It basically comes down to conclusions based on anecdotal evidence vs. conclusions based on properly tested and supported theories. Since the entire transit of an audio signal through an amplifier/receiver/preamp/what-have-you must obey the laws of physics -- incidentally, these "laws" are actually theories -- any two amplifiers measured to have linear frequency responses, THD below the threshhold of human hearing, etc. should produce the same output given the same input.
Many -- and dare I say most -- perceived differences between amplifiers having linear specs are a result of the brain's not-to-be-discounted capability of influencing perception. In order to do a proper comparison, you need to deprive the brain of as much extraneous information as possible. Simply knowing which device is playing will influence your perception. This is why a double-blind test is vital to getting any sort of definitive answer to whether or not sonic differences exist between any two (or more) amplifiers.
