tea.mathoverflow.net - Discussion Feed (Minimum reputation for asking a question?)

Joseph O'Rourke comments on "Minimum reputation for asking a question?" (16194)

2011-09-22T15:54:17-07:00

Thanks, Anton, for providing data for this discussion! (That Vaughan Jones exchange is priceless!)

Anton Geraschenko comments on "Minimum reputation for asking a question?" (16191)

2011-09-22T10:40:18-07:00

Here's some data.

As of the last database dump, there were 9831 users who had posted on MO. Of those, 6236 posted a question before posting any answers. Of those first questions, 1604 of them were closed (about 26% versus the global average of about 11%).

So first questions account for about 3/5 of all closed questions. On the other hand, if you ignore users who start out by asking a question that gets closed, it's still true that most users first start using MO by asking a question. Offhand, I happen to remember that this includes desirable-but-not-very-active users Michael Freedman and Vaughan Jones. I think this argues strongly against a minimum reputation for asking questions.

Harry Gindi comments on "Minimum reputation for asking a question?" (16189)

2011-09-22T09:35:11-07:00

I am opposed to such a policy, since it would mess up the question-asking economy. Some of the best answers would never have been submitted if some questions by people with no answers had never been asked.

Such a decision would slowly kill the site, since it would create an imbalance in the number of people with research-level questions and those with research-level answers. Then, when those answerers find no questions to answer, they drift away and don't come back.

Anton Geraschenko comments on "Minimum reputation for asking a question?" (16188)

2011-09-22T09:22:22-07:00

@Joseph O'Rourke: As of the last database dump, there were 22531 questions, of which 2571 were closed, so the 10% figure is about right historically. This doesn't take into account deleted questions, but I don't think that should skew the numbers much.

Joseph O'Rourke comments on "Minimum reputation for asking a question?" (16176)

2011-09-21T07:45:24-07:00

That's an interesting, and surprising observation, that one-quarter of all questions are closed! There might be a large variance in this statistic, because just this moment, I see 10% closed. And, as someone mentioned in another thread, there is a September effect of many new users posting as they return to their universities.

quid comments on "Minimum reputation for asking a question?" (16175)

2011-09-21T07:43:48-07:00

a-fortiori: I am not sure I get your point. In my experience question that are clearly off-topic, in the sense of being on obviously too simple maths, typically get closed quickly without much or any discussion (the worst that can happen is that the questioner complains, but even this is rare). By contrast, those questions that generate real discussion are (in some sense by definition) not clearly off-topic, since if they were there'd be noone (or no critical mass) to argue in their favor.

I can vividly remember various somewhat heated discussion on question, but in very few cases would the questioner have been impeded from asking the question by your or any other of the suggested mechanisms (in some cases technically by yours, but the questioner typically could have answered something easily and I doubt they'd been convinced in the process not to ask the question).

To me all these suggested measures would at best remove something that I considered a very minor problem. Somebody just asked a group-theory excercise, so what? Costed me ten seconds to vote to close; if I'd been the first on the question perhaps a minute to type the closing message. If I would not want to deal with this at all I'd just read question that 'survive' for a couple of hours.

But the really more problematic situations would not be avoided, since the 'problem' is that the opinions among frequent users are split.

a-fortiori comments on "Minimum reputation for asking a question?" (16174)

2011-09-21T06:27:40-07:00

By browsing the question lists from some days ago like http://mathoverflow.net/questions?page=15&sort=newest , I have arrived at the estimate that about 1/4 of all questions are closed.

I agree that a new user who just wants to ask a question may be annoyed by having to answer some other question first. In my opinion, any other model like approval by moderators like on sci.math.research would be even better. I suggested using the reputation system because it seemed in the spirit of the software and would work automatically. One could have hoped that the voting system would have worked that way, but at least in my perception, the rating of questions (and answers, for that matter) is not a reliable indicator.

Finally, I think that constantly having discussions about off-topic questions on the first page hurts the image of MO more than a small barrier.

thierryzell comments on "Minimum reputation for asking a question?" (16165)

2011-09-20T14:02:57-07:00

Though I did feel the need to answer some questions before asking my first, I'm also not a big fan of forcing the issue: some users are question-only, some answers-only, and that's fine. It is true that among the question-only, a small minority are users who would be better off taking their questions elsewhere, but they form a small minority: so long as the moderators do not signal that they are overwhelmed by having to handle such cases, I don't see why we should draw up arbitrary rules that would make us appear less than welcoming to everybody else.

quid comments on "Minimum reputation for asking a question?" (16164)

2011-09-20T10:21:40-07:00

In brief and as far as I understand: the software is out of control of MO so it seems impossible to implement what you suggest. The current version is not changed anymore, a switch to a newer version is under discussion, however I doubt, from the little I understand on the general philosophy behind that software, that a change along the lines you suggest would be considered. A key-goal is to have as low a barrier to participation as possible. (There is something of a somewhat opposite mechanism, namely that individual questions can be 'protected' which means only users with non-minimal reputation can answer; this refers to the newer version of the software and is not available on MO.) So, on technical grounds, this change seems essentially impossible.

There were suggestion in a similar spirit see here and here, for example. In particular, see grp's suggestion of a two-tier system in the second thread [however, this is an idea feasible only assuming somebody would develop a custom version of the software for MO; and indeed came about in the context of such a discussuion].

Finally my personal opinion, assuming for the sake of argument it would be possible: I am not at all sure I'd be in favor of it (I say this as somebody not having asked a single question, but having given close to 100 answers). For example, a side-effect could be creating low-quality/redundant answer given for the sole purpose of 'finding' some initial up-vote. And, there are some users that asked numerous legitimate question before ever gving an answer, some never gave an answer so far. Or, I would also not have liked to somehow be forced to ask something first before being able to answer (could also suggest this, after all a low-quality answer can be more harmful than a low-quality question).

a-fortiori comments on "Minimum reputation for asking a question?" (16163)

2011-09-20T09:12:28-07:00

Has it ever been considered to require a very low reputation threshold (say, 2) for asking a question? Achieving this would be no problem for any reasonable user, but it would probably drastically reduce the number of off-topic questions since any new user would actually have to find out what this site is about (and at least contribute something before she chooses to ignore the rules and ask an off-topic question anyway).