Proposal: do quality voting with upvotes, downvotes and no-votes

Hey folks. This is my first post in the forum so I hope my late start doesn’t give me a false start… This is a proposal which extends this existing discussion about voting but I have so much to say that I decided to make it stand alone.

So, this is relevant only with the assumption that quality voting is going to be a thing in the new site. Which I do believe is the common consensus.

The problem

Let me address up/downvoting on “that other site”; I fully support its existence yet I find it too restrictive for its own good. You can either vote up, vote down or not vote at all. We’re all still human and as humans we have needs to express ourselves in a manner that has the appropriate amount of tact and nuance. You might argue “You can comment!”… but we also know that this can be like tip-toeing through a minefield. Especially if downvotes have already landed.

Downvotes are an issue, they have a bit of a false reputation problem (that’s kind of punny, isn’t it). I personally think people hammer too much on downvotes when they are talking about something being bad. Just look at the daily meta posts which take the form of “Why was this downvoted!?”. No no I then think. Why was it not upvoted?

A lack of upvotes to me weighs in just as much as the existence of downvotes. People are not moved to upvote it, why not? And that is where my story begins, because how exactly do you demonstrate factually that besides people downvoting there are also people which are actively choosing to not vote? They could have been potential upvoters that make sunshine and unicorns appear, but they were held back by something. We can do that by giving them the option to record their decision.

The suggestion

For this I draw inspiration from a Dutch technology website that I frequent called tweakers.net. This is not based on the Q&A section of the site, but just posts on news items. On that site, reply post quality voting is done with a range of options. Now this range of options works because on that site post scores are not summed, they are averaged and can range from -2 to +3. The voting range goes as follows.

  • -2 = offensive / needs to be removed sooner rather than later. An option only available to people with elevated site privileges, if it is applied the post becomes invisible by default.
  • -1 = unwanted
  • +0 = off-topic or chatty, but otherwise not harmful
  • +1 = on-topic / good
  • +2 = expert information
  • +3 = spotlight

I took a little poetic liberty there with the translations.

I am not suggesting that we adopt this range as-is because that doesn’t really work for a Q&A site, at least not in a way that I can think of. You’d also have to switch to the averaged post score system I think and that will have ripple effects on other important site features. However I am really fond of having the ability to explicitly do what I will refer to as a no-vote, with which you can basically say that it’s not good enough to upvote, not bad enough to downvote but there is a good possibility the scale tips towards upvote territory if just a little more effort is invested.

That’s what I really miss on SO, tangible evidence that people chose to not vote, likely because they thought “meh”. I think “meh” a lot when perusing questions and answers. I would be more inclined to post a comment with some tips if the no-vote is a thing. I’d even be okay with the vote not being anonymous as then people can really see where my comment is coming from.

So a post score breakdown would be like this, translated into neutral wording:

  • x people voted that this content is of good quality
  • x people voted that this content is incomplete or not entirely correct
  • x people voted that this content is not of good quality or largely incorrect

Something like that, the exact wording should be decided with scrutiny. But I hope you can agree that having that extra layer in the voting breakdown can really make a difference in how you read it as a whole, especially when you’re the author of the content. It adds a lot to the neutrality of the voting process.

Bonus

I also don’t like the harsh time limit on SE where you cannot change your vote anymore, until the content is edited. I am in favour of there being a grace period to prevent fraudulent behaviour long after the fact, but I’d like to see that time limit set to at least 24 hours especially if the no-vote option is considered to be a good idea. The site needs to take into account that people go to the toilet, go out for a smoke, take a shower and/or go to bed and change their mind in any of those places.

9 Likes

After reading your post, I’m still not clear on what problem exactly your 0-vote is intended to solve. I suppose it differentiates between those that thought the post was meh, and those that didn’t bother to vote at all. As I understand it, the 0-vote doesn’t actually change the score of the post, so I’m not sure what the benefit is other than perhaps keeping statistics and getting a better insight on general user behavior.

We have kicked around the idea of having two types of downvotes, anonymous and signed. Anonymous downvotes still effect the score of an answer, and thereby its sort order. However, they don’t have any effect on the author. You can’t use anonymous downvotes as retribution, for example. That takes away a major objection to the SE voting system.

Signed downvotes leave a comment or some statement as to the reason, and there is some mechanism to for all to see who cast these votes. These do have some effect on the author’s rep.

This mechanism was mostly proposed to deal with the problem of anonymous downvotes on SE, so it hadn’t really been talked about for upvotes. We did discuss the concept of an expert stamp of approval, but I’m not sure what, if any, mechanism was decided on. I suppose signed upvotes could be used for that. Ordinary anonymous upvotes change the answer score, and increase the author’s rep some. Signed upvotes by recognized experts (ignoring for now how that is determined) give a higher bump in score, a higher rep bump to the author, and also some visible expert stamp of approval.

So with these mechanisms there are already 4 different types of votes, 5 if you differentiate between signed upvote by expert and ordinary user. I’m not sure what the 0-vote type would add to all this. I don’t have any real objection to it, but I don’t see much advantage either. Given that, it seems to me that the extra complexity it creates and the cost of implementation isn’t worth it.

3 Likes

When I see what I thought was a good post that doesn’t seem to be getting traction, I sometimes wonder whether that’s people saying “meh” or people not even reading it. On SE you can see a question’s view count, but that doesn’t help with answers. I like the idea of being able to communicate “meh”.

As a bonus, sometimes on SE I come across old posts that I’ve already decided were “meh” and don’t rate a vote, but it’s been long enough that I’ve forgotten and I end up rereading and re-evaluating. Whether tracking my “meh” vote is a plus or minus in this scenario is unclear – sometimes I wish I’d had the reminder so I wouldn’t invest the time doing the same thing I already did, but on the other hand, sometimes I reread something and change my vote because I have different knowledge and perspectives than I did when I first saw it.

10 Likes

It sounds like they can still get retribution by affecting “the score of an answer” x100. (It sounds theoretically impossible for both “downvotes do something negative” and “downvotes cannot be used for retribution” to be simultaneously true.)

I favour anonymous downvotes: they prevent arguments. I can’t imagine ever wanting to sign a downvote—I don’t want my name associated with reading rubbish.

It says “I could have upvoted this post, but I chose not to.” It indicates that the post was reviewed, and considered not worth downvoting, but not good enough to upvote.

It’s also handy to mark a post as “read”. I could imagine systematically going through some smaller sites and reading every single post.

I guess there’s also robo-voters to consider (e.g. to obtain badges), and maybe there’s less harm if they robo-“meh”-vote.

Many answers on language sites are of the form “I’m a native-Swahili speaker, and [XYZ] is correct”. They’re “meh” answers: maybe they answer is correct, but it’s nevertheless unreliable.

2 Likes

They can still do some damage, but it is more limited. Having a small effect or sort order isn’t as big a deal as dinging a user.

I can’t imagine casting a lot of anonymous votes. If I think something is wrong, I’ll be able to justify it. I would have no objection to SE making all downvotes public retroactively. I’m willing to stand behind what I said.

The real problem with any anonymous activity is that it’s a haven for vandals. This has been rather irksome on SE. Maybe you haven’t been the target of such things, but I certainly have been many times. I would love to force all votes to be public. However, I recognize there is strong objection to that out there. A compromise is to have both anonymous and signed vote types. The voter gets to chose. You can still anonymously indicate you think something is bad. However, the system will limit the damage vandalism and retribution can cause. A vote you’re willing to public stand behind will have more weight.

1 Like

I think we have clear consensus that votes will be anonymous by default. Mandating public votes would be a huge change and would require that a substantial case be made and agreed to. I don’t see that happening. It sure hasn’t happened yet.

What has been discussed (inconclusively, I think) is allowing people to make specific votes public. The use case here is an expert being able to say, more prominently than in comment #37 in that thread, that this answer is right (or wrong). It’s up to the reader to evaluate that vote, ascribing whatever weight one feels is due to that particular user’s opinion.

4 Likes

Yes. Maybe that’s what this 0-vote concept really leads to.

One problem with the SE system is that there is no way to indicate you have reviewed a post while you’re there reading it normally. I read most new questions as they came in. Occasionally I’d try going thru review queues only to find mostly the questions I’d already looked at. In the end, I just ignored the review queues.

This problem would have been fixed with a “Reviewed” button by each post. That would not only keep them out of my review queue, but the system would then know I’d done review work. I did lots of review work, just that SE didn’t think so because I didn’t do it via the (useless to me) review queues.

5 Likes

Yes, and I thought we’d take it a little further. Anonymous downvotes only effect the question score and sort order, not the author’s rep.

Agreed. I tried Review Queues when I got the privilege. I quickly decided they were poorly designed and only occasionally check them. But for the communities where I’m active I read almost every question anyway and “review” as I read them.

5 Likes

People are always free to claim downvotes by leaving a comment explaining what is wrong, that tends to lead to backlash and a negative reaction from the OP so I won’t ever do it, but if others want to then that option is always there.

2 Likes

Oh, you just mean rep. Ok, we still haven’t really worked out what we’re doing with rep. I don’t think somebody with one decent post and a zillion bad ones should show up as comparable to someone with one decent post and no crap, but we can sort that out on a thread about rep.

3 Likes

It sounds to me like a “reviewed” button and a “zero-vote” button serve exactly the same purpose. After you’ve read the content, you can mark it “plus”, “minus”, or “fine, I guess”.

I like this idea, regardless of what we call it. It helps the user keep track of what they’ve read, and let’s future readers (and the OP) know that the content could be improved.

6 Likes

“Reviewed” sounds good to me too, it’s less suggestive. Originally I was calling it the zero-vote because it basically doesn’t affect the score (+0) but that too is suggestive, like “this has no value”.

2 Likes

I actually prefer “meh”, but I don’t think we need that much honesty :grin:

2 Likes

Could the percentage of anonymous to commented votes be used to influence the ability to vote one way or the other? e.g. 100% anonymous = cannot vote anonymously 100% commented = can vote anonymously. So a self-regulating system to a fashion. Note sure but I suspect commented votes would always be allowed.

Down votes, when cast for the right reasons, are just as important as up votes. Also, a lot of people (even though I’m not one of them) really want to be able to cast anonymous votes. We have to let them. What we can do is limit the damage that a vandal or revenge voter can cause with downvotes.

What I think we are converging on is that anonymous downvotes count towards the votes of an answer, which effect the sort order. They do not, however, harm the author, like by decreasing rep or something. Signed downvotes do both. If the signed voter is somehow recognized as a domain expert, then there might be some additional “stamp of approval” or “stamp of disapproval” on the answer.

1 Like

@Olin OOOps I actually meant to say anonymous v commented instead of up/down. Will edit the post accordingly.

That’s fine for me—I don’t want to end up in arguments; the lack of chit-chat and focus on Q&A is why I like this format.

In fact, this would probably encourage me to downvote more, particularly content that I feel is to high on the list (e.g. old broken links). It depends if we have a “meh” vote.

A “stamp of disapproval” may be quite the burn. Perhaps this is best reserved for the “mistaken” posts that look correct (and thus receive multiple upvotes) but are wrong.

1 Like

SE has the concept of counting rep for given tags. Given enough relative rep - or accepted answers - for tags you’ve answered questions for, you could get a tag domain expert badge, bronze, silver or gold.

I like the “Meh” vote idea… we should encourage all users to vote on every question that they feel qualified to evaluate, even if they are not experts, separate to a reviewed flag. I also like the idea of a signed vote, perhaps also linked to a comment, so you can see that Joe Expert voted +1, for reason “x”. Joe might not leave a text comment, but his vote should be seen there, and not vanish into chatland if comments are migrated.

3 Likes