29
$\begingroup$

In October 2016, Grace noted that registration to ask a question had been implemented on math.stackexchange. I would like to know whether this has had any effect on our site.

$\endgroup$
  • 1
    $\begingroup$ Gerry, just curious: what motivates this question, curiosity? Or skepticism, or seeking transparency, or just because? Or any of many other reasons you may have had that I am oblivious to? Don't get me wrong; it's a fine question, and the answer informative. Just asked out of my own curiosity! $\endgroup$ – Namaste Nov 4 '17 at 22:35
  • 9
    $\begingroup$ @amWhy, if you look back linked question, you'll see the registration question raised a lot of discussion at one time. I was not involved in that discussion, but it seemed to me that, the situation having been resolved, there should be some follow-up. So I guess it was mostly a sense of a lack of closure. $\endgroup$ – Gerry Myerson Nov 5 '17 at 1:01
  • 1
    $\begingroup$ Well, ironically, I was just talking about (yesterday or the day before) the registration requirement. I knew it existed, but did not know that only users wanting to ask a question must register prior to asking. No such requirement for walk-in editors or answerers. Anyway, I was pleased to see your question! $\endgroup$ – Namaste Nov 5 '17 at 22:26
  • 2
    $\begingroup$ Hmmm, I wonder if absurd and spammy answers would be curbed by requiring answerers to register before answering, and wanna-be editors register before futzing with site content. A feature request in-the-making... $\endgroup$ – Namaste Nov 5 '17 at 22:29
  • 2
    $\begingroup$ @amWhy: I have always wondered why on earth they didn't require registration for both questions and answers. Totally agree that it will improve site quality. I have seen so many bad/crank answers by new users. $\endgroup$ – user21820 Nov 8 '17 at 13:33
  • 1
    $\begingroup$ @user21820 Some actions which increase average post quality also deter growth of the site (that is, number of future posts, quality or otherwise). Such action should not be undertaken lightly. $\endgroup$ – jpaugh Nov 11 '17 at 4:40
  • 1
    $\begingroup$ @jpaugh: According to Wikipedia, a 2008 study found that Yahoo! Answers is suboptimal for questions requiring factual answers and that the quality decreases as the number of users increases. So as far as I can tell from general Q&A sites, you're just helping to support my point, unless you have statistical evidence specific to SE. =) $\endgroup$ – user21820 Nov 11 '17 at 9:28
  • 1
    $\begingroup$ @user, speaking purely hypothetically, suppose that on taking no action a site would get posts of qualities 6, 5, 1, 1, and 1, while after taking action they just get the one post of quality 5. They have raised the average quality from 2.8 to 5, but they have lost the best post of the lot. So maybe it's not always good to increase the average quality of posts. $\endgroup$ – Gerry Myerson Nov 11 '17 at 10:20
  • $\begingroup$ @GerryMyerson: In my opinion any method of increasing average quality will also increase best quality in general. Just to follow along with your pure hypothesis, let P be the person who would have posted the quality 6 post in the original site. After a few years of watching the site quality degrade as with Yahoo Answers, P is likely to stop posting high quality answers as frequently as before, since the asker anyway picked the quality 5 post because he/she wasn't good enough to see that P's was better. [continued] $\endgroup$ – user21820 Nov 11 '17 at 10:42
  • $\begingroup$ [continued] Meanwhile, in the parallel site that had only permitted the quality 5 post to that particular question, P was too lazy to register and so did not post what would have been the quality 6 post, but after a few years of observing the site grow (slowly) with mostly high quality posts, P decides that his/her contributions would indeed be of great benefit and would effectively reach the intended audience through that site (unlike with other sites such as Yahoo Answers), and so eventually decides to register, thus further improving the average quality in the near future. $\endgroup$ – user21820 Nov 11 '17 at 10:45
  • 5
    $\begingroup$ P is not an immutable object, but a person who grows. P doesn't start out in life with the ability to post 6-quality questions; merely the potential. It isn't enough to increase the quality of the posts; one must also increase the quality of the poster, in order to have lasting value. The tension that exists between experienced and new users can actually be a good thing overall (it promotes quality posts & growth in the newer posters), as long as it doesn't discourage too many from either crowd from coming back. $\endgroup$ – jpaugh Nov 11 '17 at 15:18
25
$\begingroup$

The registration requirement had a big impact on the issues related to users losing access to their unregistered account (points 3,4,5 in the feature request).

It is now a rare exception that a user needs to be told about merging their accounts. Before the change it was a common occurrence that a user posted a question from an unregistered account, later created a new account (often but not always), but lost their cookies so that they didn't have access to their first account any more, and then posted clarifications to their questions in response to comments - or requests for elaboration on a point in an answer - as an answer from the new account.

It still happens that users post non-answers instead of editing their question or posting a comment, but that's much rarer than it used to be.

Also deletion attempts by defacing or blanking the questions are now rarer, though their number didn't decrease as much as the other issues. That's not surprising, since such attempts are often made when the asker can't delete the question themselves anyway due to answers posted. Also, many users use the android app which doesn't provide a delete link, and thus don't have an obvious way to delete their post even when they technically can.

I don't see any big effect on anything else, except of course a big increase of the number of newly registered accounts [per time unit] since then.

The post counts are generally a bit higher than in the corresponding month of the year before registration was required to ask questions (the December dip led a bit lower than usual in 2016). post counts from Jan 2016 to Oct 2017

It is of course possible that the growth would have been larger without the change, but the site growth had already slowed before.

$\endgroup$
  • 1
    $\begingroup$ On the last line, comparing questions asked to the various "visit" statistics, it seems to me there quite possibly is some minor damping effect on new questions (which is not necessarily a bad thing). Tangentially there is a 'shadow' in your screen shot in the top right corner. $\endgroup$ – quid Nov 4 '17 at 14:28
  • 1
    $\begingroup$ Yes, the number of visits and page views has increased way more than the number of posts. How much of that is due to the registration requirement and how much due to people finding their question already asked and answered before, I have no idea. $\endgroup$ – Daniel Fischer Nov 4 '17 at 14:58
  • $\begingroup$ "many users use the android app which doesn't provide a delete link..." Is that something the Stack Exchange powers-that-be can do something about? Has there been any discussion of this feature? $\endgroup$ – Gerry Myerson Nov 4 '17 at 21:12
  • 1
    $\begingroup$ @GerryMyerson There's a feature request on über Meta. It seems that implementing it isn't trivial if one wants to do it well [I'm not sure what the problems are exactly]. $\endgroup$ – Daniel Fischer Nov 4 '17 at 21:25
  • $\begingroup$ @DanielFischer Is there any current site in the SE network that requires wanna-be answerers, who haven't already registered when they asked a question, to register before being able to answer a question? If such a feature were implemented, I suspect the number of NAN posts would significantly decline, and it would help curb spam and/or offensive and/or nonsense posts. $\endgroup$ – Namaste Nov 6 '17 at 22:36
  • 1
    $\begingroup$ @amWhy As far as I know, answering is still possible as a guest everywhere. I guess requiring registration to answer would reduce the sheer nonsense posts, but actual spammers do already register quite often. It would probably reduce the number of "I have a related problem" posted as answers, but those aren't a huge problem anyway, they can be deleted quickly (and if it's a somewhat reasonable question, one comments that they should ask a new question). $\endgroup$ – Daniel Fischer Nov 7 '17 at 13:10

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .