Old Research Indicates Pair Programming Doesn't Save Time
Blogger Steven Thomas wanted to base his adoption or non-adoption of pair programming on research. The best research on the topic that he found was from 6 years ago in Norway. While the results were underwhelming, I had some thoughts on the research.
First, pair programming's biggest draw is improvement in quality, especially for complex projects where two sets of eyes are better than one. This Norwegian research used simple projects and complex projects. The complex projects had a higher success rate, even if it took longer sometimes. It wasn't always helpful for simple projects, but that's ok.
Right now, this research is some of the best we have on Pair Programming, but it's not enough of an efficiency booster for Thomas to start using it. He's happy to look at any other research on professional programmers if readers have any to offer.