoreinthefuIIsampiewas0．87；itrangedfrom 0．63to0．90inthedi仟erentsubgroupsofsubjects． ExpIoratoryfactoranaIysisinarandom haIfofthesamPIeidentifiedasinglefactor(eigenvaIue=4．21) thatexpIained84％ ofthetotaIvariance．ConfirmatOrvfactoranaIysisinthesecondhaIfofthesample confirmedtheunidimensionaImodeI：themodelfitmeasuresoftheone—factormodeIusingthe5item sc