The acceptance of benefit-cost analysis is so broad that it virtually defines the field of study concerned with the economic evaluation of extra-market goods. Because of this wide acceptance, debate has moved on to such issues as how to best measure benefits· (Bradford) or how to apply them (Kalter) or which of several benefit measures is theoretically correct to use (Seckler). It is with some trepidation, therefore, that we suggest an approach to the assessment of preferences for publically-provided goods which is outside the benefit-cost framework. The approach is distinctive because it is budget-oriented rather than benefit-oriented. We believe it to be as promising as any other existing method when answers are desired to a particular type of public planning problem. One of the purposes of this paper is to support this contention. Another is to give a conceptual description of the suggested technique. Both of these aims depend ultimately, however, I on a clear understanding of what we mean by "government good" by "public planning problem" and by "budget-oriented" method.