A Lightweight Systematic Literature Review of Studies about the use of Pair Programming to Teach Introductory Programming

Background: Our research group is developing and documenting procedures for undertaking systematic literature reviews (SLRs) within the software engineering domain. A question that has arisen is whether the procedures are suitable for use by students over a relatively short period of time. A further question is related to the effectiveness of pair programming for undergraduate students. Aims: The aims of this research are twofold: to investigate the applicability of the SLR process within the constraints of a 13-week master’s level project and to aggregate evidence about the effectiveness of pair programming for teaching introductory programming. Methodology: To address the first aim a case study approach was taken with a single student applying the SLR methodology under the supervision of an expert reviewer (and member of academic staff). The process was adapted to fit the time available. For the second aim, a modified SLR method was used, based around an analysis of a random sample of the included studies. Results: The case study found that, with certain modifications to the process, it was possible to undertake an SLR within a limited time period and to produce valid results. As a novice researcher was undertaking the process, it was found that training was required in certain aspects in order to effectively undertake the review. In particular, the distinction between conference and journal publications and the relationship between publications and studies were sometimes unclear to the student. In terms of the results of the SLR, 28 publications reporting empirical studies of pair programming were selected for inclusion, of which nine publications were used for data extraction and analysis. The preliminary evidence from the review suggests that pair programming can have a positive impact on pass and retention rates as well as the students’ confidence and enjoyment of programming. However, the evidence also indicates that pair programming does not have a significant effect on the marks obtained for examinations and assignments. Conclusions: The preliminary results are positive, both for the use of pair programming in introductory undergraduate programming courses and for the use of the SLR process for master’s level project students. The evidence from the SLR is that pair programming can significantly improve the success and retention rates on programming modules, as well as student’s confidence in their work. In terms of the process, this study has demonstrated that it is possible for a student to follow guidelines for conducting SLRs and, with only minor modifications, deliver a valuable project report.