OK, so maybe everyone is "totally biased and expects everything to work like their beloved Windows box" and maybe that's "wrong" in some sense. But which is more likely to succeed: trying to convince 95% of desktop computer users that their way of doing things is flawed and that they should learn a whole new system *or* closely emulating the existing system *while* introducing new features and functionality unobtrusively? One is almost certainly a losing proposition and thus is hardly worth considering for any serious advocate, *regardless* of the "righteousness" of the position. Being right doesn't make the world go around and it doesn't always get things done.
The point is the statistics/studies are not useless, in fact quite the opposite, they are just not useful *to everyone*.
If Linux does what you want then that's great and if you don't care if anyone else uses it then you needn't worry about such studies or their validity. As long as it works for you then hoorah!
This ignores the fact that the level of platform support is directly influences how much software is available, how many hardware devices have driver support, etc. - you clearly need some critical mass. But Linux does very well even as a desktop OS despite the comparatively small installed user base.
In any case many people do not share the personally-focused perspective. They want Linux to succeed and to succeed big. Whether that just means "beating Micro$oft" (sic) or if they really think Linux could be a better desktop OS for the average user doesn't matter too much - in general the steps to get to the end goal are similar, at least for now: Make Linux a better *replacement for Windows*.
Why does a successful desktop OS for the average person need to replace Windows? Because Windows has ~95% market share and that's what everyone is used to, even if they don't like it. Those that really don't like it can move to a Mac if they want. So although Apple's numbers are rising, I think the still extreme dominance of the Windows platform says that there is *something* about it that makes it successful. *That* is what the Linux community needs to emulate if they hope to succeed in challenging Windows for the desktop.
Again not everyone shares that goal and if that's not you then this study isn't of interest. But it does have validity and could be useful to anyone trying to make Linux a better desktop solution. Arguments about whether a GUI(s) tested in this particular case are "standard" or "the best one" only lend weight to the argument that Linux just isn't ready for the average user - 99% people don't want to have to choose.