Behavior Change Analysis

As part of a project to enhance our understanding of smoking patterns during cessation attempts among adult smokers, we published Repeated Measures Latent Class Analysis (RMLCA) of daily smoking status (any vs. no smoking) over the first 27 days of a quit attempt in four datasets. See:

This research was supported by the National Institute on Drug Abuse of the National Institutes of Health under Award Number R01DA033303. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Our objective in applying repeated measures latent class analysis to daily smoking status data was to describe the patterns of smoking behavior over the first few weeks of a quit attempt in a person-centered way without imposing assumptions of linear growth or change processes.  This approach allowed us to identify latent classes of smokers based on their patterns of smoking across the first 27 days of the quit attempt.  The patterns could be simple, as in the case of people who never smoked in the target period, or quite complex, as in the case of people who smoked intermittently initially and then returned to daily smoking.  We sought to identify the common patterns that emerged and then to examine the correlates of these patterns using an analytic method that is capable of identifying very complex patterns, even those that do not follow a smooth trajectory.  After identifying latent class patterns, we examined associations between latent class membership and baseline individual differences and long-term smoking status outcomes.  We also sought to test the effects of pharmacotherapies on smoking patterns in the early phase of quitting smoking. 

This website provides pragmatic answers to questions about how to implement a repeated measures latent class analysis in the study of behavior change.  More extensive guidance about latent class analysis and latent transition analyses is available through other sites.  We particularly recommend the following

  • The book on latent class and latent transition analysis written by Linda Collins and Stephanie Lanza:

Collins, L. M. & Lanza, S. T. (2010).  Latent Class Analysis and Latent Transition Analysis: With Applications in the Social, Behavioral, and Health Sciences.  Hoboken, NJ: Wiley, http://methodology.psu.edu/latentclassbook/

  • The training in latent class analysis and latent transition analysis provided by Stephanie Lanza through Statistical Horizons, LLC (http://statisticalhorizons.com/) is very helpful.