A pilot project sought to learn how, at the time of admission, graduate schools of social work can begin to identify levels of potential and flair for, and interest in, doing research. Data of 353 students were obtained before these students had any classroom training in research. Data on 263 students were obtained after the course was completed. The training plan exploited a number of naturally occurring differences among the five schools that influenced student learning. These included such factors as the relative emphasis on various aspects of content, differences in methods of teaching, and in learning experiences, variations in methods of evaluating learning achieved, the time and place of research in the curriculum, and teacher characteristics. Data on the objectives, methods of teaching research, course content, and other details were obtained. Student characteristics and biographical data were also obtained. An instrument, the "Measurement of Attitudes and Research Knowledge" (MARK), assessed student attitude to, and knowledge of, research prior to the course and subsequent to its completion. Students also evaluated the teaching. The study found that various levels of potential for learning research can be identified, and learning achieved by students can be predicted. (Author/EK)