Measuring the Effectiveness of Nondefense Public Programs

This paper suggests that current public programs have been measured in grossly inadequate ways. It argues against the use at all levels of government of workload measures, physical standards, and cost criteria to represent outputs. It discusses the need for output measurement (1) that uses measures which are directly related to basic governmental objectives, (2) that includes multiple measures (since governmental programs seldom can be adequately evaluated by attempts to compress multiple measures into a single criterion such as a single monetary measure), (3) that identifies the differential effects on different population subgroups, (4) that does not excessively vise “proxy” or “lower-level” criteria without at least briefly examining the relation of the proxies to the basic objectives, and (5) that, where appropriate, uses qualitative, subjective, measures rather than neglecting important effects. Measurability does not require the quantification of absolute numbers, but can make use of many subjective approaches such as ratings, rankings, and the like.