Mplus培训手册(2)

LatentVariableModelingUsingMplus:Day1BengtMuth´en&LindaMuth´enMplus´en&LindaMuth´enMplusModeling1/186TableofContentsI11.MplusBackground22.MediationPathAnalysis2.1Example:MediationOfFetalAlcoholSyndrome2.2Example:ModeratedMediationOfAggressiveBehavior2.3Causally-DefinedEffectsInMediationAnalysis2.4Two-LevelPathAnalysisWithABinaryOutcome:HighSchoolDropout33.BayesianAnalysis3.1BayesianMediationModelingWithNon-InformativePriors:TheMacKinnonATLASExample44.FactorAnalysis4.1EFAOfHolzinger-SwinefordMentalAbilitiesData4.2Bi-FactorModelingOverview4.2.1Bi-FactorModelingOfThe24-VariableHolzinger-SwinefordDataBengtMuth´en&LindaMuth´enMplusModeling2/186TableofContentsII4.3TheESEMFactorAnalysisApproach:Multiple-GroupEFAOfAggressiveBehaviorOfMalesAndFemales4.4TheBSEMFactorAnalysisApproach4.4.1BSEMForHolzinger-Swineford19Variables4.5.1OtherFactorModels:Second-OrderFactorModel4.5.2OtherFactorModels:Multi-Trait,Multi-Method(MTMM)Model4.5.3OtherFactorModels:LongitudinalFactorAnalysisModel4.5.4OtherFactorModels:ClassicACETwinModel55.MeasurementInvarianceAndPopulationHeterogeneity5.1CFAWithCovariates(MIMIC):NELSData5.2Multiple-GroupAnalysis66.StructuralEquationModeling(SEM):ClassicWheatonEtAl.SEM6.1ModelingIssuesInSEM77.GrowthModeling:TypicalExamplesBengtMuth´en&LindaMuth´enMplusModeling3/186TableofContentsIII7.1ModelingIdeas:IndividualDevelopmentOverTime7.2LSAYGrowthModelingWithTime-InvariantCovariates7.3LSAYGrowthModelingWithRandomSlopes7.4SixWaysToModelNon-LinearGrowth7.5PiecewiseGrowthModeling7.6GrowthModelingWithMultipleProcesses7.7Two-PartGrowthModeling7.8AdvancesInMultipleIndicatorGrowthModeling7.8.1BSEMforAggressive-DisruptiveBehaviorintheClassroom7.9AdvantagesOfGrowthModelingInALatentVariableFrameworkBengtMuth´en&LindaMuth´enMplusModeling4/186TheMapOfTheMplusTeamBengtMuth´en&LindaMuth´enMplusModeling5/186TheOtherMembersOfTheMplusTeamThuyMichelleSarahBengtMuth´en&LindaMuth´enMplusModeling6/1861.MplusBackgroundInefficientdisseminationofstatisticalmethods:Manygoodmethodscontributionsfrombiostatistics,psychometrics,etcareunderutilizedinpracticeFragmentedpresentationofmethods:TechnicaldescriptionsinmanydifferentjournalsManydifferentpiecesoflimitedsoftwareMplus:IntegrationofmethodsinoneframeworkEasytouse:Simple,non-technicallanguage,graphicsPowerful:GeneralmodelingcapabilitiesBengtMuth´en&LindaMuth´enMplusModeling7/186MplusIntegratesAMultitudeOfAnalysisTypesUsingTheUnifyingThemeOfLatentVariablesExploratoryfactoranalysisStructuralequationmodelingItemresponsetheoryanalysisGrowthmodelingLatentclassanalysisLatenttransitionanalysis(HiddenMarkovmodeling)GrowthmixturemodelingSurvivalanalysisMissingdatamodelingMultilevelanalysisComplexsurveydataanalysisBayesiananalysisCausalinferenceBengtMuth´en&LindaMuth´enMplusModeling8/186MplusIntegratesAMultitudeOfAnalysisTypesUsingTheUnifyingThemeOfLatentVariablesExploratoryfactoranalysisStructuralequationmodelingItemresponsetheoryanalysisGrowthmodelingLatentclassanalysisLatenttransitionanalysis(HiddenMarkovmodeling)GrowthmixturemodelingSurvivalanalysisMissingdatamodelingMultilevelanalysisComplexsurveydataanalysisBayesiananalysisCausalinferenceBengtMuth´en&LindaMuth´enMplusModeling8/186MplusIntegratesAMultitudeOfAnalysisTypesUsingTheUnifyingThemeOfLatentVariablesExploratoryfactoranalysisStructuralequationmodelingItemresponsetheoryanalysisGrowthmodelingLatentclassanalysisLatenttransitionanalysis(HiddenMarkovmodeling)GrowthmixturemodelingSurvivalanalysisMissingdatamodelingMultilevelanalysisComplexsurveydataanalysisBayesiananalysisCausalinferenceBengtMuth´en&LindaMuth´enMplusModeling8/186MplusIntegratesAMultitudeOfAnalysisTypesUsingTheUnifyingThemeOfLatentVariablesExploratoryfactoranalysisStructuralequationmodelingItemresponsetheoryanalysisGrowthmodelingLatentclassanalysisLatenttransitionanalysis(HiddenMarkovmodeling)GrowthmixturemodelingSurvivalanalysisMissingdatamodelingMultilevelanalysisComplexsurveydataanalysisBayesiananalysisCausalinferenceBengtMuth´en&LindaMuth´enMplusModeling8/186MplusIntegratesAMultitudeOfAnalysisTypesUsingTheUnifyingThemeOfLatentVariablesExploratoryfactoranalysisStructuralequationmodelingItemresponsetheoryanalysisGrowthmodelingLatentclassanalysisLatenttransitionanalysis(HiddenMarkovmodeling)GrowthmixturemodelingSurvivalanalysisMissingdatamodelingMultilevelanalysisComplexsurveydataanalysisBayesiananalysisCausalinferenceBengtMuth´en&LindaMuth´enMplusModeling8/186MplusIntegratesAMultitudeOfAnalysisTypesUsingTheUnifyingThemeOfLatentVariablesExploratoryfactoranalysisStructuralequationmodelingItemresponsetheoryanalysisGrowthmodelingLatentclassanalysisLatenttransitionanalysis(HiddenMarkovmodeling)GrowthmixturemodelingSurvivalanalysisMissingdatamodelingMultilevelanalysisComplexsurveydataanalysisBayesiananalysisCausalinferenceBengtMuth´en&LindaMuth´enMplusModeling8/186TheMplusGeneralLatentVariableModelingFrameworkObservedvariablesxbackgroundvariables(nomodelstructure)ycontinuousandcensoredoutcomevariablesucategorical(dichotomous,ordinal,nominal)a