Oh, if only more people would realize that. The data are the data; the further away from the actual performance you move, whether it be through math or opinion, the further you are away from objective reality.
We see it all the time here. When you "normalize" data, you are taking out essential variables that cannot be normalized. It makes comparison easier, but it's also not real (it's hypothetical). Treating it like it is real is a type I error. But most people do not have a grounding in statistics or math and gobble up the values as if they are facts. Don't get me started on using ordinal measures as data.