The controversy over evaluation methods is based on concern about how to increase the chances that program administrators will base decisions at least in part on evaluation results. The dominant view in evaluation literature is that experimentation based on the hypothetico-deductive approach is crucial to the quality and the use of results. Federal agencies appear to be committed to the use of this approach, which is affected, however, by such problems as ethical neutrality and the absence of attention to the unintended consequences of a program. To resolve these and other problems, some researchers propose an active-reactive-adaptive approach relying on more subjective, nonexperimental methods. In this approach, the evaluator begins not by clarifying goals but rather by identifying key decisionmakers and information users and asking them what they would like to know about their program. Although many evaluators believe that these two basic approaches represent competing paradigms, finding a middle ground is possible. The alternatives are not mutually exclusive either in theory or in the minds of researchers designing evaluations. The two methods may be combined via synthesizing procedures or via searching for compromises between two alternatives. The research setting and the research subject will influence the choice of research methods and design. Among factors to be considered are the layers of administration, resources, program visibility, and ethical and legal factors. Overall, the active-reactive-adaptive approach can supplement the experimental approach without having to replace it. Thirty-four references are included.