Research assessed the effectiveness of the Utah Pupil/Teacher Self-Concept Program in changing specific teacher verbal behaviors related to self concept. Subjects were secondary school teachers in the Logan City and Cache County School Systems, Utah. An experimental pre-post control group design was employed with 12 teachers in the experimental group and 10 teachers in the control group. Pre-treatment observations were conducted in all teachers' classrooms for one full day each. Experimental teachers then participated in a 10-week training class that taught them to use the program's verbal behaviors. At the end of this class, post-treatment observations were conducted in all teachers' classrooms. The independent variable was the experimental treatment (the program course) while dependent variables were the use frequencies for specific teacher behaviors. Analyses of covariance (with the pre-scores as the covariate each time) were used to analyze the observation data. Eight significant differences in group performance were found in favor of the experimental group. Experimental teachers learned to recognize their individual use of judgmental behaviors and to replace those behaviors with appropriate descriptive strategies in a high incidence of pertinent interaction situations. (Author/JMK)