Without the two cameras in the exact same location at the exact moment, with the exact same lens, etc., all of which is physically impossible, you are always changing more than one variable at a time. IR has one of the best test methodologies available, certainly better than dpr and more relevant than DXO, IMHO.

At the end of the day, it's a data point. It don't proclaim the 7D as the king of all DSLRs on one test. Anyone is free to proclaim every test as poorly designed and thus no way to every prove any camera is better than any other camera (or any other physical device, for that matter). I tend to be more pragmatic and believe that this, plus a couple of other similar results, show a believable comparison with pretty high confidence.

Personally, I don't shoot sports or BIF, so don't really need the fps or increased pixels/'reach'/cropping ability, and haven't had focus issues with my 40D, shooting mostly static nature shots. So it will join the 5DII as a nice piece of technology to read about as I continue to focus on content.