Latent Class Analysis (LCA) is a classification method that produces a categorical latent factor rather than a continuous one, as in analytic techniques such as factor analysis and Item Response Theory (IRT) (Hagenaars and McCutcheon, 2002). LCA calculates the probability of being in a particular group or class conditional on a pattern of item responses. Each class has a specific profile of item endorsement unique to that class and classes are independent of each other. LCA can be a valuable tool in addressing the fundamental issue facing DSM-V concerning whether subtypes of alcohol abuse and dependence exist and whether these subtypes can be superimposed upon a continuous underlying dimension.
