Latent Class Analysis

Latent class analysis (LCA) is similar to ordinary cluster analysis. Both techniques divide participants in a sample into groups based on their standing on a predetermined set of variables of interest to the researcher (e.g., political views, music preferences). The result is that individuals within a class or cluster are as similar to each other on the relevant variables as possible, but each class or cluster is as different, on average, from each of the other ones as possible. LCA is sometimes known as mixture modeling, for the variety of classes that emerge. The remainder of this entry discusses distinctions between LCA and related techniques and reviews practical aspects of running LCA models. Examples—both from the literature and hypothetical ones—appear throughout.