Longitudinal data in the study of causal processes: Some uses and some pitfalls.