In a microsupervision workshop experiment designed to investigate group differences in acquiring an indirect style of supervisory conference behavior, 18 elementary school inservice teachers were randomly assigned to a video-modeling, written-modeling and performance-feedback treatment or non-treatment. Criterion performances were the frequency and quality of the supervisory conference behavior used on two separate microsupervision sessions and on five observational instruments. Results showed: (1) experimental microsupervision did not yield significantly higher nor more varied performance frequencies than the control group; (2) the reliability trial results for the conference and teaching category systems were considered highly satisfactory; and (3) a high positive and negative relationship between teaching and supervisory conference behavior in the experimental group was attained. A discussion is presented on the findings, and on the linear relationship between personal and professional characteristics of teachers. A stepwise multiple regression analysis of a teaching behavior by supervisory conference ratio factors indicated that the teaching behavior, "Data Recall," was significantly predicted by several supervision ratio factors. (Author/JD)