Reported is a study to ascertain if experienced science teachers would modify their teaching behavior after observing and studying an instructional model; to ascertain if changes in teacher behavior after this experience would be in the direction of the model; to determine the relative effectiveness of different displays of the instructional model on modification of teacher behavior; to assess the effect of the instructional models; and to determine if the use of television techniques in a microteaching format effected change in teacher behavior of the 27 experienced science teachers involved in the study. The teachers, assigned randomly to five treatment groups, each taught the same Earth Science Curriculum Project investigation to two different groups of students in a microteaching format. Video-tapes of pre- and post-lab discussions were made. A modified form of the Flanders' Interaction Analysis system was used to obtain IR scores. The findings showed that teachers did change their investigative teaching behavior after seeing an instructional model. Teachers viewing their own teaching performance and then viewing a video-tape instructional model had the most variability of change. The effect of the models in the pre- and post-lab sessions was similar. Television techniques in a microteaching format were found effective in changing teacher behavior. It was concluded that these techniques for changing teacher behavior have potential for any teacher training program. (Author/EB)