Patton (2008a) and Stufflebeam have both indicated that full meta-evaluations are not necessary for every evaluation. If they are costly, they should be reserved for evaluations that are what Patton calls "high stakes. ’, He proposes the following guideline: