Well, these are attempts to develop predictive models. The method is commendable, but the global result should not be treated as gospel as it often his. The authors seem well aware of that and I imagine this is going to be published next.
https://www.researchgate.net/public...ir_Preferred_Headphone_Sound_Quality_Profiles
Significance should be relatively low (compared to other fields with larger Ns) but that is probably the best we are going to get in this field. But this does not automatically means objective/subjective should be systematically opposed.
One can objectively say things like
- most people prefer that target.
- most people in segment A prefer another target.
- the segments are clearly different (or not).
- etc
A subjectivist can perfectly say, objectively
- it seems I do not belong to the overall majority.
- it seems I belong to another segment.
- etc
Where things fall apart is when the objectivist tells the subjectivist that he shouldn't like some response or the subjectivist claims his preferred response is some kind of absolute best response.
So there are, imho, two very different aspects in the objectivist/subjectivist debate
The first one is just when taste is the dominant issue (archetype: response curve). It is perfectly OK not to like what the majority has been proven to like. If that is not OK, I would suggest a blind registration test on this board and to forbid people who don't like the Harman curves from joining
The second one is when claims that can't be measured or blind tested are made (archetype: fancy cables). There, obviously, objectivism calls for zero tolerance, especially if those bogus claims are used in daylight robbery schemes.
PS: segmentation should be of great interest to manufacturers who, no doubt, would love to see their products well received by their target market. The recent archimago blind test hints that demographics play a significant role in preferences imho.