To be fair, sampling from a very specific subset of programmers (ones who would participate in a code jam) to do this study just seems like it would introduce v

To be fair, sampling from a very specific subset of programmers (ones who would participate in a code jam) to do this study just seems like it would introduce various kinds if sampling bias. I would like to see someone try to figure out a better sampling method before I would even consider looking at the results of such a study.