matter-precise designs??that permit a complete specification with the distribution of the result by using random consequences to capture within-subject correlations. The selection as to which method of choose partly will depend on the purpose in the investigate and the desired interpretation from the approximated consequences (population-ordinary v