In his Economic Opportunity Message to the Congress on February 19, 1969, President Nixon mentioned briefly that the preliminary results of a Westinghouse Learning Corporation-Ohio University evaluation indicated that "the long-term effect of Head Start appears to be extremely weak." This terse announcement triggered a major public controversy that ranged over the Congress, the executive branch, and the educational research community. Much of the debate focused on the esoteric techniques of modern statistical analysis, but the issues were far larger than the particular study. In conflict were two basic premises—one concerned with how to start programs and the other concerned with how to analyze them— that emerged independently in the mid-1960's. For the notion underlying much of the war on poverty—that effective programs could be developed quickly and launched full-scale (and Head Start was a prime case)—was being called into question by the type of evaluative analysis that lay at the base of the Planning, Programming, Budgeting System initiated in late 1965. The outcome of the clash will have profound implications for governmental procedures for developing new large-scale programs and measuring their results. This paper traces both the events that led up to the controversy and the controversy itself in order to look at the implications for future policy.