subject matter-specific types??that permit a full specification on the distribution of the end result by using random outcomes to seize in-subject matter correlations. The choice regarding which method of pick out partly will depend on the goal on the investigation and the desired interpretation from the estimated effects (populace-common compared