Reported is a study designed to determine whether changes in verbal interaction patterns would be demonstrated in classrooms of teachers who received observation training and used it, using a structured checklist of classroom interaction. Twenty-six teachers, from different grade levels, were divided evenly into Experimental (E) and Control (C) groups. Teachers in the E group received three days' training in the use of an observation sign system checklist and subsequently made several 25-minute classroom observations in a period of six days. Teachers in the C group conducted their classes without the training or observation experiences. Data obtained from tape recordings were analyzed and coded, using Flanders' Interaction Analysis System. Ten categories of verbal behavior were identified. Twenty indices of teacher influence and pupil response were used to measure changes in the verbal interaction patterns of the teachers. Pre- and post-treatment means were compared and subjected to an analysis of covariance to test the significance of differences found. Teachers in the E group demonstrated significant changes in their verbal behaviors. Pupil participation increased significantly in the classes of the specially trained teachers (E group). Evidence was obtained that teachers can modify their own classroom verbal behavior patterns. (Author/EB)