This study determined the effects of two data collection techniques on response rate, response bias, response validity and respondent evaluation. Administration of a computer survey was compared to a machine readable paper survey. Cover letters were sent to 249 college sophomores who were randomly selected and proportionally stratified by intended major. They were alternately assigned either a computer-assisted survey or an optically-scanned paper and pencil survey. Both questionnaires contained 31 identical items. After eliminating four nondeliverable surveys, the response rate was 29 percent for the computer-group and 49 percent for the scanner-group. There was no difference in response rates between applied and non-applied majors. Significantly more computer-group members had had previous computer experience. There was no difference in overall satisfaction means between groups. Students in the scanner-group were more likely to overestimate their actual grade point average. In evaluating the survey, both groups considered the survey convenient. Computer-group members were more likely to recommend it to a friend and felt that the survey was too short. From this study it is concluded that the practical and psychological advantages of administering an attitude survey by computer outweigh the disadvantages associated with its lower response rate. (Author/PN)