When selecting research for this blog, I sometimes fall into the trap of looking primarily at larger studies. However, exploratory studies with a relatively small sample can certainly have value as well (see also here). At least, if you take the limitations into account. This is also true today with a study on group work. Previous research often looks at who speaks a lot or a little. But maybe it’s more about quality than quantity? After all, what exactly do students say? Do they ask questions? Do they develop ideas further? Or do they mainly agree with others? Or do they try to steer the conversation in a particular direction? And, maybe more importantly, how do the other students react? And what about gender differences in group work?
A new study by Kang and colleagues, published in Learning and Instruction, wants to help us to better understand this. The researchers were particularly interested in possible differences between boys and girls during group work. Therefore, they examined each student’s individual contributions and how conversations developed.
For their research, they followed (only) sixteen Finnish tenth-grade pupils, nine boys and seven girls. The pupils worked in four mixed groups on an open research assignment on sustainability. The pupils chose their own research topic and developed an approach together as a group. The researchers analysed the first two hours of the group discussions.
This resulted in 1,489 utterances. They then divided these into categories such as asking questions, agreeing, developing ideas, steering the conversation, and straying from the assignment.
The researchers then applied Transition Network Analysis. It enabled them to measure the frequency of a given contribution type and analyse transitions between contributions, i.e., what follows a given contribution. For instance, does a question lead to an explanatory contribution? Does a proposal develop? Or does the conversation get stuck at some point because comments become irrelevant to the task?
What did the researchers find?
To begin with, further developing ideas seemed central to the group discussions. A guiding contribution was followed by substantive elaboration in 48 per cent of cases. After agreement, this was 40 per cent, and after a question, 35 per cent. However, the researchers also found differences between boys and girls.
Among boys, 45 per cent of contributions involved further developing or explaining ideas, compared with 36 per cent among girls. Girls, by contrast, asked questions slightly more often (18 versus 14 per cent) and agreed with others slightly more often (18 versus 15 per cent). Boys, in turn, took the initiative to steer the conversation slightly more often.
But there was more to see here. When boys spoke in succession, the researchers observed longer sequences in which they further developed ideas. When two girls spoke in succession, questioning and agreement were more prominent. This does not mean girls did not develop ideas; rather, their contributions followed different patterns in this study.
Variation in conversation patterns was most evident in interactions between boys and girls. At the same time, once these mixed-gender interactions went off-task, they were more likely to remain off-task for several consecutive turns. In simple counts, girls went slightly off-task more frequently than boys. However, when looking at the sequence of contribution patterns, these were mostly related to interactions between boys and girls.
This suggests that differences in participation in group work are not necessarily solely the result of individual student characteristics. They can also arise from interaction between students. Here, too, the collective seems to influence individual behaviour.
A fairly significant limitation
Although 1,489 utterances might sound impressive, they are of course still utterances by only sixteen students, divided into four groups. And that is really too few to draw general conclusions about differences between boys and girls.
The researchers use various statistical techniques, including bootstrapping and permutation tests, to analyse the patterns found. But that does not eliminate the fundamental problem. Those 1,489 utterances are not 1,489 independent observations. The same students are always talking to and influencing each other.
Moreover, students chose their own groups. Existing friendships, interpersonal relationships, and the personalities of a few students may therefore have significantly influenced the results. You will likely have already noticed that, for example, in a group of four students, one very talkative student can be quite decisive in the conversation. That conversation, incidentally, was limited to the planning phase of one specific, fairly open assignment.
And I have to make a fuss about a completely different point. After all, developing more ideas isn’t necessarily the same as learning more. The researchers measured no learning gains. A student who primarily asks good questions can be at least as important to the quality of the conversation as someone who speaks at length.
Furthermore, straying from the assignment is not always negative. Especially when students are just starting a group assignment, conversations that don’t immediately focus on the content can help build relationships or reach agreements. Incidentally, the authors already mention this in their own work.
What can we learn from this, then?
To me, this is an interesting study, particularly because of the method used and the attention paid to conversational dynamics. However, for now, we must describe the conclusions about gender differences as exploratory, an invitation for further research.