Xu Jun
2008-Nov-05 15:13 UTC
[R] latent class analysis of nominal and continuous indicators
Dear R-help listers, I am a new convert to R. I am trying to use a r package to conduct latent class analysis as a triangulation check of my cluster analysis using the cluster package in R. I have about 30 cases and 6 indicators, some of which are binary indicators and others are ratio-level variables (percentages). I looked around for information in flexmix, lca, and poLCA, and couldn't find anything relevant or probably I missed them. I checked MPlus, and it seems to me MPlus can handle latent class analysis with indicators of different measurement levels (continous or nominal). And my intuition is that R ought to be able to it. Any ideas about which package I should use and any examples that I can model after. As a long-time user of Stata, I am also used to making easy tables (of regression results like those in published journal articles) using the estout command. I am not sure if R has similar functionalities. Thanks! Jun Xu, Ph.D. Assistant Professor Ball State University Muncie, IN 47306 [[alternative HTML version deleted]]