One is tested on a 16MP imager, the other on a 12MP imager. Also they use imatest and don't specify if they are doing it in RAW or JPEG, in which case the JPEG engine plays as much role as the optics.
Looking through their lens reviews I immediately notice their undocumented test methods, their graphs with unlabeled scales, and the fact that they test with random camera bodies. All of those indicate you can safely ignore any graph, plot or conclusion drawn from their test data when compared to another lens. Don't even bother comparing one test of theirs to another - it would be total garbage. Now, their text about the lens and what not may still be useful. And for a given lens the graphs might help you figure out where it is sharpest, or how bad the corners are in a relative sense. But don't compare lenses between their reviews - they aren't set up to do that at all.
--
Ken W
See profile for equipment list