A latent trait and a latent class model for mixed observed variables

Presents a latent trait and a latent class model for analyzing the relationships among a set of mixed manifest variables using 1 or more latent variables. The set of manifest variables contains metric (continuous or discrete) and binary items. For the latent trait model the latent variables are assumed to follow a multivariate standard normal distribution. The method gives maximum likelihood estimates of the model parameters and standard errors of the estimates by analyzing the data as they are without using any underlying variables. The mixed latent trait and latent class models are fitted using an EM algorithm. Three sets of survey data are analyzed to illustrate the application of the mixed model. The advantage of handling simultaneously binary and continuous observed variables within the framework of a latent variable model over other methods is that response patterns of the individual can be analyzed. (PsycINFO Database Record (c) 2008 APA, all rights reserved)