Jun Xu posted on Tuesday, October 07, 2008 - 7:49 am
I am working a data-reduction project. I have about 30 countries, each with a series of binary indicators and some ratio variables bounded between 0 and 1 (e.g., educational expenses as percent of GDP). I intend to find latent classes based on these indicators. Based on the readings that I had, LCA only applies to categorical or ordinal indicators, not a mixture of these indicators. Does MPlus tackle this problem and/or can you point me to some relevant references. Thanks.
Jun Xu, Ph.D. Assistant Professor Department of Sociology Ball State University Muncie, IN 47306
In Mplus, the latent class indicators can be continuous, censored, binary, ordered categorical (ordinal), unordered categorical (nominal), counts, or combinations of these variable types.
Jun Xu posted on Tuesday, October 07, 2008 - 8:59 am
Thanks for your answer. Another concern that I have before plunging into MPlus is sample size. Is 30 observations with about 6-8 variables a large enough sample size for the EM algorithm. Or you would suggest I go on another route?