The matching of treatment groups in terms of means or other summary characteristics of the matching variables. Has been shown to be almost as efficient as the matching of individual subjects in some circumstances.