Evaluating Collaborative Practices in Acquiring Programming Skills: Findings of a Controlled Experiment

Abstract

[Context] Collaborative programming is achieved when two or more programmers develop software together. Pair Programming and Coding Dojo Randori are two increasingly adopted practices for collaborative programming. While the former encourages the collaboration in pairs, the latter promotes collaboration in groups. However, there is no broad understanding about the impact of these practices on the acquisition of programming skills. [Goal] In this study, we empirically compare the influence of both collaborative practices on two essential aspects of skill acquisition: motivation and learning. [Method] We conducted a controlled experiment with novice programmers applying solo programming and both collaborative practices to three different programming exercises using a crossed design. [Results] Our results showed that, while both practices outperformed solo programming, they also presented complementary benefits on acquiring programming skills. For instance, the programmers inserted less code anomalies in Coding Dojo Randori sessions than in Pair Programming sessions. On the other hand, the motivation was often considered to be stronger in the latter than in the former. [Conclusions] Our results suggest that the use of collaborative practices is particularly promising for acquiring programming skills, when programmers have little or no practical experience with software development.