The ability to examine within-person change is essential to test process-based theories in sport and exercise psychology. Longitudinal data, whether experimental or observational, are prerequisites to be able to examine change processes, but most longitudinal studies in sport and exercise psychology focus solely on between-person/group differences, not on within-person change. In this review, we (1) provide researchers in the sport and exercise psychology field with a framework for longitudinal research that focuses on within-person change; (2) provide an overview of how researchers in sport and exercise psychology currently analyse longitudinal data, which showed that most longitudinal studies focus on between-person/group differences; and (3) provide examples of statistical models for analysing longitudinal data that correspond to the framework for longitudinal research. In the examples, we focus on latent variable modelling, such as latent growth-curve modelling and latent change-score modelling, which capture within-person change. We argue that there is a need for stronger emphasis on the match among theory of change, temporal design, and statistical models when designing longitudinal studies in sport and exercise psychology. © 2016 Informa UK Limited, trading as Taylor & Francis Group