This paper describes a field study of the "Personalized System of Instruction" (PSI) as it was used in an undergraduate communications research course. On the basis of statistics from two years of the program's implementation and from student responses after completing the course, the study draws five conclusions. (1) Most students did not make optimal progress in passing the sequential unit mastery tests. (2) Most students found that they had difficulty in understanding or remembering course material, even in the self-paced format. (3) The PSI approach produced a lower grade distribution than did the traditional lecture-discussion method. (4) Many of the students were not satisfied with the PSI approach and considered that no significant learning had taken place. (5) Although some students seemed to learn well and were satisfied with the PSI method, the instructors could not tell if it was the method that helped or if these were merely good students. The paper concludes that PSI has many advantages as an instructional method but must be carefully adapted to fit the needs of particular teaching situations or course objectives. (RL)