the screen or keyboard and, of course, typing on the keyboard. Some pioneering methods of capturing nonverbal communication have been developed and carried out by Trowbridge, Bork, and colleagues ( Potter, 1982; Trowbridge & Bork, 1981; Trowbridge & Durnin, 1984a, 1984b). By recording students' behavior while working together (video plus audio) and all key pushes of the computer keyboard simultaneously, they can produce simultaneous images of students working together (with audio) and what appeared on the computer screen. Such methods have great promise for the study of peer interaction and learning of computer programming.
The present research also did not clarify the optimal role of the instructor. The best way to clarify this question is to conduct a randomized experiment that compares different styles of instructor interaction with students.
Finally, the present research did not investigate the mechanisms by which peer interaction influences learning. The introduction of this chapter suggests that peer interaction might produce cognitive restructuring in a variety of ways, for example, through conflict resolution, modeling, internalizing the group's strategies and solutions, and reorganizing and clarifying material during explanations. As Forman and Cazden ( 1985) pointed out, however, the studies proposing these mechanisms typically did not systematically observe peer interaction. Future research needs to systematically analyze peer interaction for evidence of conflict resolution, clarification and restructuring of material, and internalization of groupork (e.g., students who initially have difficulty solving problems adopt strategies proposed by the group and apply them successfully to later problems). Forman and Cazden ( 1985) have done pioneering work in this direction by tracing how problem-solving strategies that first appear during social interaction later become internalized. In pioneering work testing the model of sociocognitive conflict, Bearison et al. ( 1986) systematically examined the frequency of different kinds of disagreement between students (e.g., verbal, gestural, whether accompanied by justifications or explanations) as predictors of cognitive change. These kinds of systematic analyses of interaction between students are essential to help clarify how peer interaction influences learning.
In conclusion, current research shows how a social context for learning computer programming may be beneficial for achievement. The next steps are to better understand how learning takes place in these settings and, consequently, to refine the social context of instruction to maximize the learning of all participants.
Artzt, A. ( 1979). "Student teams in mathematics class". Mathematics Teacher, 72, 505-508.
Atwood, M. E., Jeffries, R., & Polson, P. G. ( 1980). "Studies in plan construction". 1: Analysis of an extended protocol (Tech. Rep. No. SAI-80-028-DEN). Englewood, CO: Science Applications.
Bandura, A. ( 1969). Principles of behavior modification. New York: Holt, Rinehart & Winston.