下载

0下载券

加入VIP
  • 专属下载特权
  • 现金文档折扣购买
  • VIP免费专区
  • 千万文档免费下载

上传资料

关闭

关闭

关闭

封号提示

内容

首页 信度分析(Reliability Analysis)

信度分析(Reliability Analysis).doc

信度分析(Reliability Analysis)

Tina弯弯
2018-10-14 0人阅读 举报 0 0 0 暂无简介

简介:本文档为《信度分析(Reliability Analysis)doc》,可适用于社会民生领域

信度分析(ReliabilityAnalysis)信度分析(ReliabilityAnalysis)Prepareaquestionnairesurvey,inordertofurthertestthereliabilityandvalidityofthequestionnaire,namelytodoreliabilityanalysis(ReliabilityAnalsis),andthemeasurementresultsareaccurateandreliabilityisindependentornot,itsfunctionistotestwhetherthemeasurementitselfisstableReliabilityreferstoameasureofaccuracyoraccuracy,reliability,includingstabilityandconsistencyscholarKerlingerbelievesthatreliabilitycanmeasurereliability,consistency,andstabilityofthetoolsThetestreliabilityishigh,saidthetestresultsmorecredible,butalsocannotexpectthetwotestresultsareconsistent,inadditiontoaffectingthequalityoftestreliability,butalsobymanyothersubjectsfactors,thereisnotatestisabsolutelyreliableReliabilityisonlyasizedifferenceTheconsistencyofthequestionnaireisonlythesamegroupofpeopleacceptedthesamenature,thesametypeofquestions,thesamepurposeofallkindsofquestionnairesurvey,inthemeasurementresultsshowastrongpositivecorrelationAtoolofhighstabilitymeansthatagroupofpeoplereceivethesamemeasuringtoolsindifferenttimeandspace,buttheresultisverysmallTothereasonoferrorsmeasurementreliability,includingthefactorsfromtheresearcher:measuringthecontent(words,form),impropersituation(thelengthoftime,inthatatmosphere,etc)andresearchersneglect(mishear,memoryetc)andfactorsfromrespondentsmaybebecauseoftheirpersonality,age,educationlevel,socialclassandotherpsychologicalfactors,whichaffecttheaccuracyoftheanswerThetypeofquestionnaireandtheinfluenceoftimeintervalarethetwomainfactorsthataffectthereliabilityThroughthetestofreliabilityandvalidity,theresearcherscanknowwhetherthequestionnaireitselfisgoodornot,asabasistoimprovethecorrectionandavoidmakingwrongjudgmentsInaddition,therelationshipbetweenvalidityandreliability:reliabilityisthenecessaryandnotsufficientconditionofvalidityBothvalidityandreliability,butreliabilityisnotnecessarilyvalidTherearemanywaystoexaminereliabilityAmongthem,thefourthmostcommonlyusedCronbachalphacoefficientsarethefollowingfourtypes:,(RetestMethod)method:usingthesamequestionnairetothesamegroupofparticipants,atdifferenttimes,beforeandafterthetesttwotimes,calculatethecorrelationcoefficientoftwopoints,thecoefficientisalsocalledstabilitycoefficient(CoefficientofStability)Itshouldbenotedthatthecorrelationcoefficientishigh,whichmeansthatthereliabilityofthistestishigh,andthetimebetweenthetwotestsshouldbeappropriateIfthetwotestintervalistooshort,thesubjectscanusuallyscorewillincrease,butifenoughnumberofquestionscanbeavoidedbutifthetwotestintervalistoolong,thesubjectsofmentalgrowth,thestabilitycoefficientmayalsobereduced,copycorrelationmethod(EquivalentFormsMethod):thecopyissimilar,thedifficultyoftestistwo,tothesamegroupofparticipants,forthefirsttimeinacontinuoustest,seconduseacopy,twocopiesofthecorrelationcoefficientofthescoresforduplicatecoefficient(CoefficientofForms)CoefficientofEquivalence(orequivalentcoefficient)Iftwotestsarenotimplementedatthesametime,theycanbetestedagainforaperiodoftime,sothatthecorrelationcoefficientsarestableandequivalentThereplicationcorrelationmethodisthebestmethodformeasuringthereliabilityofthetest,butitisverydifficulttomakeaduplicatetestMoreover,thereplicacorrelationmethodisnotaffectedbymemoryutility,andthecorrelationofmeasurementerrorislowerthanthatofrecorrelationmethod,thebinarymethod(SplitHalfMethod)andcopycorrelationmethodisverysimilar,thebinarymethodistestedatthesametime,thebestpropertiesofcontent,ontwoquestiondifficultyintoconsideration,thetwoproblemsareconsistentasfaraspossibleSplithalfreliabilitycoefficient(splithalfcoefficient):thesamescaletestsubject(projectcontentsimilarity),foldintwo(odd,evennumber),thecorrelationcoefficientforthetwohalfofthetestscore,KeBiQucanalpha(Cronbachalpha):Cronbachproposedthealphacoefficient,splithalfmethodtoovercomesomeshortcomings,iscurrentlythemostcommonlyusedreliabilityofsocialscienceresearchMeasurethereliabilityofasetofsynonymousorparalleltestsIfalltheitemsinthescalereflectthesametrait,thereshouldbearealcorrelationbetweentheitemsIfthereisnocorrelationbetweenotheritemsinaprojectandscale,itisindicatedthattheprojectdoesnotbelongtothisscale,butshouldbeeliminatedAslongasyoucandothequestionnairereliabilityanalysis,provideobjectiveindexesastestandscalespecificevidenceofqualitydegreeCorrelation:thehigherthecorrelationcoefficient,thehigherthecorrelationInternalconsistency:thehigherthecorrelationcoefficient,thehighertheinternalconsistencyDiscriminationofreliabilityvalue:theproductdifferencecorrelationreachedsignificantlevel,thecorrelationcoefficientwasexpressedbya*,theproductdifferencecorrelationresultwassignificantlevel,andthecorrelationcoefficientwastwo*Formula:Alpha=k(k){(sigmaI)Sigmat}K:thenumberofitemsinthescaleSigmaI:varianceofscoresforallrespondentsintheprojectI(i=,,(K)SigmaT:varianceoftotalscoresofallrespondents,ThetotalscoreofeachintervieweereferstothesumofthescoresoftherespondentsineachitemIftherelativecoefficientofthetotalscoreofthisitemistoolow,itmaybeconsideredtobedeletedAfterdeletingthisquestion,thescaleacoefficient,ifsuddenlybecometoolarge,meansthatafterdeletingthequestion,thescaleacoefficientcanbeimprovedCalculateeachitemscoreandtotalcorrelation,correlationcoefficientofeachiteminorder,wherethecorrelationcoefficientclosetooftheprojectcanberemoved,acorrelationcoefficientorasuddendropintheprojectcanalsobetakenoutWhenthecoefficientofalphaislow,iftheitemisenough,itmeansthatsomeitemsarenothomogeneous,soitshouldbeeliminatedTabletwotableofreliabilityandCronbachcoefficientreliabilityCronbachalphacoefficientUntrustworthyCronbachcoefficientislessthanBarelycredible<<Cronbachalphacoefficientcredible<<CronbachalphacoefficientIt'sbelievable(themostcommon)<<CronbachalphacoefficientVerybelievable(frequentlyseen)<<CronbachalphacoefficientVerycredible=CronbachalphacoefficientTheoperationstepsofreliabilityanalysisinSPSS:Step:Analyze,ScaleReliability,AnalysisStep:Alpha,ModelmodeselectionvariableslisttopicvariablesintheleftpointprojectStep:accordingtoStatistics,inwhichScaleifitemdeleted,clickonContinue,OK

用户评价(0)

关闭

新课改视野下建构高中语文教学实验成果报告(32KB)

抱歉,积分不足下载失败,请稍后再试!

提示

试读已结束,如需要继续阅读或者下载,敬请购买!

评分:

/6

VIP

在线
客服

免费
邮箱

爱问共享资料服务号

扫描关注领取更多福利