Latent Class Models

Hwan Chung

Many quantities of interest in sociology, psychology, public health, and medicine are unobservable but well-conceived characteristics such as attitudes, temperament, psychological diagnoses, and health behaviors. Such constructs can be measured indirectly by using multiple items as indicators of the unobservable characteristics. The latent class (LC) model can classify individuals into population subgroups based on these unobservable characteristics, by using their responses to questionnaire items that are related to those characteristics. Suppose, for example, we are interested in the construct of nicotine dependence and want to classify respondents into groups corresponding to different types of nicotine dependence we believe are indicated by a number of behaviors relating to cigarette use. We can collect data on several items pertaining to cigarette use and then apply an LC model to this data to identify two or more nicotine dependence types to which smokers might belong. We can use this LC model to classify ...

Users without subscription are not able to see the full content on this title. Please, subscribe or login to access all content on this website.