The 3-year research and development project described in this paper was conducted (1) to develop new knowledge about sex equity in classroom interactions, and (2) to develop successful techniques for reducing or eliminating sex-biased interaction. Two interventions were developed to train teachers in more equitable instructional behaviors. In the Washington, D.C., metropolitan area, teachers from fourth-, sixth-, and eighth-grade classrooms were trained according to an intervention based on a microteaching model. In New England, teachers of the same grades participated in a training intervention based on a collegial problem-solving model. A comparable group of teachers composed the control group. Overall, the sample consisted of 102 classrooms located in six school districts. All classrooms were observed for 45-minute periods of active interaction by raters trained in the Interactions for Sex Equity in Classroom Teaching Observation System (INTERSECT). Primary analysis of observational data focused on the nature of interaction patterns and the distribution of interaction between male and female students. In the second phase, differences in teacher interaction with boys and girls across treatment groups were examined. Statistically significant differences among conditions and between male and female students were consistent across all approaches to data analysis. (A summary describes selected findings about general characteristics of classroom interaction and about bias reflected in classroom interaction and treatment and control differences. The INTERSECT observation form is also appended.) (RH)