Stack Exchange has recently rolled out the new user indicator. Given that this comes with a reminder on the comments and answer sections, this is presumably intended to get people to be nicer to newer contributors in their answers and comments. I also take this as an indicator that Stack Exchange is looking to improve retention of new users.

So are rude comments actually the problem for new users? I posit that it's actually worse, from a user retention perspective, to down-vote a new user and not say anything than to give a rude yet informative comment. Even something as rudimentary as pointing them to the FAQ or saying that opinion questions are out of scope gives them feedback that they can use to improve rather than telling them that they suck with no other information (down-votes with no comments). Additionally, downvotes can snowball. When you read answers with even a single downvote, you likely come in with the bias that the answer is bad, and are therefore less likely to read it in a favorable light.

I'm not proposing requiring a comment with every downvote. Instead, I'd like to consider something much more specific: requiring a comment when casting the first downvote on a post by a new user.

Do the data support this idea?

Ideal data would be numbers on the following three things:

  1. The number of users whose first post was downvoted who then made additional posts over a week later

  2. The users in (1) with the additional condition that there was a comment on their first post

  3. The fraction of downvoted posts that have comments

EDIT: I'm looking for data not opinions. This is not yet a feature request. I'm already aware that people don't like the idea of mandatory comments, as I did my research before asking. I'm looking for whether or not there is data on user behavior on stackexchange if possible or a related site if not.

  • 4
    Downvotes do not mean “you suck”, they mean “This question [or answer] does not show any research effort; it is unclear or not useful”, as the tool-tip says. Down-votes are directed towards the post, not the user. Sep 12 '18 at 13:33
  • 3
    How would a required comment explaining the first downvote solve the problems that come with requiring to explain every downvote? In other words, why is your proposition better than the many propositions before? Sep 12 '18 at 13:36
  • 2
    @Xufox: You know that and I know that, but as we've seen time and time again new users (and some veterans) take downvotes too personally. I don't like the proposal here, but I think some instant "Why are people downvoting this?" feedback to a new user might be helpful.
    – ale
    Sep 12 '18 at 13:45
  • 1
    "requiring a comment when casting the first downvote on a post by a new user." - Now new users won't just get a downvote, but a prefab comment (probably handled by a userscript at some point) that reads "Bad question, visit idownvotedbecau.se to find out more" as well. I can't imagine this backfiring in any way shape or form.
    – Magisch
    Sep 12 '18 at 16:00
  • @Zags: What kind of data are you looking for here? Are you asking if someone has done some kind of study or experiment? And to what end is this inquiry being made? Sep 13 '18 at 0:20
  • 4
    Well, maybe, but the comment should be anonymous, (except to mods/CM). I no longer comment on new accounts in order to avoid personal backfire and getting dragged into a street fight I cannot possibly win. Users who open new accounts can safely say anything they like and, when they get suspended/banned, just open another account. Those with non-trivial rep cannot. I do not want to engage in any way with bad questions from new accounts; experience has shown, over and over again, that it's not worth the effort. A downvote means no r/a flag, no emailed moans and no blog & titter risk. Sep 13 '18 at 2:01

requiring a comment when casting the first downvote on a post by a new user.

This may turn out not too useful:

  1. No vote (up- or down-vote) comes without an explanation, as you'll notice when you hover the mouse over the respective buttons.
  2. Still requiring such a first comment may have the effect that a down-voter just writes down what that very same tooltip says. And this would even still be a helpful comment. A sufficient first comment might also be "Hi and welcome to ....SE!" (and even be very welcoming). Also, what about posting even nonsense like "vefjebvje" and deleting the comment right afterwards?
  3. Comments are noise and the info is often already to be found (e. g. when you take the tour, have a look at the help center (what's on- or off-topic etc.) etc.), so why duplicating it?
  4. As noted by @Xufox in a comment, why should an explained first down-vote be especially useful to the OP, but not all the following ones? Or is the hidden purpose to make people withhold casting a down-vote because the first one would have to write a comment, too?
  5. Votes are anonymous and this violates that core principle of SE.

So, is this a good idea?

I don't think so.

This post is an answer to the first revision of this question.

  • Do you have any data of the variety I requested or is this just an opinion?
    – Zags
    Sep 13 '18 at 14:49
  • @Zags What sort of data are you looking for? You can't just demand it without giving some idea of what, exactly, you're looking for.
    – fbueckert
    Sep 13 '18 at 18:17
  • @fbueckert The question already has a list of data that I'm looking for
    – Zags
    Sep 13 '18 at 19:16
  • @Zags No, it doesn't. All it has is a vague list of criteria, worth no context on why you want it, or how it should be presented. You need to give us more to go off of.
    – fbueckert
    Sep 13 '18 at 21:03
  • @fbueckert There are three very precisely defined things I'm looking for data on with plenty of preamble to motivate the type of feature that could be warranted if the data support it. Is there something in particular I can clarify?
    – Zags
    Sep 13 '18 at 21:11
  • 1
    @Zags The fact that you edited in, "I'm looking for data not opinions" after the fact offers nothing for future readers to latch on to. It invalidates this answer, as you're providing no inkling of what would satisfy you. Are you going to edit it again after someone does provide data that you don't like, narrowing the confines again? Why bother answering if you're just going to move the goalposts yet again? You need to clarify what, exactly you want as an answer, and why the duplicate isn't good enough.
    – fbueckert
    Sep 13 '18 at 21:19
  • @fbueckert At the time this answer was posted, there were extremely clear goalposts for the data I'm looking for. Those have not changed. All I did was remove the part of the question that made it appear to be an opinion question. This answer provided an opinion with no quantitative data, and as such is not a satisfactory answer to the question. The rest of your remarks seem directed at my question rather than this answer and as such are inappropriate for the comments section of this answer.
    – Zags
    Sep 13 '18 at 22:09

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