It has been hypothesized that a student will develop more fully academically if grouped with students of similar ability. A study was undertaken at Harcum Junior College to assess the effect upon academic performance of the assignment of roommates with varying high school academic records. For investigation purposes students were identified as above average, average, and below average. Four groups were then formed: Group I paired above average students with average students; Group II, average with average; Group III, above average with below average; and Group IV, below average with below average. Analysis of results indicated that significantly higher levels of academic achievement could be attained by average and below average achievers by assigning them as roommates to above average academic achievers. (CK)