Longitudinal measurement invariance of psychosocial measures in physical activity research: an application to adolescent data


Correspondence concerning this article should be addressed to Scott C. Roesch, Department of Psychology, San Diego State University, 5500 Campanile Drive, San Diego, CA 92182-4611, USA. E-mail: sroesch@mail.sdsu.edu


The current study served as a practical and substantive guide to establishing longitudinal measurement invariance of psychosocial measures commonly used in adolescent physical activity (PA) research. Psychosocial data on an initial sample of 878 adolescents (ages 11–15) recruited through primary care providers were provided at baseline, 6, 12, and 24 months. The target measures included family support, peer support, decisional balance (pros, cons), self-efficacy, and behavioral strategies. Five of the six psychosocial measures exhibited strict longitudinal measurement invariance, with the sixth measure (self-efficacy) exhibiting strong longitudinal measurement invariance. These findings support the equivalence of these measures across time, and provide the foundation to substantively interpret group differences and associations involving these measures and PA.