Motivation and Objectives.

Group work is an integral component of many job descriptions in the computing profession. To prepare students for this, and also to be able to scale course offerings, many instructors require students to work on homework assignments in small groups. The importance of group work has contributed to an increasing interest of CSE researchers in regulatory strategies and processes occuring during collaborative learning situations. However, so far only limited work has been done on the concept of social regulation, or more precisely co-regulation and socially shared regulation. In this paper, we extend the nomological network of regulation in CS1 group work by incorporating the students’ self-efficacy and frequency of group work participation. We also report on barriers to groups work participation reported by the students.

Methods.

We conducted a mixed-methods study to understand barriers, effects, and affordances of group work in an introductory programming class. For this, we assessed three different types of regulation - self-regulation, socially shared regulation, and co-regulation - as well as general self-efficacy at three times during the semester. In addition, we collected data on the frequency of participation in group work and, at the end of the semester, the group’s collective performance. We complement this quantitative data by qualitative data on reasons for not - or not frequently - participating in group work.

Results.

Extending previous results on regulation in CS1 group work, we found self-efficacy to moderate the correlation between self-regulation and group performance. We also found self-efficacy to be correlated with self-regulation and co-regulation, but not with socially shared regulation. Moreover, we found that students who reported to frequently, but not exclusively work in groups exhibited the highest levels of regulation. The analysis of the qualitative data revealed a broad spectrum of reasons, some of which are specific to first-year courses while others are connected to perceived self-efficacy.

Discussion.

Our results indicate that the nomological network of regulation in CS1 group work should be extended to also include self-efficacy, even more so because it is essential in explaining the relationship between regulation and performance. Given that algorithms for optimizing group composition usually require data about prior performance and learner characteristics, both of which are not available at the beginning of the first semester, our results also suggest to not enforce working exclusively in groups in similar contexts.