A preliminary model was developed for predicting mail survey response rates of college bound high school students using student demographic and academic ability data. Two samples of high school juniors and seniors who took the American College Testing (ACT) Assessment between October 1987 and February 1988 were selected. One sample (n=810) included students selecting agriculture or a related area as an academic major at the time they completed the ACT Assessment. The other sample (n=2,549) consisted of students who had selected college majors in non-agriculture related areas or were undecided about majors. The questionnaire contained a series of questions related to choice of major and to agricultural interests. It was hypothesized that students from Sample 1 would respond at a higher rate than would those in Sample 2. The overall response rate for both samples was 40.7%; 1,367 usable surveys were returned. There was no significant difference in response rate between Sample 1 and Sample 2. Responders were more likely to be females, with high ACT scores, good high school grades, and no intent to obtain financial aid to help pay for college. Results indicate that knowledge of students' demographic and academic characteristics can help survey researchers design mail surveys. A model is constructed, using effects for background characteristics, academic characteristics, personality characteristics, interest in survey content, and effects for survey characteristics. This model is proposed as a basis for future research. Six tables and one figure present study data. (SLD)