Structural Equation ModelingAKASEM in a nutshellJargonSlide 5Slide 6Slide 7Diagram elementsPath DiagramSEM questionsSlide 11Slide 12Slide 13SEM limitationsSlide 15Slide 16Slide 17Slide 18Structural Equation ModelingIntro to SEMPsy 524AinsworthAKASEM – Structural Equation ModelingCSA – Covariance Structure AnalysisCausal ModelsSimultaneous EquationsPath AnalysisConfirmatory Factor AnalysisSEM in a nutshellCombination of factor analysis and regressionContinuous and discrete predictors and outcomesRelationships among measured or latent variablesDirect link between Path Diagrams and equations and fit statisticsModels contain both measurement and path modelsJargonMeasured variableObserved variables, indicators or manifest variables in an SEM designPredictors and outcomes in path analysisSquares in the diagramLatent VariableUn-observable variable in the model, factor, constructConstruct driving measured variables in the measurement modelCircles in the diagramJargonError or EVariance left over after prediction of a measured variableDisturbance or DVariance left over after prediction of a factorExogenous VariableVariable that predicts other variablesEndogenous VariablesA variable that is predicted by another variableA predicted variable is endogenous even if it in turn predicts another variableJargonMeasurement ModelThe part of the model that relates indicators to latent factorsThe measurement model is the factor analytic part of SEMPath modelThis is the part of the model that relates variable or factors to one another (prediction)If no factors are in the model then only path model exists between indicatorsJargonDirect EffectRegression coefficients of direct predictionIndirect EffectMediating effect of x1 on y through x2Confirmatory Factor AnalysisCovariance StructureRelationships based on variance and covarianceMean StructureIncludes means (intercepts) into the modelDiagram elementsSingle-headed arrow →This is predictionRegression Coefficient or factor loadingDouble headed arrow ↔This is correlationMissing PathsHypothesized absence of relationshipCan also set path to zeroPath DiagramDepressionBDICES-DZDRSNegative ParentalInfluenceDep parentInsecureAttachmentNeglectGenderEEEDEEESEM questionsDoes the model produce an estimated population covariance matrix that “fits” the sample data?SEM calculates many indices of fit; close fit, absolute fit, etc.Which model best fits the data?What is the percent of variance in the variables explained by the factors?What is the reliability of the indicators?What are the parameter estimates from the model?SEM questionsAre there any indirect or mediating effects in the model?Are there group differences?Multigroup modelsCan change in the variance (or mean) be tracked over time?Growth Curve or Latent Growth Curve AnalysisSEM questionsCan a model be estimated with individual and group level components?Multilevel ModelsCan latent categorical variables be estimated?Mixture modelsCan a latent group membership be estimated from continuous and discrete variables?Latent Class AnalysisSEM questionsCan we predict the rate at which people will drop out of a study or end treatment?Discrete-time survival mixture analysisCan these techniques be combined into a huge mess?Multiple group multilevel growth curve latent class analysis???????SEM limitationsSEM is a confirmatory approachYou need to have established theory about the relationshipsCannot be used to explore possible relationships when you have more than a handful of variablesExploratory methods (e.g. model modification) can be used on top of the original theorySEM is not causal; experimental design = causeSEM limitationsSEM is often thought of as strictly correlational but can be used (like regression) with experimental data if you know how to use it.Mediation and manipulation can be tested SEM is by far a very fancy technique but this does not make up for a bad experiment and the data can only be generalized to the population at handSEM limitationsBiggest limitation is sample sizeIt needs to be large to get stable estimates of the covariances/correlations200 subjects for small to medium sized modelA minimum of 10 subjects per estimated parameterAlso affected by effect size and required powerSEM limitationsMissing dataCan be dealt with in the typical ways (e.g. regression, EM algorithm, etc.) through SPSS and data screeningMost SEM programs will estimate missing data and run the model simultaneouslyMultivariate Normality and no outliersScreen for univariate and multivariate outliersSEM programs have tests for multi-normalitySEM programs have corrected estimators when there’s a violationSEM limitationsLinearityNo multicollinearity/singularityResiduals Covariances (R minus reproduced R)Should be smallCentered around zeroSymmetric distribution of errorsIf asymmetric than some covariances are being estimated better than
View Full Document