Let's stop displaying a user's accept rate

Question

My feature request (or anti-feature-request perhaps) is the following:
Let's stop displaying a user's accept rate.

For those of you who know "Fawlty Towers", the whole issue of a visible accept rate and our behavior towards it starts to feel like "Don't mention the war!".

We generally seem to agree that harassing a user about it is not appropriate and that our decision to answer a question should not depend on it, yet we display it in varying terrorist-threat-levels of color. Yes dear users, please don't pay any attention to the brightly orange colored percentage you see there. And whatever you do, don't bring it up.

But as long as you're displaying the accept rate, people will comment on it, criticize users for it and possibly reconsider answering because of it. And while I'm of the opinion that a low accept rate should not stop you from answering (we're here to make the internet better after all) and that the possible reputation gain (or lack thereof) should not factor in either, I can't blame users for considering it. It's in their face after all. And we're only people.

So let's take the issue off the table. Or at least partially. Information about a user's acceptance behavior can still always be gathered from the profile. But at least that would be less in-your-face.

"But it helps me to see if a user is a help vampire!"
I'm still wondering if those are of the sparkly variety as well, but I digress. So what? If the question is bad, downvote it. Evaluate a question on its individual value. There is no need to take a user's history into account. There might be a correlation between the quality of a question and a user's accept rate, but there doesn't have to be. And if there is, the bigger problem is the question's quality.

"Exactly! A user with a low accept rate might indicate there's a problem!"
I'm not arguing we should get rid of the accept-rate value within SO/SE. After all, this Q&A works based on up and downvotes, as well as the acceptance of an answer. But why not use it behind the scenes to (very occasionally) nudge the user with a statement like:

"You have received several answers to your questions, yet have not accepted an answer as correct or helpful in a while. You might want to evaluate them and where possible accept them as correct or helpful".

After all, in the various questions with regards to comments on low accept rates, if any comments are encouraged at all, then they are of this soft variety. So why not let the system take care of that?

I feel this would take the whole issue of accept rate off the table, while still providing the user with the information he might have missed or needs. (As a bonus, it would reduce the frequency with which the topic is brought up here on Meta) And those users who would ignore this information are most likely not convinced by a displayed percentage either.

Answer 1

Starting with the next build accept rate will no longer be shown.

We're still keeping track of it on the backend for various things, but the negative behavior its display encourages outweighs its benefits.

Answer 2

Alternate request:

Show the accept-rate to the owner of the question only, and only on question user-cards. That way they get a gentle reminder that they're rewarded for ticking the box, and nobody else has to see it.

And move that public metric to their profile

---

As pointed out below by bluefeet, this would probably help contribute to the Summer of Love ;-)

Answer 3

I prefer seeing the accept rate.

Being a good SO citizen includes contributing to those who attempt to help.

Not everyone with a low accept rate is a Bad Citizen. People who pay too much attention to it, or don't bother investigating to see why the poster's accept rate might be low, also fall down.

Potential answerers who use the accept rate as a reason not to assist at all are also missing the point. I think the majority of consistent answerers wouldn't use that as the sole reason to move on to the next question. It might alter the amount of effort put in to an answer, and I'm not sure that's bad.

Removing it might solve that problem, but I'd rather encourage people to become better SO citizens over the long term, balanced against the risk that someone with a low accept rate won't come back.

Edit; minor clarification from my comments:

I believe allowing users to avoid playing the "SO game" diminishes what SO is--my opinion.

I use the accept rate as a cue to look at their questions; perhaps they can be improved if they're not being answered, perhaps they're OT, who knows. I use it as a way to see if a user might be struggling to use SO in the way I believe it's meant to be used.

Answer 4

I agree: Remove the accept rate from the user card. No user should ever see another's accept rate.

My rationale is that there are no positives to showing the accept rate, and many negatives. I've gone into more detail below.

---

First let me address the only real positive I've seen claimed about the accept rate:

[I]ntroducing accept rates heavily increased the amount of accepted answers in the system. Which in turn helps close loops and motivate the community.

• You can't measure motivation. But at a user level, I've never seem anything but discouragement happening because of the accept rate display.
• It seems to me that a lot of answers get accepted prematurely and, specifically due to this feature, for no reason other than to increase the accept rate. That basically makes the acceptance mark unhelpful and inaccurate, and certainly the fact that the number of accepts increased doesn't mean anything except that shaming people works.

Another quote from that post:

[Accept rate] is a one-dimentional metric that does not really give me enough information about how upstanding a user is in the community. It is used frequently to bully users and leaves a horrible taste.

I fail to see how the benefit is worth the horribleness.

And it is horrible; I reject the claim some have made that having a low accept rate displayed is not a form of punishment and shaming. To quote a comment made on an answer to another question:

Accept rate isn't to shame people, but to allow answerers to concentrate more time on those people who engage more in the site – Casebash May 10 '10 at 5:32

The purpose is irrelevant. What are the actual effects of labelling someone as "not worth your time"?

1. Users make comments intended to shame those with low accept rates.
2. Users make polite comments intended to gently remind a user to accept, which may still cause the user shame. You've probably noticed that people can get awkward when you tell them they've got food in their teeth or something. Now imagine that everyone in the room heard you tell them (equivalent of publicly posting a comment). It would be perfectly natural to be embarrassed in such a situation.
3. Users who notice or are told that other users refuse to spend time on their questions due to the accept rate are likely to feel rejection and/or shame and/or any number of other negative emotions.

There may be the rare case where a user is genuinely glad to be reminded if they had intended to accept an answer and forgotten. In all other cases, the emotional effect is negative. (You may argue that this is desired, but I am not making any claim about that yet so it's irrelevant.)

Given this negative effect, let's enumerate what it causes in turn:

1. It discourages people from asking about difficult problems that are less likely to be solved. Stack Exchange is about expertise. Expert questions should be encouraged.
2. It punishes people who have in fact asked expert questions that haven't yet been adequately answered. Not only with feeling rejected or whatever, but in not getting further answers.
3. As a result of the previous two, it pushes away experts who want to do more than answer, or want to participate in a site where they and other experts could do more than answer.
4. It pushes away people who want to participate in a site where people are treated with respect indiscriminately.
5. It aids and encourages rep whores. Yes, everyone is free to participate as they wish, including playing the rep game. But we should encourage excellent and constructive behavior, not just acceptable behavior. Encouraging people to give all the attention to the easy-rep questions doesn't make this the valuable site for real development issues that we want it to be.
6. It pushes away new users who weren't aware of the feature until they received a nasty welcome.
7. Users who see negative comments think they're acceptable.

Now yes, there are users who just come to leech solutions and don't care to accept answers that they should. Should they be punished for having a low accept rate? Some points:

1. Their questions may still be useful to others if answered.
2. It's easier to use a throwaway account each time than get engaged in the site; the punishment may be ineffective.
3. Have you seen how many people continually post crap without learning their lesson, even if their questions go unanswered? Again, the punishment is ineffective.
4. You'll always have people like me who will answer a decent question if they can regardless of who asked it, why they asked it, or whether they've accepted previous answers. The punishment is ineffective.
5. If the punishment's ineffective, the only point in executing it is vindictiveness. I don't want to participate in a site that promotes vindictive behavior.
6. How can it be valid to ignore questions but invalid to ignore answers (not accept them)? This behavior seems inherently hypocritical to me. We don't stamp "ignores questions from low accept rate users" on user cards, why should we be stamping "low accept rate" on user cards either? Questions require effort and have value too, not just answers. Ignoring a question due to an external factor harms the site.

Yes, we should always encourage people to leave polite and constructive comments. But no matter how we encourage good behavior or discourage bad behavior, these comments will always be made if the accept rate is displayed. When it comes to the accept rate, there's a simple and effective option to reduce abuse. Get rid of it. The downsides are prominent and the upside is questionable as to its existence and entirely dubious as to its ethics. I certainly feel that the benefits of removing it greatly outweigh the benefits we purportedly get from it.

---

Some examples of how useless the accept rate is:

• Mentioning the accept rate is considered so unconstructive that mods will delete such comments and so prevalent that flagging them auto-deletes them

• Skipping questions from users with a low accept rate is discouraged

• The rate gives no indication of why it is what it is:

  • "There are occasions where accepting isn't practical".
  • Some users accept answers solely to increase their rate
    
    

• "[A]ccepting answers is not required" (unless you want people to treat you with respect and answer your question, apparently)

Answer 5

Before reading this I had no idea what "acceptance rate" meant - I had assumed that Stack Overflow was telling me how many of my answers had been accepted by others; I had no idea it meant how many questions I had accepted answers to.

So when someone asked "Why the low acceptance rate" on one of my questions I had no idea what they were talking about. If they had not done so I would still be in the dark.

But there is a really simply solution!

When someone starts to ask a new question, if their acceptance rate is below X% have the website show a red warning message above the Title input which states:

Warning: You have accepted answers to only Y% of your questions, this may deter people from offering answers.

Answer 6

I have always found the accept rate display to be stupid, useless and prone to causing noise and fighting instead of promoting better questions and answers.

There are people who go around following people with low accept rates and commenting things like "Go accept answers or I won't answer your question" and other garbage.

Answer 7

The accept rate is a proxy for a person's ability to ask good questions and encourage good answers. With a large enough statistical sample, I think it's a valid proxy. If someone asks 100 questions, and accepts only 5 answers, the odds are that there's a problem.

People providing answers on Stack Overflow are volunteers. I don't get paid for it, and neither do most of you. (Yes, yes...SO has employees. Don't get sidetracked.) A volunteer's time is a valuable and limited commodity, and anything that ignores this basic premise is doomed to failure from a psychological standpoint.

There's certainly a valid case to be made that small samples lead to skewed results. If you're a new user with only 2 questions and 1 accepted, that leaves you with a 50% accept rate. Perhaps a red/green metric with a minimum threshold of asked/unanswered questions before it can go red is a better option than a displayed percentage.

In the end, though, people will apply a metric whether you provide them one or not. Whether it's accept rate, overall reputation, or upvotes/downvotes on a question, many people will anchor their evaluations on something before determining whether to read a question in-depth. You won't be able to remove this basic instinct; at best, you will shift it to some other discriminator.

I'll upvote any sensible suggestion for improving a proxy metric, but I simply can't agree with anything that ignores basic sociology or psychology. Removing the accept rate just moves the cheese; it doesn't get you out of the trap.

Answer 8

I was going to propose this, then I ran out of time, so I'm glad to see it's been proposed and generally welcomed. In the context of an otherwise fine question, a comment such as this:

You really need to work on your accept rate. Go accept some answers on your other questions if you want an answer for this one

... is little more than a stain. Additionally, 'accept rate proding' (aka bullying, aka nagging) is one of the largest sources of comment flags we receive. They are either just noise, or obsolete noise when the person does as asked. Sometimes, people accept answers that didn't really help them just to try to comply. That's unfortunate. Displaying it,at least in my opinion seems like a tacit endorsement of the problematic behavior it encourages. That's also unfortunate.

I would like to see the actual calculated metric vanish almost entirely. People can still take a glance at your profile and see how many questions you've accepted answers on, but doing so would probably take more effort than just answering the current question to begin with. I don't see anything wrong with showing it to the question author only if it's consistently low.

Anyway, as Bill dropped a 100 rep bounty on this, I'm going to chime in and say that moderators would not be at all displeased if that metric went away.

Answer 9

About the only context in which the accept rate is marginally useful is when it is at 0%: in the overwhelming majority of cases it tells you that the poster is not aware of the purpose of the check mark outline next to the answers, and may benefit from a brief explanation. Once the user accepts an answer for the first time, mentioning the accept rate is almost certainly counterproductive.

However, the absence of the "Scholar" badge already tells you that the user does not know how to accept answers, so displaying the accept rate is redundant.

Moreover, the process of making the user aware of the "accept" feature could be automated with a gentle reminder to consider accepting an answer, if this is the behavior the sites would like to promote. A very gentle reminder can be sent only once to users without the "Scholar" badge, and only if they have at least one answer with the score of at least two that has not been accepted for three or more days.

Answer 10

I think accept rate shouldn't have anything to do with whether a user answers a question or not. You're talking about 15 reputation, while upvotes from a good question can easily exceed that.

So since it shouldn't have an effect on whether a user answers a question or not, why not just take it away then? What's the point in having a statistic in front of the user that says hey, don't answer this, move on? I don't think it's a good idea to put a sign post on otherwise good questions that says don't bother.

So, if you have to display accept rate, just put it in a user's profile screen, where you have to do quite a bit of click through to find this number out.

Should this be public information? Arguments go for and against. If it is public information, then only people who really care about accept rate can find this information at a glance, instead of having to really troll through the user's previous posts to find the answer to this question.

Answer 11

Stack Exchange is a feedback loop and there is a difference of opinion as to the entry point. The varying communities would be a fascinating study in group dynamics.

The topic at hand:

There is one theory that views the charity of the says that if we're nice, we'll attract more quality answerers and we'll perpetually be an ever-growing group of experts motivated by simply helping a peer in need. In this theory, acceptance rate is perceived as ammunition to attack a user and should be removed.

There is an alternate theory that views the shared community values as the entry point. This one says that the extant experts are all too often left unrewarded for their volunteer efforts. In this theory, "work on acceptance rate" as an example of coaching a user on how to be a "productive" member of a community. If you want to be a member of the community, there is an initial grace period, but eventually you need to accept and espouse the community's values.

I lean slightly toward the latter, honestly. I am not an expert. I am not a power user. I am here to learn and I understand that the folks here are not indebted to me to answer my question. As such, I wear my acceptance rate with pride because it shows how much I value this community. That is how I understand Acceptance Rate. It is a metric that demonstrates the community's value to me.

Answer 12

Further to the answer by jcolebrand :

Show the accept-rate to the owner of the question only, and only on question user-cards. That way they get a gentle reminder that they're rewarded for ticking the box, and nobody else has to see it.

And move that public metric to their profile

An automated peroidic prompt when the answer accept rate falls below a given threshold.

Answer 13

I think it is a good motivation for users to reward and identify correctness in answers, but would prefer to instead see:

'8 unaccepted questions' as opposed to '74% acceptance rate'

this would inherently remind users they have a finite number of questions they can go back and review to accept/bounty, and may encourage others to view those 8 questions (assuming that number is a link, which I believe it should be regardless of the phrasing)

so rather than a percent which I view as a grade (out of 100%) , it should have more of an informative todo list for the user.

74% unaccepted doesn't mean much if you've only asked 4 questions (other than incorrect math)

100 unaccepted questions is a lot even if you've asked 1000 questions here...

I hope this makes sense to someone else!

Answer 14

The problem here is the actual accept rate, not the display of the accept rate. If people asking questions cared enough and were courteous to their fellow users, they would accept answers. It's not about points. It's about saying "thank you".

Answer 15

The real, root problem is that the ratio of points granted for accepted vs. upvoted does not correspond to the value we place on them as a community.

The questions and answers belong to the community, over time, in a very real sense. And, the community votes - from disinterested third parties - are more meaningful than one vote from the questioner. Since having been chosen as the correct answer already imparts psychological karma, an accepted action should grant fewer votes to the answerer than an upvote.

Removing the visible accept rate hides the user's behavior and makes the community more anonymous. These are the exact kinds of community attributes that result in the worst behavior. E.g., driving in traffic; commenting on the Internet in general. Anonymity and lack of accountability foster bad behavior every time. Public knowledge is a positive social force that keeps societies friendly. Witness the anonymity of large cities, lack of public / social repercussions for bad behavior and the results vs. small towns with cohesive social networks.

Answer 16

I have had a similar discussion recently on the meta site for the SharePoint Stack Exchange site: Acceptance Rate And Unanswered questions.

To resume my post, I was wondering what I should do with some of my old, unanswered questions. I have very few question (about five) so each one has a big impact on the total score. Yet, I still don't feel that I should be "forced" to accept an answer and give the impression a question is "resolved" only because the score felt to 25%.

Have a look at this question for example (yes, it is the same reference I gived on the other meta): Purpose of the <MAPPINGS> CAML element in a custom Choice Field definition. Valid, could be answerable, but yet... no one really knows or can give a solid evidence that the feature was dropped/don't work. I may even say that my question contanins more in-deepth information than the answer, and the information I found are pointing opposite directions.

Now, for a similar question I have the following choices:

• Start a bounty, and probably lose it. It is pretty evident no one know this (except maybe Microsoft, and I'm not so sure...)
• Delete the question because "it is not answerable" with our knowledge. This would mean destroy all the info that it may contain.
• Chose the most voted answer as an aswer, even if it doesn't resolve the problem. I am fine with up-voting useful answer that can help other even if they don't actually answer the question, I am not so fine in marking them as answer while they didn't fix the problem.
• Create a placeholder answer (like I did) to tell others what I have done for now, and then edit/delete it if something better is posted.

As you may see, I have gone for the last option. But this only gives the impression that that question is resolved, which in truth isn't.

I see at least two basic errors in the current system:

• We are inclined to upvote useful answer. This means that if the answer is not a complete solution, but shows effort and helps, we will upvote it. This in turn make it a candidate answer for the accept rate calc script. Trad: I must not vote an answer that I won't accept, and hope that others don't do that too... Pretty lame, isn't it?
• The score doesn't matter. No, it doesn't, but let's display it red, will ya? Please don't look at that 0%, but let's display it, OK? See it? Please ignore me... Sounds like Glados. The point is: if we see that score, we will make consideration on it. If we see that the poster never upvote an accept valid answer we will skip it. It is lame, but it is real.

So now what can we do? Let's think about it. The system should be there to remember people to accept answer, but as now is used by people to decide if they should ignore a question. So we need a way to still remember people they should accept valid answer, but at the same time make it less invasive, so that other people won't get distracted by it.

I belive that hidding it from the question can be a good start, as other suggested. I would actually go further. Make the score more visible based on the ammount of unmarked questions? That is, we have only two options there: if the score should be ignored, then please remove it - forever - and never talk about it again - give the user some message, some prompt but remove the score from the question. If otherwise it should be looked at, then make it more visible the more the user seems to ignore upvoting/marking. An user with four questions (and 9999 answers) and 2 of it unmarked is not the same of an user with 0 answer, 9999 question and 5000 of it unmarked.

(I noticed this question only now, and I hope this won't be seen as necroposting, but I felt I needed to add my point of view.)

Answer 17

Another way to answer this:

What is the steady state of displaying accept rate, but discouraging anyone else from commenting on it?

The steady state is this: New users with a low accept rate will probably fail to get as many good answers to their questions as people with higher accept rates, but won't really know why (because no one is telling them).

Which points bask to the question. Just get rid of the accept rate if it's considered bad form to say anything about it, or even take it into consideration as to whether or not to answer a question.

Answer 18

We have been through this before. The site owners need to make a policy because otherwise, we are going to keep flip-flopping between showing it and not showing it.

The problem is no one is going to be happy whichever way this goes AGAIN.

Do a search on META and you'll find hundreds of discussions on the Accept Rate.

Answer 19

Just a bit of a personal perspective - this is meta, after all.

Funny how the word "meritocracy" has taken on a bad connotation in certain contexts in today's world. At first superficial glance, one would swear that that's exactly what SE is, and what it was always intended to be - a meritocracy with top respondents gaining a reputation par excellence. And rightly so, given the value they add to the state of globally-available knowledge. Productivity and quality communities come first, and thereby, individuals benefit most. Those who spend considerable time putting quality knowledge into SE deserve fair rewards, to encourage them to continue to act, possibly producing more masters (programmers/writers/mathematicians/what have you). "The needs of the many outweigh the needs of the few." So support those who ultimately help the uncounted many.

Sadly, what's good for the goose apparently isn't good for the gander. It is more than acceptable, encouraged in fact, for top respondents to put their time (often considerable amounts of it) into answering questions, but to expect the simple courtesy of a single vote or an accept is apparently somehow absurd because this creates an atmosphere of mutual monitoring. When in society have we ever not monitored others? Respondents are supposed to be altruistic, expecting nothing in return.

FWIW I'm no. 12 on gamedev.SE for the moment, and the number of answers I've given that have received zero attention, even in cases where there are no other answers, is considerable. The same goes for many other top users there, at least those who don't observe the user carefully before answering (you see, perhaps it is I who is the charitable fool that likes to help the brand new users...and I who pays the price for it). This has increased noticeably since the accept rate vanished from easy public view. I've begun to feel that the typical user has the attention span of an orange. I mention my rank only because it implies I'm no idiot as far as the sites have a few thousand points on. I get 0 or if I'm lucky, 1 vote, and those are most often never accepted. Rarely, we hit the jackpot with unremarkable answers and get tons of votes; the imbalance here is utterly disproportionate. I have to explicitly ask, months or years after the fact, for the checkmark on such answers, provided no one else received it or I got the most votes (often just marginally). About 1 in 6 such users then responds apologetically and with thanks. That's rare, regardless of the assistance provided.

What's more, I find it exceptionally odd that I have been able, with no extraordinary effort, to maintain a 92% accept rate on SO and a 95% accept rate on gamedev, for much of my SE career. It's just... common courtesy. Maybe you're not always thrilled with the answers you get, but after some time - several weeks, months or even years - hell, just deliver a reward to those most helpful to you.

In conclusion: Lack of a visible accept rate is gradually making me more and more jaded and less willing to help new and/or low acceptance users, in nearly every instance. I feel making this change was regrettable indeed, an overly simplistic, knee-jerk reaction to a complex and worsening problem. This affects the output of users who would otherwise add considerable value to the SE experience.

The game, so to speak, is broken.

No one can force votes, but public pressure to conform is part of human society, for better or for worse. The general public can and should demand decent and polite behaviour that acknowledges the positive actions of others - i.e. that askers place the checkmark when it is deemed appropriate to do so. While I don't condone abusive remarks, we certainly should bring back the accept rate at least as visible to higher-ranked, trusted users (say at the 5000 points mark) who otherwise use the system with care and respect. This also encourages the starting run which for so many has a snowball effect in bringing them back to the site - good, steady rewards elicit good, steady engagement.

P.S. Perhaps some will say I'm complaining unnecessarily; I'm really not and am very fortunate to have gained the rep I have on gamedev specifically. But given this state of affairs, I will henceforth be far more frugal with the questions I actually choose to answer, as it clear that the average SE user is not being encouraged sufficiently to thank those who spend their time responding.