Exactly. A proper test would have a control group - perhaps Windows XP with a different theme.
Also the applications wouldn't be selected for a particular outcome, eg, running on Vista.
I think that misses the point of this MARKETING CAMPAIGN ;) They don't want a real study, they just want something SHINY to show people...
I'd love to see the results of a usability test between OSX, various flavors of Linux, and varios versions of Windows. Asking first time users to acheive simple tasks like sending and reading email, using the web, storing and sharing pictures and listening to music.