Distal Outcomes in Latent Transition Analysis: Linking Later Outcomes to Transition Histories

Latent transition analysis (LTA) is widely used to study movement among latent classes over time, yet distal outcomes are often modeled as a function of final class memberships, resulting in a relatively static representation of the dynamic developmental process. Although distal outcomes can, in principle, be conditioned on combinations of latent class variables or transition histories, such specifications may be underused, difficult to implement, or insufficiently documented. This paper presents a practical, theory-guided approach for relating distal outcomes to longitudinal transition histories in LTA. We focus on a grouped-pattern approach in which theoretically meaningful transition patterns are specified using Mplus dot-language constraints while adjusting for classification uncertainty. Using math and science attitudes across high school, we illustrate how stability and change relate to later STEM career attainment.