The interactive influence of student characteristics and eight measures of student-faculty relationships on the prediction of freshman year voluntary persistence/withdrawal decisions was studied. A questionnaire was sent to the total population of incoming freshmen at a large, independent, residential university in New York State. Usable responses from 1,457 students provided information on expectations of the college experience and selected background information. A second questionnaire on the reality of the college experience was sent to the students during the spring semester of the following year. Of the 1,457 students, usable responses were received from 773 freshmen. Student-faculty relationships were measured on eight scales; six of the scales were concerned with the frequency of student-faculty informal contact. The remaining two scales assessed students' perceptions of the quality and impact of their informal interactions with faculty and faculty concern for teaching and student development. Eight measures of student-faculty relationships explained a statistically significant additional 14.4 percent of the variance in persistence/withdrawal decisions, after 14 pre-enrollment characteristics and five measures of freshman year social and academic integration had been statistically controlled. (SW)