Latent class analysis methods and software

This halfday short course introduces the concept of latent class analysis, which is a model-based statistical approach to determine hidden classes in data. The session will be illustrated with examples from the social sciences (particularly criminology), and will also describe a number of software products which can be used to fit these models. The focus of the session will be on understanding and
interpretation. Part one will focus on the concept of simple latent class models for categorical data and the practical issues in fitting such models. The second part will extend these models into longitudinal data.