I actually agree with nearly all your examples above (except AZ which I would definitely choose over LH especially long haul despite, as you rightly point out, the rubbish back office work and OS which I would choose over LH on long haul faster than it takes Usain Bolt to run 1 metre
the new OS J is one my top European one alongside AZ), but less good than LX, OS, TK, OS, SU, BA, and AZ for me -- better than KL, AF, IB, SK, SN (note: I would choose SN long haul J over LH long haul J any day too), EI makes LH 'middle of the road' in my view rather than 11th world airline
My point is double. The first (and minor) point is not actually that larger airline will tend to score higher but rather than smaller airlines will be penalised because they are less well known.
The second (and more important) point is that for well-known psychological reasons which I won't go into here, when you give people a chance to evaluate multiple things, they tend to evaluate not just what they "really" know but also what they don't know directly but think they know (ie what they hold prejudice about). Your example of elections above is very pertinent here. My second point is thus that this is not a survey which will specifically reflect 'experience' of airlines but also, and to a large extent, 'prejudices' about airlines. In this sense, I have no doubt that many people who have never flown AZ (or not for many years) have answered questions about AZ and scored it badly because in their view, AZ 'personifies' the bad airline. Conversely, I wouldn't be surprised if part of the good score of LH was due to people "thinking" that LH is good, well-organised, and efficient, even when they have not tried it or not tried it recently. Let me give you examples. If you ask people to rank airlines on the basis of how good they are at being on time, I would bet anything that the majority of people would give very high scores to LH and LX, and very low scores to U2 or UX (Air Europa, a Spanish former charter airline) or Greek airlines while in fact, Greek airlines have long held the record of the most on time in Europe (both Aegean and Olympic!) for many years, and U2 and UX also have a much better on time performance than LH and a vastly superior one to LX which is one of the very poor players in that field (in fact, AF, IB, TK, KL and most other European airlines... except BA have much better on time record than LX!) but hey, when you think Switzerland, you think precision, organisation, watches, and timeliness, which is in fact very much true about the country... but not about the national airline. I'm sure as discussed in another thread on the BA forum, that many people also choose to answer questions about AF's P lounge saying it's rubbish because the cues they use is that the other AF lounges are not good (but what they ignore is that particular one is in fact excellent even though the J lounges are everything but!) or conversely many people would say that LHR has horrible delays (because we are used to circling around it) while in fact, LHR has much better on time performance than CDG, ARN or IST, let alone ORD or HKG.
So this is the sort of phenomenon which I think is largely triggered by the type of votes skytrax organises - echoing prejudice in addition to echoing experience, and that means that "valid data" and "error" which are hard to dissociate. That would be consistent with much of what we know in social science research, and it is a problem that we can very effectively avoid, but there are techniques involved to do that that I do not believe skytrax is using at all (nor are commercial surveys typically). My feeling is that this type of error will be maximal for summary judgements ('best airline' etc) and less problematic for specific judgements ('best premium economy', etc).