4

Both a moderator and I made comments on an answer to this question today to the effect that it (i) did not address the OP's question in any way and (ii) proposed lying to avoid having to face the issue posed in the question. The author of the answer has acknowledged the need for an edit but two hours later had not made any edit, meanwhile it still accrued votes.

Before any edit was made the answer had received 55 votes (41 upvotes, 14 downvotes) and has since accrued more. I believe that an edit to an answer allows previous voters to change their vote (assuming they become aware of the edit and can be bothered re-reading it and changing their vote), but still the votes remain.

My first concern is that an answer could be substantially changed - every word of it - after it has received votes and retain those. This could work either way - a good answer could receive lots of positive votes and then later be changed to a terrible or inappropriate answer. Alternatively a poor answer could receive a flood of downvotes and then later be improved, but now it has to fight the previous downvotes.

I appreciate that a contributer has the option of posting a completely new answer rather than edit their old one, but my concern is more that it could be misused. There was similar concern in recent history over how Facebook posts can be completely edited but retain "likes"; for example a photo of a kitten could accrue thousands of "likes" and then be switched for a photo of Hitler. That might be an extreme and unlikely example for IPS - but what if someone wrote an answer which wasn't completely inflammatory but tapped into the general population's feelings over an emotive topic (take your pick - gender or sexuality issues, whatever), gained a high number of "knee-jerk" upvotes, and then later appends or rewrites it to be far more right-wing. If the previous upvotes had ensured this had risen to the top as the most popular answer it would be a very bad representation of what the site is about. If it was deeply unpleasant then sure, mods could flag for deletion, but it would have to have crossed a line, and I have seen answers skirt the edges.

My secondary concern from this is that for all the strict moderating of questions and answers to ensure that this site is not "opinion based", votes clearly are opinion-based (obvious I suppose). But further to that, votes clearly ignore the guidelines of what a "good" answer is. Allowing votes to remain on answers which have changed from "good" to "bad" or vice-versa surely misrepresents the standards of the site?

With these points in mind, should anything be done to prevent answers, especially ones that do not meet the site guidelines, from retaining accrued votes after substantial edits or complete re-writes?

  • 7
    This is an interesting question. I'm wondering though, are you sure this belongs to IPS.meta rather than on MSE? Seems to me like a quite generic issue. – avazula Oct 9 '18 at 12:49
  • 5
    @avazula I did wonder that - it might not be unique to IPS but I don't think it affects all SE sites. For example answers to tech questions on stack overflow are either right or wrong. People tend to only vote for tried and tested solutions. It is really only on sites like this where we filter out "opinion based" questions but the votes are more likely to be based on reader's opinions and gut-instinct than actually putting them into practice themselves. In my specific example, I doubt that 55 people all told their friend they'd been cheated on by another friend and then came back to vote. – Astralbee Oct 9 '18 at 12:54
  • 2
    meta.stackexchange.com/q/135626/369802 might be an interesting related point. But lets keep this here, as I think it points out some important stuff. I'd like to see the answers it gets here. – Tinkeringbell Oct 9 '18 at 13:34
  • 2
    I was wondering the exact same thing the other day about that answer who got deleted (with ~250 upvotes) then heavily edited (I'm thinking ~90%) before undeletion. Feel free to include this example in your question. – Ælis Oct 9 '18 at 14:52
2

I believe that an edit to an answer allows previous voters to change their vote (assuming they become aware of the edit and can be bothered re-reading it and changing their vote), but still the votes remain.

This is true. Once a post is edited, votes can be retracted/reversed. More information on the exact circumstances can be found on meta.stackexchange:

  • In general, once you have voted, you cannot change your vote. There are two exceptions.

    • Exception one: you may change your vote a practically unlimited number of times within a five minutes window from the first vote you cast on that post. Note that after voting and undoing your vote 30 times, it will also be locked in.

    • Exception two: you may change your vote after every time the post is edited. A new window starts with the first vote you cast after each edit.

For a question that hit the Hot Network Questions list though, don't expect much voting to be undone or changed. People read the post, and are more likely to vote for agreement/disagreement then actually consider whether or not a post fits the guidelines on IPS.


There was similar concern in recent history over how Facebook posts can be completely edited but retain "likes"; for example a photo of a kitten could accrue thousands of "likes" and then be switched for a photo of Hitler.

This may not be the best example here. If you EVER come across a post on IPS that went from kittens (being an actual, good answer to the question) to a picture of Hitler (spam, rude, abusive, asking 'delete me', or other nonsense) feel free to flag such a post for moderator intervention. It's then likely it should be deleted and/or the edit rolled back. Practice due diligence when flagging this stuff, look if the author and the editor are the same people and when the edits happened for example.

We're actually talking about the opposite happening here. We're seeing answers, holding opinions that people agree with, getting upvotes before the community has the time to point out that something in the answer doesn't hold up to community guidelines. That can happen, some posts reach the Hot Network Questions list very rapidly and there may not be anyone around to notice the answer until it has reached quite the number of upvotes and someone notices 'hey, nice opinion, but could you explain how this answers the question and perhaps add some expertise'?

I've expressed my personal opinion on this before in chat:

This is my, purely personal, opinion about the reputation 'issue' with ArtificialSoul's answer: We don't really take into account votes when deleting answers, so we probably shouldn't worry about them when undeleting either. If the OP provides the effort to turn it into a proper answer, they can have some sort of 'reward' for it.

Furthermore, now that it's edited, people can undo/reverse their votes. So if anyone comes back and sees that this is 'not what they voted for', they can fix that themselves ;)

I see editing your answer to improve it, to actually address the question, to include expertise, as something that may be 'rewarded' by keeping your votes. If anyone disagrees, like I said, they can fix it themselves by unvoting.


My secondary concern from this is that for all the strict moderating of questions and answers to ensure that this site is not "opinion based", votes clearly are opinion-based (obvious I suppose). But further to that, votes clearly ignore the guidelines of what a "good" answer is. Allowing votes to remain on answers which have changed from "good" to "bad" or vice-versa surely misrepresents the standards of the site?

As I pointed out before, you're probably right. Especially after a question has been through HNQ, answers votes don't represent usefulness or site standards, but agreement.

As for answers changing from good to bad, see what I wrote above. You're free to flag such posts and bring them to moderator attention if there's a major decline in quality, or undo your vote own vote if applicable.

As for answers going from bad to good, I personally don't mind them keeping their (positive) scores. Sure, it may not be what people voted on, but at least we now have one good answer with a good score that actually represents site standards. For negative scores this is of course a bit more difficult, but it's not unknown for answers to, over time, go from very negative to really positive scores. Our job, when we see a downvoted answer that meets all the guidelines of this site, is to acknowledge this by refraining from adding our own downvote OR upvoting if the information presented in those answers is actually useful. An answer may meet all the guidelines, but still be 'wrong' and thus be downvoted

  • Thanks for understanding my concerns, not sure if this answers them though. My Facebook example was extreme - here it would be more likely going from an answer which isn't inflammatory but perhaps taps into people's unrest over something and gains knee-jerk upvotes, then gets appended and rewritten to be far more right-wing and unpleasant - but it already has so much support it has risen to the top as the most popular answer. I have seen examples of this in the past. – Astralbee Oct 10 '18 at 11:34
1

I fear that the fact people keep they reputation after a heavily edit increase the "Fastest Gun in the West Problem".

Scenario

  • I see a question that is likely to end-up on NHQ.

  • I write a quick, not so good, answer so that I can be the first to post.

  • I wait.

  • If someone says something, I try to improve my answer. Otherwise, why bother?

  • My answer has +100 upvote now, but people say that there are major quality issues with it and that it will likely be deleted. Since it's not deleted yet, I decide to wait-and-see.

  • My answer is deleted, I lose all my reputation. In order to have it back again, I heavily edit it and it works: the answer is undeleted, my reputation comes back (even if nobody voted for what the answer currently is), everything is fine.

  • I go look for the next NHQ question.


I agree that people should be rewarded if they improve their answer. But, if people upvoted the answer when it didn't meet the quality guidelines, shouldn't the reward be to not have your answer deleted?

If the upvote (and downvote) count is erased after a heavy edit, then people might stop thinking they can just write a bad answer, get the upvote and work for a better version later if it's really needed. Hopefully, they will write a good answer from the beginning!

  • You make some good points. I'm a bit guilty of 'fastest gun' myself sometimes, i think we all are to a degree. I'd never write a deliberately bad answer, but if I think it's rounded enough I'll post and then continue to tweak it. But this doesn't really cover my issue. I think a downright rubbish answer gets downvotes. My concern is over opinion-based answers which court popular opinion and get upvotes from people who don't care about that filter, then get substantially changed, maybe even re-written from the ground up. – Astralbee Oct 10 '18 at 12:33
  • @Astralbee Isn't it the same problem? People posting opinion-based answer who then re-written them because they gain a lot of rep and want to keep it and that editing is the only way to do it? – Ælis Oct 10 '18 at 12:36
  • @Astralbee do you perhaps have a concrete example of such an edit? I'm still thinking that if the popular opinion gets edited to be more substantial, to honour the premise of the question, to add expertise... the popular opinion is still there? I don't think I've ever seen an answer edited in such a way that it went from popular opinion to unpopular opinion. And if it did, something is wrong with moderation, as we're only supposed to focus on improving answer quality, not on changing the writers opinion... – Tinkeringbell Oct 11 '18 at 18:03
  • So yes, perhaps we ask the answerer to show how the popular opinion will reach the same goal, or to add more explanation of why the unpopular answer won't help... but I sure hope we never told anyone 'your answer is wrong, it's now deleted', and that we've always stayed on the line of 'your answer is lacking, please improve it' – Tinkeringbell Oct 11 '18 at 18:05

You must log in to answer this question.

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