>Despite an impressively vast sample size, the survey received some flak for the inherent bias it produced by surveying its own, less-than-comprehensive user ba

>Despite an impressively vast sample size, the survey received some flak for the inherent bias it produced by surveying its own, less-than-comprehensive user base. Frankly, Stack Overflow is probably the most comprehensive developer site on the web. If you're a programmer of any stripe you can be expected by numbers to use it. It's not like there's a 'no girls' sign on the front, assuming their survey is representative is damn well justified.[0] But yes, forgetting selection bias in your analysis[1] is a bit silly on their part, and not even raising the hypothesis suggests whoever wrote the analysis wasn't thinking about it as deeply as they should have. Moreover attaching the Stack Overflow brand to conclusions that are based on not-super-rigorous analysis and could have a large impact on already quite plausibly disparately impacted groups is grossly irresponsible. [0]: Was this survey only open to members? If so that was a foolish decision, learning why people lurk and what demographics lurk your site as opposed to contribute would have avoided this exact accusation, as well as providing some very interesting business data for StackOverflow themselves. [1]: Comparing to base rates helps.