This paper reviews the past 25 years of applied social research, with an emphasis on evaluation research, whose intent is the estimation of the net impacts or effects of social programs. The 1960s and early 1970s represent the "Golden Age" of evaluation, an age that ended with the Reagan Administration. During this period, the evaluation field was dominated by the randomized, controlled experimental paradigm. Accordingly, several of the major field experiments of the era are reviewed in some detail. Recently, various alternatives to this paradigm have been advanced; they too are discussed. The key lesson from the Golden Age is that the expected effects of social programs hover near zero, a devastating discovery for the social reformers of the time. Some reasons why this is the case are presented. One important