Evaluation is a necessary and useful way of assessing policymaking. However, evaluation is both an art and a science. Not only the data but the underlying assumptions must be carefully interpreted. Evaluation is a constant process, whether conscious or unconscious, and activities can be assessed even where nonquantifiable values are involved. Programs and projects should be designed to permit evaluation of their effectiveness, significance, efficiency, compliance with the original design, and unanticipated effects. Evaluators recognize the need for quantitative measures of performance as well as the difficulty of developing such measures. Many issues have not been resolved by evaluation: police effectiveness in controlling crime, the value of management procedures in police departments and prisons, the benefits of differing allocations of police resources, and the role of corrections. However, evaluations of a number of vital issues indicate that added probation resources do not affect reconviction rates, the quality of prison life varies greatly across States, crime control is not simply a matter of police manpower, rapid police response does not necessarily affect arrest rates, and extra lighting may reduce street crime. Evaluators should be aware of the variables involved in ongoing and retrospective evaluations, as well as political factors in all evaluations. Footnotes are provided.