This study compared results of measuring student attitudes regarding race, gender, and religious climate using pencil-and-paper administration in a group setting with computer-administrated measurement over a local area network. Approximately 2,000 U.S. Air Force Academy cadets in February 1996 and 1,400 in November 1996 were administered the Cadet Social Climate Survey. The survey instrument contained demographic questions, Likert-scaled items regarding attitudes about school climate issues, and questions about personal experiences. The instrument used in November was modified to include items from the Balanced Inventory of Desirable Responding. Computer administration took less time for a student to complete, could be done in the privacy of the student's room, was easier to administer, and provided a ready data set for analysis. The comparison showed a significantly higher rate of return of valid responses using the computer version and found no significant differences in the pattern of responses between the two versions. However, significant differences were found on socially desirable responding in that computer administration appeared to encourage a socially desirable response set, probably due to concerns regarding anonymity. (Contains 14 references.) (JLS)