For one, it is plagued by selection bias - but it never claimed otherwise: it's a survey of people who frequent SO and are likely to take a survey ;)