It seems we have a problem. Specifically, a downvoting problem.

Early in the private beta, I asked Should this be a high-voting site? which received (relatively speaking) massive community support. A passage from the accepted answer on that question seems to have gone by largely unnoticed, however:

Downvotes are also important! Sometimes answers have problems, sometimes questions have problems. One of the most important things for a new Stack Exchange site is attracting experts. The number one thing that turns off experts is noise. If someone has a lot of experience, and takes a look at our front-page, we want them to immediately think "That is an interesting question," or "Oooh, I was just struggling with that, let me share my knowledge," not "Wow, ok. What a bunch of amateurs. Seriously, 10 upvotes? Did any thought go into that?"

It appears that most people are only, or almost exclusively, voting up. Voting up quality content is great, but we also need downvotes for bad or "not useful" content.

When I look at the moderator Analytics data, the specifics of which cannot be shared in public, there are peaks and valleys on the graph of upvotes, whereas the downvotes barely register on the graph when viewed using the same scale. This Analytics data tells me that in the past 28 days, of all votes cast, about 97% were upvotes.

Stack Exchange Data Explorer allows looking into database dumps. Some aggregate statistics for all sites say that we, on average, vote up posts 48.5 times for every downvote cast.

Another SEDE query tells us that on average, with 10079 posts (not including deleted posts) each post on average receives about 5.40 upvotes and 0.16 downvotes, for a 34x ratio. This agrees well with the Analytics data. Note that this actually means that on average, less than one post in six is ever downvoted.

What is perhaps even more alarming is that we actually have a substantial number of deleted, upvoted answers. I used the search feature to get a list of all deleted answers with a net vote score of +1 or higher. At the time of writing this, it returned 35 hits. Expanding the search to answers voted net zero and higher and deleted, it returned 263 hits. That is, out of our some 10,000 answers, a full 2.5% or so was deemed bad enough to be delete-worthy, but were not deemed "not useful". I occasionally come across answers when simply browsing the site that have been deleted by someone other than the owner, but which are not negatively voted.

Content curation requires both upvotes and downvotes. It's how the Stack Exchange Q&A model is able to consistently produce high-quality questions and answers that actually match each other.

• Why aren't people downvoting more? If the content is bad enough to warrant outright deletion, surely it is bad enough to warrant a downvote as well; but there can also be plenty of content which warrants a downvote but not a delete vote. Delete votes are not super-downvotes, just like close votes aren't super-downvotes; the actions are separate for a reason.
• We have a highly active community on this site, which is great. What can we do to encourage that community to actually exercise the downvote button more?

Keep in mind that a downvote isn't permanent; it gets locked in a few minutes after being cast, but even once it has been locked in, if the post is edited (by anyone) then the vote can be retracted and even reversed.

• Oh, the irony that this post got upvoted. :-) – a CVn Apr 12 '15 at 21:48
• Content evaluation can be based on upvotes only. Why only all the people believe this SE "be rude" brainwashing? – yo' Apr 12 '15 at 22:05
• @yo' Downvotes are not rude. Deleting about one post out of 40 which has not been judged by the community to be "not useful" seems to me to be rude. – a CVn Apr 12 '15 at 22:08
• I see your point: WB.SE seems to be much more subjective than TeX.SE. We usually leave the answerer some time to correct the post, and if the post is still wrong, we delete it. – yo' Apr 12 '15 at 22:11
• Wow; meta's busy today. – HDE 226868 Apr 12 '15 at 22:26
• As a new member, I didn't find the number of up-votes to be particularly alluring. Nor the number of 'views'. Is it your experience that people are being deterred from WB, because of up-votes for 'bad' questions? I suspect this is a behavior where up-voting something gives the feeling that 'more people up-vote, so I should get one next', but the numbers don't affect my click on the question (again, I'm new, just commentin'). – Mikey Apr 13 '15 at 6:49
• Yet another example - a low quality answer, yet the vote tally reveals a +1. Admittedly there is also -2, but the poster's got 6 rep off that which while not a lot, could give some people the impression that they've posted something good. That answer as it stands is not. – ArtOfCode Apr 13 '15 at 16:13
• Fantastic post. However, because I agree, -1 – Fund Monica's Lawsuit Apr 13 '15 at 19:53
• Here's a SEDE query to match possible downvote-worthy answers (thanks to Michael for his assistance): use own judgement – ArtOfCode Apr 13 '15 at 21:39
• All too often the answers I want to downvote are from new users, and I don't want to give them a bad first impression (I may add a comment, but I usually leave that for people who are better at that sort of thing). – DaaaahWhoosh Apr 14 '15 at 15:50
• @DaaaahWhoosh Well, now you have a few pro forma comments that you can use where appropriate. See meta.worldbuilding.stackexchange.com/a/1999/29 – a CVn Apr 14 '15 at 16:19
• @MichaelKjörling I was wondering where those came from, thanks! – DaaaahWhoosh Apr 14 '15 at 16:25
• Tangentially related - the number of positively voted questions on the first page impacts my opinion of the SE's overall friendliness towards new users. I had a bit of a bad experience on a different SE, and that SE has what seems to me an extremely high number of down-voted questions on the first page. I see that SE community as a bit... bitter... towards new users. I have a much more positive opinion of this community. I'm not sure how much my opinion is influenced by my experiences, though. :) – Jerenda Jul 16 '15 at 5:41

Why aren't people downvoting more?

Here are some possible explanations:

• Many questions are subjective, and have no "wrong" answer. This is not the case on most other Stack Exchange sites. For example, on Mathematics, an answer saying that $x+x=\frac{x}{2}, x \in \mathbb{R}$ will get downvoted to hell because it's wrong.1 On Worldbuilding, you don't see many cases where answers can be blatantly wrong.

A lot of answers that aren't so good are still "useful". Unfortunately.

• The community is too soft. I notice a huge mood change when I swing from, say, Physics over to Worldbuilding. We're rather lighthearted - perhaps too much so - and more accepting of ideas that aren't so good. As a group, Worldbuilding's users welcome pretty much anything. We seem to have subconsciously developed lower standards than other sites.

People are going to jump on me for that last bit. Yes, I'm talking about science questions. I've harped on it before, and I'll use it as an example: we don't downvote crap science as much as we should. By "crap science", I ineloquently mean ideas that are only somewhat based on science and are instead speculative. I see a lot of posts regarding the insides of black holes, parallel universes and the stuff of popular science - not scientific journals. Some of the stuff here would get downvoted on science sites because it isn't science. We let it go by.

Science is just one example.

• People don't want to lose that 1 rep. Downvoting answers loses you one rep point. While that's the same network-wide, I'd speculate that we have a much, much, much higher answer-to-question ratio than most Stack Exchange sites. A higher percentage of our posts are answers, so people are less likely to downvote. If we had more questions, I think downvotes might be slightly higher. Slightly.

What can we do to encourage that community to actually exercise the downvote button more?

• Set higher standards. No more of that "crap science". Much as I'd like to see as many questions as possible use the tag, there's no way I can do that, because it will fail just as the tag failed. It was overused. I don't want to see that happen to .

We can agree to downvote speculative stuff to bits, like a wiener dog run over by a cement mixer. I hope everyone remembers how we agreed to keep everything based in at least basic logic. I say we go out an apply that idea.

• Think. It's so easy to keep upvoting, especially on posts where there are a lot of upvotes. My sports team had the motto "Don't be a sheep!" (Ironic) Let's do the same, because I feel like this is an issue at times on Worldbuilding. I know we have some long posts - I'm trying to make mine shorter - but we should really take the time to read them in their entirety. This may lower the upvotes, even if it doesn't raise the downvotes.

I'll admit one thing: I don't downvote a lot. I probably should, but I don't I apologize if I seem to be criticizing everyone else; I am guilty of the same thing.

I should also add that just because an answer is wrong is no basis for deleting it. We can delete answers that don't actually answer the question, are spam, etc., but we can't delete wrong answers. This has always annoyed me a bit, but it's Stack Exchange policy.

1 It claims that the case holds even for $x \neq 0$

• If an answer is wrong, downvote it. We're not supposed to delete just for being wrong; that -10 wrong answer serves a purpose too. If we aren't downvoting the wrong so people will see that it's wrong, that's a problem. Deletion is for things that aren't answers per site and SE guidelines, which includes link-only and one-liner answers with insufficient explanation, but doesn't include wrong science (or wrong anything else). – Monica Cellio Apr 12 '15 at 23:48
• @MonicaCellio That's what I was trying to say. Did it come across wrong? – HDE 226868 Apr 12 '15 at 23:50
• We mostly agree - just pointing out that IMO we shouldn't delete wrong for being wrong even if SE allowed. – Monica Cellio Apr 13 '15 at 0:08
• What do you mean by crap science, can you give an example? Keep in mind that a lot of users are not scientist. – Vincent Apr 13 '15 at 1:59
• @Vincent There's a lot of "pseudo science" that has no scientific basis but keeps trying to claim it does - we see some of that recycled here from time to time. For example psychic powers, homeopathy, intelligent design, etc. There are also a lot of more confusing/subtle physics or logic fallacies that are easy to fall into if you've just heard the buzzwords but don't really understand the subject. – Tim B Apr 13 '15 at 8:36
• @Vincent That's my endearing term for science that's sort of based on science but is really just speculative. – HDE 226868 Apr 13 '15 at 14:55
• @TimB I think that's it. – HDE 226868 Apr 13 '15 at 14:56
• +1 for the comment on actually reading a post before you up/down vote it. Maybe I am just cynical but I don't think half the people up-voting the top-ranked answer have actually read it... – James Apr 13 '15 at 15:14
• @James I also think many people only read the first answer, and, when they find its a decent answer, (which it must be because so far it's at the top), +1 it and don't bother to read the others - meaning a huge gap emerges where the first answer has +[thirty-something], second has +[teens or low twenties], third-fifth steadily drops until about +[1-5] for the others. – DoubleDouble Jun 17 '15 at 21:22
• "People don't want to lose that 1 rep" - this is pretty much why I never downvote. – Dan Smolinske Jun 24 '15 at 19:03
• @DanSmolinske Yeah, that governed me for a while. – HDE 226868 Jun 24 '15 at 19:54

I am one of the worst offenders (serial upvoter, stingy downvoter), so I speak from experience. I certainly don't mean this post as guidance for other (down)voters, merely as an exploratory analysis at best.

After I got past the point threshold for voting on closing question, I noticed that I almost entirely stopped downvoting. If I can flag a question for closing at no reputation cost, cleaning the exchange and simultaneously asking for other people's input (unless a mod is involved it's a cooperative close), there's little point to also downvoting the question. I have voted to close probably about 100 questions, while my downvotes are a trickle by comparison.

Moreover, as HDE 226868 pointed out, this exchange is fairly subjective. Just because a question looks silly to me, it doesn't mean it won't get 10,000 views and 100 upvotes. Actually, it seems that the relation between what seems silly to me and what gets upvoted the most is slightly negative if nonzero. So another reason to withhold judgment right there.

That said, I will downvote any posts about non-trace oxygen-hydrogen atmospheres and other such abominations.

• Since you talk about close votes, I assume that your answer is mostly about questions. Note that downvoting questions does not incur any reputation cost on part of the downvoter; only downvoting answers does (and at -1 rep to the voter, it's not like you'd lose any privileges any time soon). – a CVn Apr 14 '15 at 7:07

I am fairly new to Stack Exchange generally and also to Worldbuilding SE. I have upvoted often and downvoted hardly at all.

It did not occur to me until quite recently that downvoting mattered as much as it does. Suggestion: better publicize the explanation of why careful downvoting is constructive. Possibly put something along those lines in the tour, or under its own heading in the help center. Or make an article on responsible downvoting pop up when a user has just gained the ability to downvote. It may be that there already is an explanation which I have missed. If so, take my not finding it as one point of evidence suggesting that it is currently too hard to find.

I grant that this very question by Michael Kjörling is itself an explanation of why you should consider downvoting. But I found it only with some difficulty.

Another area for which I think better, more prominent explanations would increase responsible downvoting is the question of whether downvotes (and votes in general) are anonymous. I now think I have worked out that, yes, they are anonymous even to users of high reputation (although I have learned that canny users can often guess who downvoted). But it took me a long time to work out that votes are probably anonymous, and I am still not sure. Again, if there is an explanation somewhere, it needs to be better publicized.

The above assumes that confidence of anonymity will increase the willingness of users to downvote. Based on my own personal wishing to be liked and/or cowardice I think that is likely to be the most usual pattern. However I tentatively put forward an opposite suggestion: that making it visible who downvoted will help make downvoting respectable. The fact that it is public who put a question on hold, and it can thus be seen that this is regularly done by respected users, serves to demonstrate that putting on hold is an accepted practice. The same could be true of downvoting.

• Votes are anonymous; only people with access to the SE database (so, developers and a very few others) can see who voted how. Sometimes there is circumstantial evidence, but mainly on very small sites or weird corner cases. It's safest to assume that anything you think you know about who downvoted you is wrong. – Monica Cellio Jun 16 '15 at 22:09
• @MonicaCellio Read the answer explaining how the downvote was guessed. In general, it's actually pretty accurate (at least on moderately low speed sites), especially when combined with other factors. – forest Feb 28 '18 at 7:21

I haven't been here very long, but I still observed that the voting systems has some bias, at least as far as our community is concerned.

What I gather from what I observed, is that people usually limits themselves to the two or three most voter answers. And do not read further down. Which increases the votes of those answers, and leave the other ones either to 0 or very low votes. And hence no one actually bothers to down-vote them.

Similarly, we are quite active, meaning that the first answers will gather some momentum very fast. And answers that come two to three days later will often be overlooked, regardless of their qualities.

As a consequence, more than once have I seen a VERY good answer in the middle range: it came too late to be up, and a few bothered to read it, which extracted that answer from the 0.

All that digression to explain why I see some general "concerns" about voting on WB.SE. Now downvotes? Examples. As any new users, I can notice that there are few downvotes around. Furthermore there are a few comments about "Why is it downvoted, care to comment?", which would then render the anonymity void. But on the other hand, it can help to identify the problem and improve the question/answer, which would avoid the need to downvote from the beginning. This can be illustrated by the number of comments which seems large here.

Some new users may care about their -1 rep. A few of those could make a difference with some given priviledges.

Finally, it might be a "Don't do to others what you would not wish to be done to you" type of problem. People don't like to be downvoted, so don't downvote the others.

• If you (anybody) see a good answer that seems to be languishing in the middle of the pack, I encourage you to draw attention to it. One way to do that is to link to it in chat. Another (albeit more-expensive) way is to put a bounty on it; "reward existing answer" is one of the bounty reasons. Oh, and if you use social media, you can always tweet/link/post on your wall/whatever. – Monica Cellio Jun 18 '15 at 20:05

Wondering about why people don't downvote enough might be useful, but I think it would be even more useful being reactive and change the site rules in order to increment the amount of downvotes.

1. remove the reputation cost. Seriously: it's stupid, it makes no sense, it directly harms the quality of the site.
2. introduce badges for downvoters. This would be a nice incentive to downvote questions you don't like.
3. downvote should be effective towards every user, remove the 1 rep low cap. If I downvote someone who has 1 rep, he'll just stay at 1 rep. If, after that, he gets an upvote, he gets the full points, so it's like I never downvoted in the first place. This is stupid.

Some more depth about these points:

1. not really, it's just plain old stupid, nothing to add.
2. if you are concerned people would just randomly downvote to get some badges, well, this is a more general site problem, this could happen anyway for the general voting badges, and this might be solved by the secret magic bots who detect "improper" (sigh) voting.
3. if you are concerned about people having negative points, don't be: everyone should start at 0 rep, and you would at worst show 0 rep instead of a negative one. But, internally, you would count down to negative. If you are concerned people would create a new account if their current one has a negative rep, well, let them do it, who cares, how many times would this happen anyway?

I don't expect this to really happen, since, frankly, in my quite arrogant opinion the SE "higher" community is crystallized in their traditions, even when they are very detrimental, and once you point out something out of their safe canon, they go berserk and insult you (politely, sure! Ha!).

Anyway, that's my 2€\$. 2 cents were not enough.

• I'm going to have to disagree with that last paragraph at least. If you want it to happen, bring it up - though not here. Meta SE is the place. – ArtOfCode Jun 22 '15 at 18:33
• Ideas 1 and 2 have been put up on Meta and have been hit by the equivalent of anti-aircraft guns, if I remember correctly. – HDE 226868 Jun 22 '15 at 19:02
• @ArtOfCode the last time I pointed out something on meta, they enraged and I got the second most downvoted post in the history of that site. No, they don't want my help, they are a lost cause. – o0'. Jun 22 '15 at 19:03
• Comments removed. Disagreeing with other people is fine, but name-calling is not. Please keep it civil. – Monica Cellio Jun 22 '15 at 20:56
• @MonicaCellio you are not helping, though. This is meta, we would be supposed to be able to freely talk, however that happens. Anyway, that's exactly the kind of attitude that has ruined the main meta: you can be evil, but you cannot say it aloud. If you are genuine and tell what you think, you get censored. If you insult other people in a convoluted way, you can do it. Bad Monica, bad. – o0'. Jun 22 '15 at 20:58
• Talk about the behavior, not the person. Calling people idiots or bigots or worse is not ok -- not anywhere on SE. On the other hand, saying that you object to such-and-such policy because of thus-and-such reason is perfectly fine. I almost never delete anything on meta because it is more free-form -- but that's not license to be rude. I deleted two comments, by the way. – Monica Cellio Jun 22 '15 at 21:05
• @MonicaCellio and that's one of the many rules I disagree with, however this only slightly. At least this one, unlike too many others, definitely has a point. – o0'. Jun 22 '15 at 21:07
• Quite fun to notice how the only answer proposing an actual solution gets downvoted, while the others who merely try to analyze this problem get voted up. – o0'. Jun 22 '15 at 21:08
• (which is basically proving my point: you just don't want to change. Even in such a small site. Religion is too deeply ingrained in the whole network) – o0'. Jun 22 '15 at 21:09
• @Lohoris You call people stupid, bring up ideas that have (rightly so) been shot down before, behave insultingly towards religion (and this coming from an atheist), and, most of all, propose ideas that will not work. Furthermore, I spent the last half of my answer proposing solutions, as did Lostinfrance and bilbo_pingoun. No wonder you get downvotes! – HDE 226868 Jun 22 '15 at 21:35
• @HDE226868 well, isn't religion about faith? They have faith in rules that happen to "always been there", and firmly reject any change. That's religion pretty much by definition. You say "my ideas to not work"… uh, faith. How do you know they do not work? – o0'. Jun 22 '15 at 21:38