Answering your own question. Perverse incentives?

Question

I have answered my own question here (as community wiki) Consumer optimum in an economy with a continuum of commodities and I can see that some other users have done so as well.

I think it is good that the site allows this to be done since it allows the correct answer to be recorded if the person posting the question later find it themselves.

But do we need some kinds of norms around this practice?

My worry is that if it becomes common to answer your own question it would attract people who would like to use this site to propagate their own views or to pose challenge problems and reduce the genuine dialogue within the community.

Answer 1

Self answering is fine and should be encouraged. But questions have to be clear and on-topic by themselves. If there's a problem with self answering it's that the questions can often be unclear, or sometimes with a refute-this focus, which is therefore often unclear what it's asking. But if the question would otherwise be fine, then there's no problem at all and I don't think any extra guidelines are needed.

Answer 2

I think there are a couple of different ways of arriving at answering your own question. There is the case where you write the question, knowing full well that you will answer it, and then there is the case like yours, where you ask the question, but find out the answer along the way (or an extension to an answer).

The second case is always good - you shouldn't be put off answering just because it is your own question, and I wouldn't have begrudged you the rep had you not made it community wiki. When self answering happens like this the original question tends to be well written, because when you wrote it you were still trying to encourage an answer from someone else.

The risk comes in the planned self answer. In these cases it is very easy for the question to be badly written. You should also question what your motive for doing it is - if the motivation is to give "your answer because your answer is best" then maybe think again. A better motivation is "this question needs to be addressed on this site somewhere". Try ask the question hoping to get someone else to answer, and then if after a while no-one gives you a good answer, submit yours. This requires a little self sacrifice that you may end up never writing that answer, but it can help the site develop. (This is what I did here: When and to what extent should the money supply of a pegged currency be included in the "parent" currency?, and the answer is undoubtably better than I would have given).

Answer 3

We should always watch out for posts done to promote an agenda. I'm not sure that it's much worse if both question and answer are from the same person. It's quite possible to generate the same problem with just a question. I've already seen a question where someone asked how to do something. My personal response was that you shouldn't do that. I squelched it because the technical question was challenging even if the example was flawed. But someone else might give a less technical spin to a similar question and draw a close for being too opinion-based.

Also consider the sock puppet problem. If we do some kind of ban against self-answers, what's to stop someone malicious from making a second account and answering with that. At least if they self-answer, we can see that it's one person.

It's also worth noting that self-answerers face increased risk. It's comparatively easy to not realize that your question is unclear if you are the one answering it. Putting your question out for others can lead to better feedback that allows you to improve the question. Also, both question and answer can easily be downvoted if perceived as agenda setting. Or someone can downvote the question because they don't like the answer.

I'm not sure that this problem requires a general rule. While bright line rules are easier to enforce, they can also be unfair and punish legitimate posts. Take the example of someone who's struggling with a problem, posts a question, relaxes from the problem, and realizes the answer. Should that person wait another day or two to post? Can we expect a new member to understand the rules well enough to be aware of that one?

It might be better to take things on a case-by-case basis until such a time as it becomes common enough to demand a rule. If it never does, then it would seem that it would be just as easy to always take things on a case-by-case basis.

Answer 4

This has happened on skeptics.se, where we had one or two users who were indeed using self-answering to grandstand their own views; they tried to co-opt StackExchange as their own personal blog for ranting, knowing they would reach a much wider audience than if they just launched their own wordpress blog like any normal person would.

It's primarily the duty of the community to flag this sort of behaviour; and the duty of moderators to help clean it up.

There is an associated discussion that I expect we'll be having around the time we move from private to public beta. As we're still at expert-level private beta, it hasn't really come up much yet. But it will. That's the discussion about positive versus normative economics. But even positive economics can be shoehorned into a platform for normative statements.

Nevertheless, self-answering is not only allowed, it's encouraged: you'll see when you ask a question that there's the option to add an answer to your own question, at the time you ask it. It can be a great way to share knowledge. But when asking a question that you will answer yourself, it's harder to make it a good question; it needs to be a lot more than a mere introduction to your prepared answer; it needs to be a good question in its own right, independent of any answer.

Answer 5

On the programming StackExchange the convention is to answer your own question, if while waiting for somebody else to answer it, further work uncovers the solution. This happens a lot with programming/computer science questions - sometimes it seems all that you have to do is ask somebody else what the problem with your code is, to suddenly realise it yourself. However, that field usually has the advantage of clear cut questions/answers. There is also a small subset of questions where somebody found something very obscure, and posted the answer to record it.

I think with Economics however there´s a much greater danger of perverse incentives, as Jyotirmoy is pointing out, and the accompanying danger of degenerating into personal grandstanding is equally problematic, and that needs to be dealt with by policy and moderation as EnergyNumbers says. But I was wondering is it possible to impose a time limit before self-answering is allowed - i.e. you can´t answer your own question for at least X days after it has been posted? That might help deal with some of the problems around this.