Student performance in CS1 with distributed pair programming

Abstract

Pair programming has been shown to provide many pedagogical benefits, particularly for students who are learning to program. One drawback with pair programming is its collocation requirement, which makes it difficult for some students to enjoy these benefits. We report on an experiment with a tool that allows students to pair program from separate locations. Our results indicate that students who experience the increased flexibility of distributed pairing perform as well in their introductory programming course as students who have to physically meet in order to pair. These students are also as confident as the collocated students. We also confirm earlier results reported by others that pair programming results in greater student confidence and increased student performance.