Several problems that an evaluator of an experimental college has to face and resolve are considered. These include: (1) the purpose of the evaluation and the role of the evaluator, (2) the utility of a research design model, (3) the specificity of program goals before the evaluation is planned, and (4) the appropriateness of standardized or local assessment instruments. Evaluation problems of the Centennial Education Program at the University of Nebraska, which included 125 college freshmen and 50 upperclassmen, are discussed. Results of the evaluation revealed some clear differences in student development between those in the experimental college and those not in the program. Students in the program were found to have closer relationships to faculty and other students, read more non-required books, showed a greater preference for participatory academic activities, showed greater interest in reflective thought, an increased fondness for novelty, and were less practical and materialistic than their counterparts in the control group. No significant difference on any index of academic achievement or ability was noted. Random assignment of students was made in establishing the experimental and control groups. The use of a research model for assessment provided reliable indices of the effectiveness of the experimental program. (AE)