Computer-based problemsolving: The effects of group composition and social skills on a cognitive, joint action task

Abstract

Research concerning joint action for problem-solving on computer-based tasks in schools shows an increasing awareness that social context factors must be accounted for if cognitive development is to be effective. This study focuses on boys and girls working in single-sex or mixed-sex groups, and training for social skills as context factors and how these foci affect performance at the beginning and end of a school term (12 weeks). An experimental/observational methodology is adopted in preference to the I-P-I methodology which has dominated the literature in this area. Interactions most likely to achieve positive outcome/cognitive scores included elaborating discussion with little off-task talk. Training in a social skills programme helped increase outcome scores and positive interactions over the term. Exploration of sex differences did not provide the expected results. Males scored highest on the computer task at the start and end of term. Girls in mixed-sex groups scored better than girls working in girl-only groups. Girls who undertook social skills training showed the highest rate of improvement over the term. Finally, end of term results were co-varied by initial attainment on the computer task and showed that social skills training was a significant factor while downgrading sex differences to non-significance. Step-wise regression found initial attainment and social skills training provided the only significant contributions to the variance within final scores. Results suggest that teachers should not group pupils of low initial attainment together, that girls may work better in mixed-sex groups and that social skills training will benefit all pupils.