33

Down-voted questions seem to fall into two categories: fixable and unfixable. Some questions are just inappropriate, easily answerable with minimal search effort, or have other problems for which nothing will make them good questions. Others are, or can be, good and useful questions but were asked poorly. Downvoting is supposed to be an incentive to improve the question.

There is a consequence, though--the downvotes don't go away when the problem is fixed. They served their original purpose. However, now you have a perfectly good question that is stigmatized by the downvotes, discouraging viewers because the downvotes are then misleading. You can also have the situation where people spend time answering the question and before the answer gets any upvotes, the OP deletes the question due to the downvotes. This disincentivizes answering.

It seems like it would be beneficial to have two types of down votes for questions. The new one would be a conditional downvote. It would be available to be cast (voter's option), in instances where the voter's objection would go away if the question's problems were fixed.

There are a number of ways the mechanics could work:

  • The voter gets an inbox message that the question was edited. If they are satisfied with the change, they un-downvote.

  • There is some type of community review where conditional downvotes can be removed if the problems are deemed fixed (conditional downvotes would be used in those situations where the problem is easily recognized and the voter is OK to let the community decide).

Both approaches would benefit from some delay (maybe 12 hours after the first edit), since improvements to questions are often done in several rounds. These two approaches are not mutually exclusive and there are probably other potential mechanisms.

  • 20
    I've wondered why "•The voter gets an inbox message that the question was edited. If they are satisfied with the change, they un-downvote" doesn't exist for a while now. Seems like a simple, reasonable thing to me. :) – Ƭᴇcʜιᴇ007 Oct 10 '14 at 20:56
  • 1
    No, downvote now requires a reason or categorization. If an upvote doesn't require such a reason or categorization, then downvote process should not be made inefficient. I also do not want an e-mail telling me a post has been fixed. Leave it alone. – Sun Oct 12 '14 at 16:18
  • 3
    BTW, shouldn't this question be migrated to meta.se as it applies to the entire network? – user80551 Oct 19 '14 at 18:32
  • 1
    Related: Allow an edit to notify downvoters: “I think I've fixed the issue now – please check”.  Score: 525.  Fifteen answers expressing support (many offering alternative approaches and other improvements), the majority of which are by users with > 13K rep.  Too many positive comments to count, with two from ♦ people.  One critical (but heavily downvoted) answer from Jeff.  Status: declined. – Scott Oct 7 '15 at 1:22
  • So, does Jeff’s veto still have the power to overrule >500 votes? – Scott Oct 7 '15 at 1:23
22

I like the concept that you're raising, but not the implementation. I feel its a bit counterintuitive to have two types of downvotes - we've already had instances of people using Close votes as a "Super Downvote", we wouldn't want to make it even more confusing.

Instead - we should stick to just having 'downvote', with one of the additional features you mentioned: Being notified when a question you've downvoted has been edited. I feel this will adequately address the main problem- which is downvoters not knowing that the question is improved in order to adjust their vote accordingly.

  • 3
    There may be some semantics here. Sunk818 didn't want to be bothered by a message when the problem question has been edited. You would need a way to make the notification optional. Selecting that choice could define whether a downvote was "conditional". That is sort of another implementation option. Do an automatic notification on all downvotes and let users opt out from ever getting such messages. The trouble is, potentially fixable questions are a minority, so it would be burdensome getting lots of messages that unfixable questions were tweaked. Hence, opt in per question. – fixer1234 Oct 12 '14 at 16:49
  • 2
    @fixer1234- Sunk818 didn't want to be bothered by an email - s/he didn't say anything about a top-bar notification. As for the rest: 1) the notification could perhaps only appear once (even if you don't change your vote, you won't get notified again). 2) The functionality you describe for 'opt-in' already exists, as the 'favourite' option (the star under the vote count). Perhaps change its name to 'Subscribe to changes'? – Robotnik Oct 12 '14 at 23:58
  • 2
    (1) @fixer1234's feature-request refers to "an inbox message", and it was this that sunk818 objected to. (Perhaps sunk818 misunderstood the post, or perhaps s/he is just using the term "e-mail" sloppily.) (2) Stack Exchange does not currently notify subscribers when their "favorite" questions are modified. – G-Man Oct 13 '14 at 19:59
  • @G-man - not via the top bar, but check your user profile - the favorites tab shows activity in the form of a little number next to it. – Robotnik Oct 13 '14 at 21:51
11

Everything you said about questions applies equally to answers.  A few days ago I posted a one-line answer (to this question) because I was in a hurry.  By the time I got back to it, 15 hours later, it had accrued two downvotes – as far as I can tell, simply because it was too short (lacked explanation).  I proceed to expand on the answer, so it’s now over 50 lines long.  I engaged in several rounds of comments with the OP, who eventually accepted the answer.  But the downvotes remain, (presumably) because the downvoters never realized that I had updated the answer.

Since my initial answer was correct, but just code-without-explanation, I would hope that people would have used conditional downvotes if they had been available.

  • First your initial revision of the answer did need more explanation but you addressed it like you shoud so going forward you will get upvotes not downvotes – Ramhound Oct 10 '14 at 0:13
  • 2
    @Ramhound that unfortunately depends on whether the question is 'popular' enough that people will open the Q&A and thus upvote the answer. If the question is not popular, the answerer might be stuck with unfair downvotes forever. – pepoluan Jan 5 '15 at 8:01
  • @pepoluan - I wouldn't say those downvotes were unfair...For instance in G-Man's case he published a low quality answer and only fixed it because of the downvotes. – Ramhound Jan 5 '15 at 12:00
  • @Ramhound hehe... They were fair at the time given. I should've written, "no longer suitable/applicable". – pepoluan Jan 5 '15 at 12:02
5

I like your idea, but I would build out the process as follows:

  • Add a new downvote mechanic with a shaded arrow called conditional downvote with the hover text of this question can be improved to be more acceptable on this site. Conditional downvotes subtract 0.5 votes from the vote score of the question, and do not affect anyone's reputation. Fractional votes are not displayed and are always rounded up towards the positive, so if you have +2 upvotes, -1 regular downvote and -3 conditional downvotes, the score would show 0, because (+2-1-(3/2)) ==> -0.5 ==> rounds up to 0.
  • Add a new review queue for questions needing improvement. Any question that receives more than one conditional downvote from a user with 10k rep or higher gets added to the list of questions needing improvement, regardless of that question's total score. If <10k rep users cast conditional downvotes, a total of five votes (any mix of zero or one 10k vote and 4 or 5 non-10k users) are needed to add it to the questions needing improvement queue.
  • An accepted edit for a question needing improvement gives the editor +2 reputation, regardless of what their reputation is, or how many edits they've made. If the user is high enough rep that they can unilaterally make edits without going through the suggested edit queue, they can receive a maximum of +2 rep per question (even if they make a series of good edits on the same question), and if their edit is rolled back, the rep gain is reversed.
  • Add another new review queue for improved questions. Any question that was formerly on the questions needing improvement queue that has an accepted edit applied to it, is removed from the questions needing improvement queue and added to the improved questions queue.
  • Reviewers who look at questions in the improved questions queue will see the revision history of the question as well as the latest revision. They may upvote, downvote, or edit the question. If two separate users (regardless of rep) upvote or click a button stating "The edited question is of good quality", all conditional downvotes that were applied to the question are reversed, and the option for conditionally downvoting the question is removed, permanently, unless the question has (or later gets) a negative score after the conditional downvotes are removed, in which case the option remains for another go-round in the questions needing improvement queue.
  • The other options for reviewers in the improved questions queue include: skip, which just goes to the next entry in the queue; and I'm done, which will apply any vote or edit changes made by the user but will not remove the question from the improved question queue unless the reviewer upvoted the question.
  • 10k users and diamond moderators have no restriction on the number of conditional downvotes they can cast per day.
  • <10k users who are not diamond moderators can cast 50 conditional downvotes per day.
  • New badges for the new queues and specifically for editing questions that are in need of improvement.
  • A notification is added to the StackExchange notifications inbox of any user who has cast a conditional downvote, whenever the question makes it through the improved questions list and has its conditional downvotes removed. This gives the user the opportunity to revisit the question and determine whether it needs more editing, a straight-up downvote, an upvote, or a vote to close.

This is a good suggestion, btw, but I doubt SE will implement it because it adds significant complexity to a system that's already reasonably complex with several existing review queues, votes to close and delete questions, the "on hold" mechanic, etc.

Also, this question might best be migrated to meta.stackexchange.com to get the entire community's feedback, not just SU -- because even if they were to implement this, it would certainly be network-wide.

  • 2
    You raised a number of nuances that make sense when you start thinking about how it would be implemented. I'm not sure it would need to be quite as complex. For example, I wouldn't use 0.5 voting; the potential fixability wouldn't change the actuality. The main point was a way to clean the slate. I'm not sure there's a need to treat conditional downvotes differently from regular downvotes in terms of quantity limits or rep differences. The new queues might not be necessary; hold/close works. – fixer1234 Oct 9 '14 at 19:33
  • 3
    It might not be necessary to introduce new badges and edit incentives. The more complex it is, the less likely it would be implemented. The simplest approach would be tacking on an available "path to redemption" on the current system with the minimum necessary changes to the infrastructure. Your suggestion would implement a question improvement system, which has a whole additional set of benefits in its own right, and in the process, would solve the problem I raised. – fixer1234 Oct 9 '14 at 19:42
  • BTW, I think the scope of your suggestion would be a great improvement if it could be implemented. – fixer1234 Oct 9 '14 at 20:41
  • I’ll want to read this again when I’m less tired, but one thing jumped out at me: users who have the privilege to make unilateral edits still get +2 rep points for editing a question that’s received conditional downvote(s). I don’t understand why we need to give away free rep points like that. From what I’ve seen, many high-rep users still do plenty of editing, just because it’s the right thing to do. – G-Man Oct 9 '14 at 23:01
  • 2
    Giving reputation for making edits to a question that needs it seems cheap and a great way to give reputation for a user that neither asks or answers questions which is the entire point – Ramhound Oct 10 '14 at 0:11
1

Generally, questions that receive a lot of downvotes will likely be closed pending review. If the question is reopened, then I think the votes (up and down) should be erased and let the cycle start again.

  • 3
    If a question has been edited then reopened people can retract their votes, I don't want my vote retracted unless I agree a question or answer is helpful. – Ramhound Oct 23 '14 at 0:31

You must log in to answer this question.

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