Though I've never been able to find the results and complete study methodology online (can anyone?), Mayer is a CS grad from Stanford, as well as head of the UI team at Google. I'm guessing she and her group know what they are doing.

I'd be more than willing to bet that they controlled for these variables you mention, else the test would be useless. Sure, it's right to be skeptical, but they do this for a living.

Given Google's stake in having accurate data, they probably bothered to do it correctly.

(And I am not usually a defender of Google.)