Group composition and its effect on female and male problem-solving in science education

Background: Cooperative learning may help students elaborate upon problem information through interpersonal discourse, and this may provoke a higher level of thinking. Interaction stimulates students to put forward and order their thoughts, and to understand the ideas or questions of their peer learner. However, partner gender is an important variable in cooperative learning. Previous research indicates that female students profit less than male students from mixed-gender cooperative learning in physics, especially where problem-solving is involved. Female and male students have different communication styles. For example, male students tend to give their opinions and explanations directly, while females tend to avoid presenting their opinion and are more likely to initiate cooperative problem-solving by asking questions. Purpose: The main aim of this study was to ascertain whether partner gender influences female students' learning to solve science problems and the role female communication style plays in the cooperative learning process. Sample: A total of 62 high schools students (31 female, 31 male) from three schools in the Netherlands participated in the study. Students were selected from three physics classes in grade 10, with a mean age of 15.6. Students came from various family backgrounds. Design and methods: An experiment was carried out to test the effect of group composition on female and male students' cooperative problem-solving in science. The students were randomly assigned to dyads and three research conditions: 15 mixed-gender pairs (MG); eight female-female pairs (FF) and eight male-male pairs (MM). Students were given training in how to solve a problem as a team, and how to complete the answer sheet. All students solved the same problems in four 50-minute sessions. In each session, students were asked to solve three new and moderately structured problems working together. Each dyad had a university student as an observer. The observer's task was to log the students' time on task and to document the interactions between the students. The observers did not interfere with the communication between the students during problem-solving. Results: Analyses of pre- and post-test performance revealed that female students in the MG condition did not learn to solve physics problems as well as male partners or as female students in all-female dyads. Analyses of interactive behaviours showed that female students in the MG condition devoted less time to actively seeking solutions and spent more time asking questions than their male partners. Conclusions: Difference in solution-seeking behaviour could explain an important part of the difference in problem-solving performance between the female and male students in this study. Female students in the all-female dyads did not differ in interactive behaviour or post-test performance from males. They had a more balanced interactive style than females in the mixed-gender dyads. Suggestions for further research are discussed. It would be interesting to examine if the findings of this study carried over to areas in which females are traditionally more comfortable, such as biology.