My own background is in vision science...
It is possible to make pretty objective measurements of things like contrast and reflectivity, though:
1) Nobody here has actually done so, as far as I recall
2) As has been pointed out, there may be between-batch and between-model differences, which means that without testing a reasonable number of units we should be cautious in making generalisations.
I think some people here have been reading too much philosophy of science though. Of course it is logically true that nothing is truly certain. But see the 'Hitchhiker's guide' for a fine example of where we might end up if we actually behaved as if that were true (the ruler of the universe who only believes in what he can see, which sometimes includes his cat). Ereaders would not have got invented, if their makers had questioned too many of their assumptions.