Investigated was the relationship between teacher and student performance on selected mathematical tasks. A measure of teacher effectiveness was obtained by comparing teacher and student performance on identical geometric tasks. The sample for this study consisted of teachers and their students from 19 fifth-grade classes, designated as either control or experimental. The six control treatment classes were presented topics in non-metric geometry by means of self-instructional reading materials. The 13 experimental treatment classes were presented the same topics by their teachers without the use of the reading materials. A pre- and post-test of selected geometric tasks was administered to students in the control and experimental classes. The same test was administered to the teachers of the experimental classes at the conclusion of the four instructional periods. Among the results reported were (1) the mean score for the experimental classes was significantly higher than for the control classes, and (2) there was a significant positive correlation between teacher test scores and class mean scores on the criterion test. (RP)