Consequences of the joint consideration of individual and aggregate data in correctional social research