In general, answers on this site are supposed to have references. This was established early on by the top answers to the following Meta post: Since it isn't in the FAQ, are we as answerers expected to appeal to any of the various English authorities?
However, JSBձոգչ's answer also describes this as a policy that is limited to questions of "correctness," saying "Not every question will be addressable in one of these ways."
I'm wondering what standards we should have in general for answers to questions about dialect differences or accent variation. In these cases, there is no single correct answer; the questioner wants information about the variety of forms that exist. (Such questions can still be on-topic because even if there is no "correct" answer, the answers can be ranked based on how informative they are and how well they summarize the "big picture").
The Stack Exchange Meta post about "the 'back it up' principle" has some discussion of how personal information can in some cases be the basis of a useful answer. I would say that this is the case for questions about dialect, but I can also see why some people might argue that such information is too personalized to be of any use.
However, if the information is truly useless, there is no reason to retain it at all. I've seen some people leaving information like this as comments, or advising others to leave it as comments. I don't think that is a good course of action, for the following reasons:
It’s not really what comments are for. Comments are supposed to be ephemeral “notes” used to give feedback about improving the question. They aren’t supposed to try to answer the question. I know there's a tradition on this site of leaving partial answers as comments, but I don't see why we should encourage this. There are known disadvantages (for example, comments are not indexed by site search and they are more easily deleted).
If it's left as an answer, people can indicate how useful it is by voting up and down. While the initial post might only reflect one person's experience, the score will give an indication of how many other people have encountered this dialect variant. Comments don't work as well for this purpose: they can only be voted up, and it's easy for them to become completely hidden beneath other comments with more upvotes. If a moderator ends up moving the comments to a chat-room, all voting information is lost.
New users cannot comment. This means that a new user who wants to answer a question about dialects by sharing their personal experience will either just give up (not share any information) or will end up posting an answer. In the latter case, I don't see why it's necessary for the answer to be converted into a comment (a process that requires moderator intervention).
So, what do you think? Should answers like this be allowed, or prohibited? If they are prohibited, should we allow the same information to be left as a comment? What is the advantage of this?
Other posts that I think are relevant:
the general Stack Exchange guide on "How do I write a good answer?":
Any answer that gets the asker going in the right direction is helpful, but do try to mention any limitations, assumptions or simplifications in your answer. Brevity is acceptable, but fuller explanations are better.
This question was prompted by a discussion about the following answer to the question "Pronunciation of vowel in vague as [æ] instead of [eɪ]":
This is odd for me too. Somehow, for me, plague could take either the vowel of vague or of bag when a noun (I would use the "vague" vowel if placing emphasis on the word), but only the vowel of bag when used as a verb. The sounds are not merged for me, and I don't know where I picked this up. I'm from Rochester NY.
I'd appreciate feedback about this specific example, but I hope people address my general question also. I can see why this answer could be borderline even for someone who generally agrees with me, since the answer talks mainly about the poster's idiolect and doesn't say if the described pronunciation is commonly used by other people in the poster's region. I think positives of this answer are that it makes it clear what the source of the information is (personal introspection, which can actually be quite a good source for information about accents), it describes an unusual pronunciation pattern that it would be difficult to learn about from other sources, and it does say what region the poster comes from.