The general purposes of this project were to test microteaching procedures for teacher training and to study relations between the process of teaching and its outcome. The project made use of "cued" video models (a short peep being superimposed whenever the model demonstrated a goal-related behavior), videotape self-feedback with self-evaluation forms to focus student attention on the particular skill, and the teaching skills of silence and nonverbal behavior. A follow-up program for a few months after treatment is planned. Of all the variables tested, only locomotion consistently showed a main effect for the factor of modeling. The contrast was between the presence and nonpresence of models, the type of model making no difference. Cuing terminal behavior was no more effective than noncuing. (Included are appropriate tables and a list of references.) (JA)