The purpose of the study was to engage college supervisors in analysis of the verbal interaction they employed in conferences with student teachers. Subjects in the study were 14 pairs of college supervisors-elementary student teachers in the Dept. of Elementary Education at a state college in New York during the spring semester of 1971. An experimental design was employed with the experimental group (n = 7 pairs): a) receiving training in analysis of verbal interaction, b) analyzing their own verbal interaction and, c) considering the results of the analysis of their verbal behavior. The control group (n = 7 pairs) did not received systematic training, did not engage in analysis of their verbal behavior, and did not have the opportunity to consider results of analysis. Data were secured by means of audio-taped conferences analyzed using the Blumberg Interaction Analysis System and a 47-item questionnaire constructed by the researchers and given to student teachers in the experimental and control groups after each of four sequential conferences. Results indicated the student teacher's perceptions were consistent in three areas: human relations, verbal process, and substance. Lower ratings were given to statements involving student output in problem solving, pursuing ways of collecting data pertinent to teacher behavior, being asked thought provoking questions, and being allowed time to think, discover and explore ideas. An extensive bibliography is included. (Author/MJM)