That's not a study of actual users, but the application of random metrics plus a bit of personal opinion (in the user friction section). It's as good or bad as the OP.

This is what I'd love to see, regardless of the results: http://www.nngroup.com/articles/windows-8-disappointing-usab...

As far as I know, the closest thing that's publicly available are customer satisfaction surveys.