This paper proposes a method for answering the question: For a given level of effort, which of the many alternative family planning programs appears preferable? The approach here is to develop a predictive model of human fertility, from which one may statistically infer from cross-sectional associations the effectiveness of family planning programs. The author presents some tentative statistical evidence on the association between regional birth rates and family planning program activity in Taiwan to illustrate a different approach in evaluating program effectiveness. The overall effects of the program as well as the mix of personnel are evaluated in terms of cost-effectiveness. This study confirms the central importance of a behavioral model, even where this model can only take account of a few of the characteristics of the parents' environment which might be responsible for differences in desired birth rates. 73 pp. Ref.