Latent Class and Latent Transition Analysis
June 29 - July 1, 2009
Linda M. Collins, Ph.D. and Stephanie T. Lanza, Ph.D.
July 2, 2009: Summer Institute Improves Research Skills by Liam Jackson
The 14th Summer Institute on Longitudinal Methods was sponsored by The Pennsylvania State University and the National Institute on Drug Abuse. It provided researchers who have varying levels of methodological training the opportunity to familiarize themselves with longitudinal data analysis. In addition, the Institute was designed to facilitate the exchange of ideas among substantive and methodological researchers. 2009's three-day Institute featured a workshop on latent class and latent transition analysis by experts Linda Collins and Stephanie Lanza of Penn State.
The goal of this workshop was to help participants gain the background and skills to be able to address interesting research questions using latent class and latent transition analysis. Latent class analysis is conceptually similar to factor analysis. However, in latent class analysis, the latent variables are categorical, and individuals are sorted into mutually exclusive and exhaustive latent classes based on a set of item responses. Latent class analysis identifies latent classes in data and estimates their prevalence, while simultaneously adjusting for measurement error. Latent class models can be used to estimate change over time in latent class membership using longitudinal data, in a variation called latent transition analysis. In addition, multiple-groups analyses can be performed, and covariates can be introduced to predict latent class membership and changes in latent class membership. Class time was spent in lecture, discussion, software demonstrations, and hands-on exercises. In addition, participants were encouraged to bring their own data to the workshop to analyze on the final day of the workshop. The software used in this course was two downloadable add-on procedures for SAS Version 9 for Windows: PROC LCA (for latent class analysis) and PROC LTA (for latent transition analysis).
Linda Collins is a Professor of Human Development and Family Studies and the Director of The Methodology Center in the College of Health and Human Development at Pennsylvania State University. This NIDA-funded Center, directed by Dr. Linda Collins, is devoted to the advancement and dissemination of statistical methodology related to research on the prevention and treatment of problem behaviors. Her graduate training took place at the J.P. Guilford Laboratory of Quantitative Psychology at the University of Southern California. Dr. Collins’ research program has focused on statistical methodology and measurement procedures for longitudinal data, particularly in the areas of human development and drug abuse prevention research. She has worked extensively on the development of latent transition analysis. More recently, she has focused on issues of design and optimization of behavioral interventions for prevention and treatment of health problems. Her work has been funded by the National Institutes of Health for over 20 years. Dr. Collins is the president-elect of the Society for Prevention Research; a Fellow of the American Psychological Association and the American Psychological Society; a recipient of the Cattell Award for outstanding early career contributions to multivariate behavioral research; a past president of the Society of Multivariate Experimental Psychology; and a recipient of Pennsylvania State University’s Faculty Scholar Medal in the Social and Behavioral Sciences. She was a Core Scientist in the Tobacco Etiology Research Network, funded by the Robert Wood Johnson Foundation. She is past Associate Editor of the journals Multivariate Behavioral Research, Prevention Science, and Journal of Educational and Behavioral Statistics, and has served on the editorial board of Psychological Methods. Dr. Collins has edited or co-edited numerous books and special issues of journals, and is currently writing a book with Dr. Lanza on latent class and latent transition analysis.
Stephanie Lanza has served as the Scientific Director of The Methodology Center at Penn State since 2004. Dr. Lanza’s interdisciplinary background includes a Masters in Applied Statistics and a Ph.D. in Human Development and Family Studies from Penn State. Her research interests include the analysis of longitudinal data, mixture models including latent class and latent transition analysis, and risk assessment. She has a NIDA-funded research grant to develop a latent-class framework for quantifying multiple, interactive risks for developmental outcomes and a NIDA-funded conference grant to continue this long-running Summer Institute on Longitudinal Methods series at Penn State. She leads the development of SAS procedures for latent class analysis (PROC LCA) and latent transition analysis (PROC LTA) and is currently writing a book with Dr. Collins on latent class and latent transition analysis. In 2008 Dr. Lanza received the Society for Prevention Research Early Career Award. Dr. Lanza plans to continue work to advance statistical methods for the social and health sciences, particularly as they apply to adolescent and early adulthood outcomes including substance use and sexual risk behavior.