The differing perspectives of developer, manager, and evaluator emphasize the themes of trust, priorities, communication, and reward. Trust seems to be a necessary part of evaluation research. At a minimum, program staff must trust the evaluator. Experience with the COPA Project suggests that, in the daily conduct of evaluations where program staff are part of the data collection process, priorities represent the biggest source of conflict. Program staff rightly place highest priority on client services, whereas the evaluator's top priority involves complete and valid data. One source of poor communication in the evaluation process is the differential training of program staff and the program evaluator. Clinicians are usually trained in positive regard, and evaluators are presumably trained in critical thinking. These differences in orientation may produce immediate differences in language and thinking. In addition, the evaluator may have a poor understanding of the clinical experience and how this affects data collection. Evaluators frequently promise program staff that the information collected will be useful to them for program improvement. Experience with the COPA Project suggests this is only partially true. From the program developer's perspective, the evaluation process was rewarding. However, the program manager did not find that evaluation feedback was directly useful for clinical purposes. 14 references