What about asking whether the StackOverflow user base is an unbiased sample of the population of programmers? Here's a possible theory explaining both the great

What about asking whether the StackOverflow user base is an unbiased sample of the population of programmers? Here's a possible theory explaining both the greater competence of old programmers and their rarity, based on the StackOverflow data (please, before tearing me apart, understand that I don't necessarily believe in this story, but it is arguably as plausible as the whole old programmers flee the industry theory): - Older programmers generally don't care or are not aware of stackOverflow and modern technologies in general. That would explain their rarity on SO. - Given the first point, it takes an older programmer an unusual dedication for his craft and curiosity to get interested in SO and new technologies. So this sample of old programmers really is some sort of "elite" of old programmers. Hence the impression that they are better. This is the problem when using proxies for populations: the accuracy of the conclusions are upper bounded by the accuracy of the proxy. In this particular case, the only reasonable conclusions that can be drawn from the study are: - Old programmers are rare on SO. - Old programmers on SO tend to be more knowledgeable than young programmers on SO.