Cohort effects: A possible limitation to the interpretation of longitudinal studies