I think it is useful to differentiate between matters of opinion and matters that do involve empirical differences but differences that we may not currently be able to accurately measure.
We do not currently measure amplifiers playing actual music and so do not really look at harmonic, IM, and other types of distortion products amplifiers actually produce making *music*. Furthermore, and much more importantly, we know only a very little bit of how to "weight" the results because we know rather little about how the ear/brain system actually works - what sorts of distortions make music unpleasant or unrealistic and what sorts really just don't matter all that much. (Of course we do know some of that, as Atmasphere & others have mentioned.)
Why some audio amplifier designs tend to sound better or more realistic may not be easy to describe empirically but it is at least not entirely a matter of opinion. Preference for dogs is. :)
(Ok, if you establish certain criteria, such a strength or cuteness, one could of course rate dogs quantitatively as well...)