Development and determination of the construct validity of a scale designed to assess aversion to computers are described. Based on a social learning model emphasizing efficacy, outcome, and reinforcement, the Computer AVersion Scale (CAVS) was developed for use with mental health clients who were high school age and older, as well as mental health professionals. The original CAVS item pool consisted of 45 true/false items. The CAVS produces four scores: (1) efficacy expectations (one's ability to perform behaviors required to operate a computer); (2) outcome expectations (one's knowledge of the required behaviors); (3) reinforcement expectations (whether outcomes produced by computer use meet one's goals); and (4) a total score, reflecting the cumulative effects of goals. A fourth summary score reflects the cumulative effects of the other three scores. Two groups of subjects were employed to reduce the 45-item CAVS to a 31-item instrument, which was then administered to 270 undergraduates enrolled in a large northeastern university. The Attitudes Toward Computers Scale was also administered to 78 subjects who completed the CAVS. The CAVS was also compared with other instruments designed to measure negative affect, including two instruments for use in diagnosing feelings about computers, as well as an instrument for assessing levels of computer experience. Results, which cover such variables as gender and educational background, indicate that the CAVS is a reliable and valid measure of computer aversion. Seven data tables, a 23-item list of references, and the CAVS are included. (TJH)