It ruins the comparison. In the Nikon Lineup you have a D3000 selected in comparison to the D3x, seriously? But wait! There's a problem there... you can't select the D3000 for that lens, but it is showing the D7000 score anyway. What's up with that??
The scores are skewed by the fact that you're comparing a crop sensor to a full frame sensor. If you select these lens and compare them on the same body the scores are a bit different. A great example is on the Nikon AF-S Nikkor 16-35mm f/4G ED VR. You stick that on a crop sensor and the "Strong vignetting" becomes less problematic than the Tokina on the same body. If you put both of those lenses on the D3x (that the Nikon was installed on) The Tokina would be crap (considering it's a DX lens), not only would there be vignetting, at the widest zoom level you'd start to get black edges.
So what the heck you guys? I mean, I love Tokina, but this test is a bit flawed, and despite coming to what would have been the right conclusion anyway, the sloppy comparison invalidate that conclusion.