The study analyzed teacher-pupil interaction in the classes of nine experimental (E) and eight control (C) teachers and their 167 mentally retarded pupils (mean age = 10.2, mean IQ = 68.76) The E-teachers were given 32 training sessions in an experimental curriculum and the inductive teaching method. Classes were tape recorded for 1 day. Analysis indicated that the distribution of the E-teachers' questions was cognitive-memory (88%), evaluative thinking (5%), convergent thinking (4%), and divergent thinking (3%). Significantly more (.001 level) cognitive-memory questions were asked by the teachers than any of the other three types of questions. A comparison of the experimental and control groups demonstrated that E-teachers did not ask significantly more productive thinking or ask for significantly more evaluations after the students' incorrect answers than C-teachers, (2) there was less consistency for E-teachers' questions irrespective of subject matter areas, and (3) there were no significant differences between the E- and C-teachers' statements in the categories studied. Additional data on teacher and student characteristics are considered. Appendixes describe the Gallagher-Aschner Classification System and provide samples of teacher-pupil interactions demonstrating aspects of the inductive method. (JD)