Using behavioral rating forms, college professors were covertly rated each day by pairs of students ostensibly enrolled in their classes. On specified days pairs of faculty observers introduced themselves to the instructor and rated him using the same forms. The purpose was to determine whether changes occurred in the instructor's behavior as a result of the intrusion of peers into the classroom. Economics, Mathematics, and Psychology classes were observed. Interrater reliabilities were very high. On peer observer days teacher-initiated talk significantly decreased and instructors received significantly higher student ratings than on baseline days. The results of the study have implications for several areas concerning faculty evaluation. (Author/GK)