10.07.2015 Views

Mplus Users Guide v6.. - Muthén & Muthén

Mplus Users Guide v6.. - Muthén & Muthén

Mplus Users Guide v6.. - Muthén & Muthén

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

CHAPTER 7zero-inflated Poisson regression equations for count latent classindicators.The structural model describes three types of relationships in one set ofmultivariate regression equations: the relationships among thecategorical latent variables, the relationships among observed variables,and the relationships between the categorical latent variables andobserved variables that are not latent class indicators. Theserelationships are described by a set of multinomial logistic regressionequations for the categorical latent dependent variables and unorderedobserved dependent variables, a set of linear regression equations forcontinuous observed dependent variables, a set of censored normal orcensored normal regression equations for censored-inflated observeddependent variables, a set of logistic regression equations for binary orordered categorical observed dependent variables, and a set of Poissonor zero-inflated Poisson regression equations for count observeddependent variables. For logistic regression, ordered categoricalvariables are modeled using the proportional odds specification.Maximum likelihood estimation is used.The general mixture model can be extended to include continuous latentvariables. The measurement and structural models for continuous latentvariables are described in Chapter 5. In the extended general mixturemodel, relationships between categorical and continuous latent variablesare allowed. These relationships are described by a set of multinomiallogistic regression equations for the categorical latent dependentvariables and a set of linear regression equations for the continuouslatent dependent variables.In mixture modeling, some starting values may result in local solutionsthat do not represent the global maximum of the likelihood. To avoidthis, different sets of starting values are automatically produced and thesolution with the best likelihood is reported.All cross-sectional mixture models can be estimated using the followingspecial features:• Single or multiple group analysis• Missing data• Complex survey data142

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!