The goal of this panel was to develop the means to improve the reliability, validity, and utility of analyses of human interaction in learning settings. Research on interaction itself, as well as its association with the variables of student learning, student characteristics, setting characteristics, and teacher education were all concerns of this panel. The panel organized its work around needs in teaching as human interaction, and proposed three research approaches based on these needs. The panel's discussion focused on these approaches as well as the organization of specific programs and projects under them. The first approach concerned research on human interaction within education settings, the goal of which was the creation of knowledge and understanding of the process of teacher-pupil and pupil-pupil interactions during teaching and learning. Research on teacher education, which builds on the knowledge developed in research on teaching as human interaction, was the second approach. The third approach dealt with issues of methodology, instrumentation, and professional communications, and focused on the special difficulties of research on teaching as human interaction. Finally, the panel emphasized the need for taking into account the ways in which teachers will adapt research-based knowledge and teaching skills to their own classrooms, students, and styles. Such adaptations will determine the validity and utility of the knowledge to be derived from research on teaching as human interaction. (BD)