An introductory discussion defined the term adversary as a conflict between the State and the citizens that is inherent at every stage of the criminal justice system and it cautions evaluators that although descended from a common ancestor, the criminal law and procedures of the States vary enormously. Criminal justice evaluators should, therefore, be cautious when generalizing from a study in one jurisdiction, or when assessing the implications of the adversary system in the locale they are studying. The examination of police procedures and of the trial system points out the various goals of the parties involved at each stage and the conflicts set up by the system that interfere with goal attainment. An example of these are search and seizure laws which prevent police from collecting evidence in the most efficient and effective manner possible. Many aspects of the system are nonadversary or cooperative in spite of the supposed adversary nature of the system. The defense and prosecuting attorneys, for instance, are two players in a game, each wanting to win the case for his client. However, each has other goals also -- to dispose of cases speedily, to make a profit, etc. To make the process proceed more smoothly and at a faster pace, attorneys actually cooperate in many ways, and much of the defense attorney's time is spent on negotiating pleas and avoiding trials. Evaluators must be aware of these kinds of things and not be quick to point out inefficencies in the system that divert from quick goal attainment. They must also recognize that the criminal justice system is a normative enterprise responsive to social and political pressures. An example of a successful research project is a 1977 study by Waldo and Chiricos of a work-release program. Finally, as participants in a system that uses adversary methods, evaluators should be prepared for an explicit contest with opposition experts. Explanatory notes, a list of relevant court cases, and 74 references are appended.