Seventy-four preservice teachers in home economics and social studies were videotaped during a 10-minute microteaching session prior to and after training. A training program in nonverbal behaviors related to encouraging classroom interaction was developed. Three training approaches to this program were constructed: programatic videotape, lecture-discussion utilizing transparencies from the videotape, and practice in a microteaching setting. A Structured-Observation Inventory was also constructed to record the frequency with which selected nonverbal behaviors were manifested prior to and after training during a 10-minute microteaching. Observers trained in the use of the inventory reached an acceptable level of agreement before rating the videotapes of the microteaching. A questionnaire to assess teachers' reactions to the training and microteaching was distributed after the last microteaching. Statistical analysis of the pre-post training data involved analysis of covariance methods by sex and training approach. Questionnaire data were tabulated and reported by percentages. There were no significant differences among training approaches on the frequency with which selected variables were manifested after training, and no significant differences between males and females. Groups trained by the programatic videotape and lecture-discussion perceived their training as more helpful than did those trained by practice. (Author)