Requirements and costs of downvoting

That seems like an outlying issue of someone who sounds psychotic. I would not use this anecdote as good reasoning.

If an OP thinks it’s their right to post low-quality questions/answers, they get downvotes, because the questions/answers are low-quality. If they attack users for downvoting, then that falls under abuse/harassment and would be judged by moderators to see if further action is necessary, assuming the victim flagged the message.

2 Likes

Furthermore I think that publishing votes discourages them. This is will harm our site, as a healthy voting culture is important in order for the quality principles and management to work.

Publishing votes will also inevitably lead to cases of harassment/abuse. Either by the downvoted or by other users “complaining” about someone’s downvote behaviours. As you say, this could be handled by manual intervention/flagging, however I’d rather protect our users by default and by design, instead of requiring manual intervention in the most cases.

Reading through the prior posts, I think - without intending to offend anyone - that one goal of publishing downvotes would be to allow them to be reviewed by others. I am strongly convinced that this will harm the site, as it would lead to (escalating) disputes about downvotes. It would make revenge-downvoting easier and it doesn’t help anyone with regards to downvote reasons, unless you require downvotes to be accompanied by a statement of reason (comment etc.), which nobody suggested here.

Hiding who voted, seems to work fine on Stack Exchange. There are no really issues with it. Even voting rings/sock puppets can be dealt with. It would require some rudimentary SQL statement, showing the vote match between two users, which could be run by any mod/admin.

4 Likes

Note that a downvote comment would not require publication of the downvoter. The corresponding comment could be shown as written by a pseudo-user “downvoter”. If a downvoter doesn’t comment only for fear of losing the anonymity of the vote, this would give an opportunity to give a reason without identifying yourself.

3 Likes

This would clearly solve most of the problems of my prior post, however I think it must be designed carefully in order to prevent abuse (writing anonymously rude stuff, no option to find out who it was).

I could imagine a system, where moderators are able to see who wrote a comment. That would need a special column in the db, such as IsAnonymousVoteComment, however IMO that is certainly non-mvp.

3 Likes

The database already contains who voted. All you would need is a field that links the comment to the corresponding vote (and is NULL for regular comments).

No it doesn’t. People who downvote in revenge don’t care that it’s obvious that they downvoted.

No, it’s common in my experience. Common enough that it makes me reluctant to comment when I downvote, or when someone else has downvoted a post.

Yes, this. Downvoting bad answers is essential.

2 Likes

I think we have enough skilled programmers and computing power to solve all the issues here with simple code.

Have 3 classes of downvotes. They are all available after registered user has received any upvotes for any purpose (same as upvote privilege which should both come very early)

Type 1: downvote and leave a comment as to why, earn 10 point
Type 2: downvote and leave no comment
Type 3: downvote and maintain anonymity must leave a comment as to why and pay 10% of current reputation

Make it clear to the user which type they are selecting and explain why the options exist the first few times and then let the user turn off the “Reason Why” pop-up box.

Incidentally the “Reason Why” pop-up box should appear for every thing that the user does so that they don’t have to feel like the rule is stupid (and it won’t be stupid because the coder asked the designer for the text to explain why) and won’t have to be pointed to a FAQ or CoC after they get angry or sad when a moderator has to intervene.

I agree with Olin’s reasons for public and commented downvotes but also see that anonymous can have value to some but it must come at a real cost to prevent frivolous abuses.

My main reason for having only serious downvotes is because I do not like or believe in closes or even worse deletion of any question. Even a pathetic question has SOME value and it costs NOTHING to leave it there for the future researcher or volunteer who wants to read or repair it. Downvotes should be used to hide bad questions, answers and comments from those who have better things to worry about but they should not be removed. Hate and illegal stuff should be moved into deep dungeons but this will be rare and easy to find in those with lots of downvotes.

2 Likes

That is probably because you decided not to leave a comment at the time when it was VERY fresh in your mind.

Make it so that a person has to click the “link to user” on the downvote comment at a COST and it will be very unlikely that anyone will ever bother to try and dig deeper than reading the downvote comments. There is not enough time in the world to worry about what others say. Just block the harassing user and ignore them in future.

Having it public is good for transparency. It is not a secret ballot where transparency can lead to vote rigging in any real way so I believe it is preferable to the abuse of hidden voting.

This is both unbalanced and highly biased against downvoters.

Voting, including downvoting, is an essential quality control measure. If you stifle the free use of up or down votes, you lose good quality control signal. The fact that downvotes cost reputation on SE is deterrent enough for many not to downvote unless it’s unavoidable; making people write words or pay percentages of their reputation will kill it off completely.

7 Likes

Assuming we have a granular numeric rep system broadly comparable to SE’s, this would be painful beyond belief. A single anonymous downvote canceling out the rep from potentially dozens or even hundreds of very valuable, highly-upvoted answers? Ridiculous. Imagine Jon Skeet losing a year’s worth of rep — more than most users ever get — because he decided to downvote one post without revealing his name.

Also, it would be very ineffective at preserving anonymity: just look for someone whose rep just dropped by 10%.

Ah, I suspect this is indeed the major point of departure. I’ve gotten 10k on a few sites and have been able to see the atrocious sorts of posts that end up deleted. Yes, there are lots that are negative-value-added for basically everyone: any time spent interacting with those posts is a dead loss. I’m not just talking about offensive posts or spam, although there are likely more of them than you think, but the distressingly common posts by people who just do not get it and are aggressively certain that they do, or (especially on SO) are simply desperate to get their problems solved in any way possible, with the sole exception of actually learning or thinking in the slightest. It’s inaccurate to say that these posts have “some value”. In a sense they do, but the value they have (barring schadenfreude) is smaller than the cost any conceivable reader would have to pay to see them. And we’re not building a notalwaysright.com clone. (Even then, most of them aren’t actually that funny.)

3 Likes

The loss of reputation once you have 100 points is not a factor to take into account.

I rarely need to downvote and normally it would only be for critical comments that could have been left unsaid.

I think we should become a community of upvotes first and downvoting as a act of desperation sort of like picking weeds and not something to tune for. If all neutral or negative answers and comments were in grey and only the upvoted ones were bright I think people would learn to upvote the good stuff in a couple of days. No other change would be required and downvotes would need to be used rarely.

No it doesn’t! Maybe that’s the point many of you aren’t getting. When you have a lot of rep, and when you generally comment why you downvote when you do, you end up the target of vandals and revengers that hide in the shadows.

I leave a comment with most of my downvotes on EE.SE because I think it’s the right thing to do. Some people don’t take that well and react childishly by going to your profile and downvoting several of your answers. I’ve had downvotes on my answers reversed by the system a number of times. However, the system catching downvote abuse is rare.

This wouldn’t be a problem if downvotes were public. Such vandalous or revenge behavior would be obvious, and could be dealt with. Because it’s obvious and likely to be dealt with, many of those doing it now wouldn’t.

The current system of anonymous votes has problems found from solid experience. The argument against open downvotes are only theoretical, with nobody being able to say for sure what would happen. Lets try it at least for a while. I think it will solve many of the existing problems, but if it turns out to be worse, we’ll actually know that, and can provide defensible reasons for reverting to secret downvotes.

1 Like

As Gilles said before:

Therefore publishing votes would only allow easier investigation of vote abuse, which could be dealt with by offering admins a simple investigation tool.

TBH I think that the fact that allowed the revenge-downvoters to start their campaign is that the vote ownership was published. As we will still want to have comments asking for clarification/critiquing the post, we need ways to revert revenge-downvoting if we detect it.

We need to do two things IMO to solve the revenge voting problem:

  1. Hide votes. When you don’t know who voted for your post, you can’t do revenge-downvoting. Of course it will be public sometimes, for example because of comments by downvoters, however that should always be a voluntary option.
  2. Detect and disallow/reverse targeted voting. This will make revenge-downvoting fully impossible, as all votes will be reverted.

I think this is the best option we have.

Would the downvoter be visible to moderators/review members? Otherwise I fear reversing targetted voting would be rather difficult.

I can imagine the following things:

  1. a dev-only, access-logged feature showing who voted on this post
  2. a mod/review board-feature showing vote-matches between users (who voted often for another user) together with some relative overlap strength.

If we have the idea of community moderators (who do not have PII access) and instance admins (who have PII access on all sites and can have stuff escalated to them by community moderators), we can give access to the votes to the admins and they can investigate those cases when needed. That’s kind of like SE, where only (certain?) CMs can see and invalidate votes.

I know votes aren’t PII, but they’re sensitive info so let’s restrict the access to those we’re already trusting with sensitive info, at least to start.

1 Like

For the purpose of figuring out what users from different brackets of expertise in the area (judged by their reputation in the tags I’ve attached to my question) is, I think the following way of viewing votes could shed more light on whose votes count more and whose should weigh less in my opinion.

Imagine you could open up a breakdown of votes by brackets of rep within each tag like this:

How to create an instance of a class knowing its name in a string, using reflection?

[c#] [reflection]

Post votes breakdown by rep in [c#]:

   0  …   50  rep   +2  −7 − don't count towards total score and not visible except here
  51  …  500  rep   +2  −1
 501  … 2500  rep   +4  −1
2500  …   10k rep   +2  −2
  10k …    ∞  rep   +1  −0

Post votes breakdown by rep in [reflection]:

   0  …   50  rep   +2  −7 − don't count towards total score and not visible except here
  51  …  500  rep   +1  −1
 501  … 2500  rep   +1  −0
2500  …   10k rep   +0  −0
  10k …    ∞  rep   +0  −0

This would tell me that users who don’t seem to know (according to their user activity on site, we don’t know their actual level of expertise in every area, but this is what we have to work with) much about [c#] or [reflection] account for the most downvotes, and users who know something about [c#] account for the majority of upvotes. This would give me confidence in my post.

————————————————————————————

As for whether/when to display votes from low rep users and how much it costs…

I think everyone (with a registered account) should be able to vote immediately, and those of us who want to see those votes should be able to.

Votes from users who have 0 rep (or unregistered users) should not be immediately visible to others until they register and earn some basic lower threshold of 50 rep, at which point their previous votes become visible to others and start counting towards total post score.

I think showing a non-intrusive message “How can OP improve the post? Care to leave a suggestion?” should be the cost of a downvote, and to make it actionable, don’t limit users to comments only (which can be a hassle to come up with) but also allow one-click reaction icons/descriptions/text votes which could be added to the post instead of text comments. These reactions could be an alternative to canned comments, and contain something like:

  • Asked and answered many times before (aka “show your research” but you strongly suspect OP probably didn’t even do a simple search) − this is one of the most common reasons why a post may gain a downvote, andin this case I would recommend changing how users vote by not downvoting but leaving this exact “reaction” as a way for OP to know that their question isn’t “bad” per se, just that it’s not valuable to have it asked here for the 100th time;
  • Badly formatted − wall of text, code not in code block, excessive markdown usage;
  • Citations needed / authoritative references needed − for when OP says “everyone knows that X” and it’s actually not something widely known;

Also, a couple other reactions to illustrate how else this system can be used:

  • Illustrations could be a great addition − this isn’t a “downvote” scenario exactly, but I think it’s a useful to have quick “reaction” for when you don’t want to write a comment with essentially the same meaning;
  • Outdated answer − this is unlikely to help anyone in present day, but could have been useful at the time of posting;

Downvoting is not abuse. Closing questions is not abuse either. Both are vital curation mechanisms.

Oh boy no. Oh no. Please spend 5 minutes in the review queues on SO. Years on Stack Exchange have taught me that there is such a thing as a stupid question.

Besides, there are many practical issues with requiring comments for downvotes, which have been discussed to death on Meta Stack Exchange. In particular:

  • If the explanation of my downvote is “read any of the other answers on this question”, why should I bother?
  • If someone else has already written a comment that explains the reason for my downvote, why should I repeat that comment?
  • If I leave a comment like “this answer is not useful”, then I haven’t provided any information whatsoever. So what was the point of the whole exercise?
3 Likes

This is blaming the victim. It also incentivizes users to not explain their downvotes. Both are bad for the site.

You must not forget that the site must be friendly to the experts providing the good content. Put yourself into the position of someone like that. You downvote when you see bad content, but also explain why. That results in a few revenge downvotes that you can’t see, can’t rebut, and the mods can’t see either. Most people end up refraining from explaining their downvotes. I explain mine anyway, and have gotten a lot of (seemingly) revenge downvotes as a result.

Think about how this feels. Anyone with minimal rep can declare something you wrote as being bad, but you can’t face your accuser or even find out what they think is wrong. You don’t know what to fix or explain why you think it is right.

The current system sucks. This was one contributing factor to my curtailing activity on SE a year ago. I still have the highest rep on the Electrical Engineering site, even after a year of inactivity. I occasionally look without logging in, and I noticed that downvotes on my post pretty much stopped as soon as I stopped posting, while upvotes on my existing posts continued. This is strong evidence that most downvotes were short-term revenge reactions to something else I did. That’s just plain wrong.

Yes, and having all votes in the open makes this much easier to judge.

But this encourages everyone to NOT explain their downvotes. Start from a different premise. Suppose most people did the right thing and explained their downvotes. Now how do you protect those people from petty revengers? That’s what we really should be focusing on.

4 Likes

I like the general idea, but 10% rep for a single downvote seems too stiff a penalty.

While I generally like the idea of rewarding people for explaining their downvotes, I’m not sure how to make the mechanics work. What keeps someone from simply typing “lkjsd jlksj dfklj sdlkj ytiu yoij we lm” as the reason? How would the system detect fraud of #2 looking like #1?

Good downvotes should be encouraged, as you are trying to do with #1 and #2.

Now consider the likely quality of anonymous versus signed downvotes. Those willing to publicly stand behind their downvotes are much more likely to have a good reason for them.

Some people here argue that disallowing anonymous downvotes will reduce downvotes, and thereby reduce an important feedback mechanism. But, consider the quality of such lost votes. These votes are cast because:

  1. Not really sure, doesn't want to look like a fool for being wrong.

  2. Revenge for downvote or close elsewhere.

  3. Doesn't want to spend the time explaining.

  4. Afraid of revenge for disagreeing.

Properly addressing the problem of revenge downvotes eliminates #4. #2 is clearly something we don’t want. That leaves #1 and #3. Some of those might be legitimate, but how much weight do you really want to give those compared to the expert that explains what is wrong and puts their reputation behind their conviction?

3 Likes