Which research design can control for practice effects, cohort effects, and dropping out?