Data analysis and modeling longitudinal processes

Abstract

This article presents a nontechnical overview of the major data analysis techniques for modeling longitudinal processes, with an explicit focus on their advantages and disadvantages as tools for drawing inferences about different specific aspects of change over time. It is argued that traditional longitudinal analysis techniques offer limited ways of addressing many specific questions about change. Recent advances in latent variable techniques, when adequately driven by theory, design, and measurement, offer a unified and flexible framework for addressing such questions.