Re:If only it were true. (Score 2, Informative)

A sample size of 400 is actually a pretty good sample. Assuming a global developer population of 10,000,000, a sample size of 400 would give you a margin of error of 5%.

With a population of the same size, in order to get a margin of error of 20%, you'd need a sample size of 25. If the North American developer population is even a third of this, you'd get around 5% margin of error as well, so I'd say that statistically speaking that 400 is a pretty good sample size.

These calculations were taken from http://americanresearchgroup.com/moe.htmlAmerican Research Group's Margin of Error Calculator