Growing dissatisfaction with conventional evaluation approaches to dealing with complex social phenomena has led evaluators to search out alternative evaluation models. Exclusive reliance on the experimental model and the logic of statistical inference has been seriously attacked on philosophic, technical, and pragmatic grounds. This paper deals with existing problems in traditional evaluation and presents a rationale for exploring alternative approaches, along with a brief case study description of six applications of adversary evaluation using the judicial and debate forms. Some common features noted in the six adversary evaluation case studies were l) better communication between evaluators and decision makers, 2) greater attention to the formulation of key evaluation issues, and 3) increased concern for meta-evaluation. The authors conclude with seven unresolved issues in adversary evaluation. (Author)