Pair programming in CS1

In academic year 2005-06, the Bren School considered incorporating pair programming into CS1, primarily because of reports it increased students' satisfaction with the course and improved their performance in it. Though not denying its benefits, objectors asserted that certain obstacles doomed pair programming to failure and so was not worth undertaking. We refuted some of these assertions sufficiently to proceed with pair programming in CS1; evidence from that offering allowed us to refute the remainder (as does evidence from subsequent ones). We contend our findings apply to programming classes generally and will help convince objectors that pair programming is worth attempting.