If the survey asked only about professional use, there would be of course the bias of developers being forced to use a specific operating system. But the survey also asked for personal use. Developers choose on their own what system they use for personal use. Of course there might be some availability issues impacting that decision. For example Apple closed ecosystem that forces developers to use MacOS to some extend. Or tooling that may be available only on Linux, or software that is running only on Windows or Mac. We can't rule that out. However, I don't see there would be any substantial bias in any direction for personal use that would change the outcome of 70 000 participant survey. Moreover, if developers use one system because it has better tooling than other ones - is it a bias, or is it a part of an informed decision proving exactly that developers choose one system for valid reasons? I think yes.