Towards Understanding Communication Structure in Pair Programming

Pair Programming has often been reported to be beneficial in software projects. To better understand where these benefits come from we evaluate the aspect of intra-pair communication. Under the assumption that the benefits stem from the information being exchanged, it is important to analyze the types of information being communicated. Based on the Goal Question Metric method we derive a set of relevant metrics and apply them in an eXtreme Programming class room project. Data covering a total of 22.9 hours of intra-pair communication was collected. We found that only 7% of the conversations were off-topic (e.g. private), 11% about requirements, 14% about design, and 68% about implementation details (e.g. syntax). Accordingly, a great share of the information being exchanged in Pair Programming is on a low level of abstraction. These results represent a first data point on what kind of information is communicated to what extent during Pair Programming.