Yes I have. To test properly you need to perform many trials for each individaul case, then find mean damages taking into consideration the standard deviation (all follow normal distribution) and then compare.
For example, regarding your "all mages have same magic critical rate" claim. Make one spellsinger, one sorceror and one spellhowler. Equip them with equal gear, buff them equally, all skill enchants equal, no clan skills etc, make everything identical. Then hit the same target with 100 spells (500 would be best case) with each mage and then compare how many times you made magic crit with each class.
Are you saying that you tested like this before making your claim?
Did you check for statistical significance? Do you even know what that is?