The epistemology of operations research directed at force composition or weapon-system selection is different from that of the physical sciences. Though tests of parts of the operations-research process are possible, and indeed required, no empirical tests of the general recommendations are possible. It is thus found that this type of operations research is not science in the sense of the physical sciences. Hence the philosophical question of the nature and validity of operations-research studies arises. The tests of validity, the criteria for judging quality, involve judgment. Several criteria for judging are presented. In practice a combination of all is required. The methodology of operations research, the subject matter of the problem, and the expertness of the analyst, all play a role in lending credence to the recommendations of a study.