Latent Transition Analysis: An Introduction

Background and Purpose: The application of developmental processes are used throughout social work research to understand time-related changes in human development. These time-related changes are referred to as transitions among stages, where stages are distinct periods of time in with characteristics that are qualitatively or quantitatively different from others. For example, a researcher may investigate how an intervention affects the Stages of Change in relation to addiction, or how skill proficiency changes as a result of specific instructional methods over a period of time. Although developmental processes are dynamic in nature, most analytic techniques use static methods to measure change. These types of methods limit the ability of the researcher to describe multiple transitions over a time period. Subsequently, these limitations often compound a researcher’s ability to describe the moderating effects of an intervention to transitions as they occur within stages of a developmental process, as well as the magnitude of the intervention upon the transitions over time. A method that alleviates these limitations is Latent Transitional Analysis (LTA).

This presentation will introduce LTA as an analysis that more accurately captures the dynamic nature of stages within a developmental process. LTA identifies latent statuses among a given set of characteristics and describes transitions among the statuses over a given period of time. A status is an empirically derived categorical latent variable that is chronologically measured, in which status assignment for specific measurement may vary given endorsement or non-endorsement to a given set of characteristics.

Methods: The presentation will: 1) Provide an introduction to LTA using non-technical language, tables and figures, 2) Provide an overview of important terminology, parameters and concepts associated with an LTA, 3) Describe important data-management techniques used in preparation of conducting an LTA, and 4) illustrate simple coding techniques of LTA using Mplus by applying the analysis to delinquent behavior among youth across the adolescent development period.

Results: By the conclusion of this presentation, participants will be able to: 1) Describe and discuss simple LTA applications, 2) Articulate and apply important terminology associated with an LTA, 3) develop basic conceptual modeling of an LTA, 4) describe important data-management functions associated with LTA, and 5) describe the step-wise method of conducting an LTA with associated parameters.

Conclusion: Without needing to account for traditional assumptions, an LTA is a statistical technique that combines a person-centered approach using cross-sectional measurement to describe categorical-latent statuses and how transitions occur among the statuses over a specified period of time. Applications of this technique for future research will be discussed.