This experiment on teacher effectiveness focuses on the causal effects of teacher behavior within the classroom recitation. Each of four teachers used one of eight treatments in teaching an ecology course to eight sixth-grade groups. The treatments differed at two levels of teacher structuring, soliciting, and reacting. High structuring consisted of reviewing, stating objectives, outlining the lesson, indicating important points, and summarizing. Low structuring was the absence of these behaviors. High soliciting consisted of asking a large percentage of "thought" questions and waiting about three seconds or more after a student's response before calling on a second student. Low soliciting asked a low percentage of "thought" questions and waited only a short time after response. High reacting consisted of praising correct responses, providing reasons for wrong answers, and prompting. Low reacting consisted of using neutral feedback after correct response and not providing reasons for wrong answers. Tests were given to the students before and after the experiment. Results suggest that uncontrolled and unmeasured teacher behaviors and characteristics influenced student achievement and attitude. Analyses indicate that student perceptions mediated the effects of structuring and reacting. A follow-up study showed that merely reading the materials contributed substantially less to student achievement than the combination of reading and teaching. Statistical tables are included. Five appendixes outline the elements of the nine-lesson unit used in the experiment. (Author/JD)