76
$\begingroup$

There was recently a (now-locked) thread about the prevalence of "do my homework" questions on math.se. I want to set aside for now the question about what to do about people posting homework questions -- it's clear the topic is controversial and there are no easy answers. Instead I want to focus on a related question that perhaps we can all get behind: how do we encourage people to answer the harder, non-elementary questions, that all too often go unseen and unanswered?

There are currently three mechanisms for promoting questions, all with flaws.

  1. Upvotes. Causes a question to appear/stay on the main page, and signals to other users that the question is likely to be interesting, or at least well-written and well-formulated. The problem with upvotes is that they are too democratic: easy questions that everybody can understand, and "soft" questions, tend to get a lot more votes than interesting, harder questions since new users are more likely to click into and upvote the easier questions.

  2. Favorite. This marks the question as interest to you, and a question with lots of favorites tends to positively signal the quality of the question, but nothing about favoriting a question directly promotes the question to others.

  3. Bounty. If you put a bounty out on a question, you promote the question and clearly indicate that you think the question in interesting. But the bounty system suffers from several unfortunate design decisions: first, you can't put a bounty on a question right away, meaning you have to remember to come back to the question after several days if you want to promote it. Second, putting a bounty on a question costs reputation, and many users are reluctant to spend any of their Internet points on questions by others.

It would be interesting to have a system that merges the best aspects of each of the above: I'll call it "endorsing" a question. I don't know the exact details and numbers that would work best, but what I have in mind is that

  1. Established users have greater power to endorse questions than new users. Maybe each user can endorse X questions per day, where X is $\frac{\textrm{reputation}}{1000}$ or some such formula; or maybe endorsement is simply a privilege restricted to users above a certain reputation cap.

  2. Endorsement promotes a question to the main site, in the same way as upvotes do. Perhaps each endorsement counts as Y votes for any algorithm used to show "new" or "hot" questions, etc.

  3. Endorsing a question rewards users for answering that question: an answer to the question that is accepted earns Z extra reputation per endorsement, as if each endorser had placed a Z-reputation bounty on the question (but without costing the endorser any reputation).

TL;DR this gives established users a way to say, "this question is interesting and deep, and even though I cannot answer it myself, I would like to promote the question and reward others for answering it," without the drawbacks of the present bounty system.

$\endgroup$
  • 7
    $\begingroup$ Upvoting a question does not cause it to appear on the main page, nor to stay there. $\endgroup$ – Tobias Kildetoft Apr 21 '14 at 19:58
  • 24
    $\begingroup$ +1 The problem is serious, and I'm all for the idea of promoting the more interesting questions here. Some users, indeed, give bounties to interesting questions/answers (IIRC Jonas Meyer and Byron Schmuland have been particularly generous). I have just started rewarding particularly delightful answers myself, but the tags I follow aren't too numerous. Many of us try to vote the way you describe, but as we are severely outnumbered the outlook is kinda grim. $\endgroup$ – Jyrki Lahtonen Apr 21 '14 at 20:40
  • 3
    $\begingroup$ Upvotes do not "bump." Upvotes are irrelevant when using the "active" tab, which most people mean by "the main page" (except that they could prevent a too negative score that makes questions dissapear, but this is more theoretical in the present context), they are however (albeit in a more complicated way than bumping) relevant for the "hot" tab. $\endgroup$ – quid Apr 21 '14 at 20:42
  • 4
    $\begingroup$ @quid Upvotes are also relevant to the "interesting" tab, described here and used as "main page" on Stack Overflow, but not here. $\endgroup$ – user127096 Apr 21 '14 at 21:00
  • 2
    $\begingroup$ I don't think that I can endorse 140 questions every day. I don't know if I look at 140 questions on a daily average. (Okay, maybe I am, but crap that's a lot of time wasted on MSE...) $\endgroup$ – Asaf Karagila Apr 21 '14 at 21:37
  • 6
    $\begingroup$ Would there be a way for users, particularly those with more rep, to rate the difficultly of a question? Then there could be 2-3 tabs filtering by difficulty. $\endgroup$ – abnry Apr 21 '14 at 23:08
  • 5
    $\begingroup$ The problem is that reputation doesn't reflect being knowledgeable. It's perfectly reasonable for someone who knows very little to get well above 10k points, and sometimes even more. Moreover, should I be the judge of any question about integration? Not at all. Should some of the high reputation users judge the difficulty and quality of questions in set theory? Not all of them, no. So you need a better mechanism, and not one which will eventually deteriorate as the number of high rep. users grow. $\endgroup$ – Asaf Karagila Apr 22 '14 at 0:03
  • 4
    $\begingroup$ @AsafKaragila Reputation is an imperfect measure but at least it is positively correlated with being knowledgeable. A more sophisticated system that looks at your tags is an interesting suggestion. $\endgroup$ – user7530 Apr 22 '14 at 2:29
  • 4
    $\begingroup$ Another thing, which is possible to do within current framework, is creating a chatroom and post interesting questions there. Owner of a chatroom can decide, who can add content to a chatroom. Nobody prevents a group of user to say: Let's create a chatroom and post there links to interesting questions from algebraic topology (or some other area we are familiar with), or even questions from any area. The question is whether other users would look at the questions posted there - the selection would have to be good to attract other users. $\endgroup$ – Martin Sleziak Apr 22 '14 at 5:17
  • 8
    $\begingroup$ It is possible that a recurring feature in a math.SE community blog could serve a similar purpose. Some other community blogs have (or had) "Question of the Week"-style blog posts (e.g., arqade.SE, security.SE, scifi.SE). Generally the community votes (on meta) for QOTW, and then a blog post related to it written up and posted. $\endgroup$ – user642796 Apr 22 '14 at 6:03
  • 17
    $\begingroup$ It seems to me that the problem is that Math.SE is split into two streams: one is interesting mathematical discussion that is too low-level for Mathoverflow, and the other is 'solve my problem for me'. Perhaps they could be separated out... $\endgroup$ – Marius Kempe Apr 22 '14 at 16:23
  • 14
    $\begingroup$ Please don't hesitate to suggest exceptionally bounty-able questions to me in chat. I like to make it rain. $\endgroup$ – Alexander Gruber Apr 22 '14 at 22:37
  • 6
    $\begingroup$ Has the site ever done a contest? Several SE sites have run contests to encourage a certain type of question(s) or behavior. $\endgroup$ – Dynamic Apr 26 '14 at 17:53
  • 8
    $\begingroup$ If StackExchange doesn't implement better filtering in the software I for one would be happy to join a new StackExchange site in between this one and MathOverflow. $\endgroup$ – user2055 Apr 30 '14 at 16:34
  • 3
    $\begingroup$ Indeed, I want to add Alexander Gruber to the list of generous benefactors. $\endgroup$ – Jyrki Lahtonen May 22 '14 at 9:43

You must log in to answer this question.

Browse other questions tagged .