The Workplace is naturally attractive to soft questions, and soft questions, more often than not, attract a lot of answers. That's good, but after a point answers start becoming horribly repetitive, adding very little, if anything, to the discussion. It's a problem we've identified before, on various Meta discussions:

Our current Area 51 stats shows that we have 3.8 answers per question, which I think is a very healthy ratio, but we shouldn't let it raise a lot. Programmers (then NPR), had 7.0 answers per question at the end of beta, and although the stats label the ratio as "excellent", it's not, it was one of the warning signs that the site was turning into a forum. And we all know what happened to NPR (for those who don't, NPR is no more).

One good way to keep our answers per question ratio reasonable, I think, would be to lower the Community Wiki automatic conversion threshold to 10 answers. Right now, the threshold is 30 answers, the default for all Stack Exchange sites, and some sites have opted for a lower threshold, Programmers and Super User are at 15. The goal of reducing the threshold is obvious, remove the reputation incentive to post yet another answer at a point were it's extremely unlikely that the question hasn't been sufficiently answered. Of course, if you really want to answer the question, you still can, the only difference being that you don't get any reputation. You'll still earn badges though ;)

In the Meta discussions I've linked to, and various other discussions in chat it has been argued that answers can sometimes play a deciding role on closing a borderline "not constructive" question. If a question is borderline to begin with (and most of our questions are), getting a ton of repetitive answers will most certainly lead to closure, curating and safeguarding an extremely popular question with 15+ answers is an extremely tedious job. Lowering the Community Wiki threshold will possibly be an easy way to avoid all these "me too" / one line answers.

These questions would have been wikified if we had a 10 answer threshold from the beginning. Notice that I'm searching for questions with 11 answers, a 10 answer threshold means that the eleventh answer triggers wikification. 6 questions out of a current total of 306, this is mostly a proactive measure for bike-shed questions than something that would massively affect the site.

Further reading:

Thoughts?

share
4  
Y U WANT TO STEAL OUR REPZ? </proactive> –  Yannis May 30 '12 at 4:58
    
I think the answer is self-explanatory. 10 is okay. –  lamwaiman1988 May 30 '12 at 10:01
2  
So would bad answer #10 eliminate the rep for good answer #1? We need a system that penalizes the new answer that adds nothing, not one that penalizes existing answers just because too many people piled on later. –  Monica Cellio May 30 '12 at 14:52
1  
@MonicaCellio It's not a penalty, it's removing an incentive. Answer #11 (one over the threshold) would turn the question to CW, yes. We have no way of knowing if answer #11 is a bad one, just that it's the eleventh, it could be a great answer and answers #1 - #10 were the bad ones. –  Yannis May 30 '12 at 14:57
1  
@MonicaCellio Also: Moderators can remove the CW status on the rare occasion that all 10+ answers are great and deserve all the rep in the world. That said, I went through all our questions and I haven't found a single one that would benefit from 10+ answers, and the ones that already have 10+ answers should be turned into CW (imho). –  Yannis May 30 '12 at 15:41
    
I started to comment but decided it should be an answer instead. See below. –  Monica Cellio May 30 '12 at 17:00
add comment

3 Answers 3

up vote 10 down vote accepted

After the discussion here brought this back to mind, I've dropped the threshold to 15.

10 seems dangerously low to me, in the range where it starts to be ripe for abuse, and unnecessarily stressful to moderate. Consider these 19 questions that would've had over 10 answers if some answers hadn't been deleted:

At 15, the number drops to 3. There are definitely some questions in that query, that I think would be better off with more curation and less kibitzing, but there are also a few that've seen quite decent late answers come in at #9, #10 or #11 (and some that should probably just be closed, but that's a separate issue).

Regardless, please review this list of questions - if there are any in it that shouldn't be CW'd, clean them up before another answer is posted.

share
add comment

To be blunt, our current tools for dealing with bad answers are terrible.

We have downvotes but no one uses them, and one upvote erases 5 downvotes in rep. We have post notices (citation needed, ect) but they're mod only and we can't practically apply a "citation needed" post notice every time a totally unsupported answer pops up. Deleting the extra answers isn't a good solution either.

What we need is something to discourage excessive, low quality answers before they're posted, and I think a lower Community Wiki threshold (and a stricter community tone) can help do that.

The lowered Community Wiki limit would help limit the problem of "me too" answers resulting in easy rep and hopefully cause posters to think more critically before answering another question that already has 5+ answers. I think this could be a helpful tool, if an indirect one, in the struggle against bad answers.

I support this.

share
2  
+1 Thanks for saving me from writing an answer. :) I also support this. –  jcmeloni May 30 '12 at 11:06
    
Will it? Or will it just cause more questions to get converted to community wiki that don't need to be converted? How exactly is everyone going to know that this site is different? I certainly don't go browse around Stack Exchange sites I'm just diving into for all the tiny rules about it like what the community wiki thresholds are... –  animuson Jun 16 '12 at 0:42
    
@animuson for the most part it wouldn't matter at all; most questions don't get more than 10 answers, and most answerers don't answer all the things. But for those few who do answer every thing in the world (there are some) this would be relative visible. Generally that's all we want to stop. –  Rarity Jun 16 '12 at 2:03
add comment

To pick one example, it seems harsh to me that the person who wrote the top answer for How can I overcome "years of experience" requirements when applying to positions? (which seems to be the 5th answer given, but by far the most popular) should be stopped from accruing points over time because I suspect that will be a question that gets hits, and helps people, for a long time to come. It is still in the recently active list now.

I also think it's unfair to suggest that all of the answers past #10 were of the quality of "Me too".

And finally, I think it's unreasonable to assume that lowering the bar on auto-CW reduces the number of those poor-quality answers. Those answers are generally given by people new to the community, who haven't a clue what Community Wiki is, how they will trigger it, and what effect it has on others.

I really don't see what you're trying to achieve here.

share
2  
In the end our goal is to create high quality content, not keep a big open game for as much rep as you can grab. In most cases rep works just fine as a motivator, but especially in the case of exactly that sort of popular question, people pile on with low quality, repeat or unhelpful answers. By removing the "free rep" motivator to answer every question ever we encourage only posting when you have a good answer, not just an opinion, as piling on opinions will just bring you closer to CW status. This has worked before on Programmers. –  Rarity Jun 10 '12 at 17:12
1  
@Rarity: Respectfully, I disagree that it has worked on Programmers. I have 6 answers on Programmers rated over 100. 4 have been auto-CWed and another one would have been if they'd reduced the cap to 10. All of them have more than 10 well-thought-out answers. I would say that I'm not particularly driven by karma points, compared to some, but there is something satisfying about seeing someone upvote a question I answered a long time ago (more than one I answered today). That encourages me to pay attention to the quality of my answers more than anything. –  pdr Jun 10 '12 at 17:32
1  
I also think it's unfair to suggest that all of the answers past #10 were of the quality of "Me too". Well, did you read them? I did, and I'm not suggesting that all answers past #10 were sub par, just that there were more sub par answers past #10 than usual. I really don't see what you're trying to achieve here. Find me one question on the Workplace or on Programmers that hasn't been sufficiently answered after 10 answers, and I'll be convinced that this isn't necessary. –  Yannis Jun 10 '12 at 20:12
1  
@YannisRizos: I did read them, although it's difficult to keep track of which order they were posted in. I do know that I upvoted a very valid answer that would have been around the #10 mark. But I think you're missing my point. I wouldn't object to the cap if I could see how it solves the problem. People still keep posting way past 15 answers on Programmers for a popular question. –  pdr Jun 10 '12 at 20:35
    
@YannisRizos: Duh. I never noticed you could sort them before. And you're right, little after answer 10 was useful. But I equally suggest that the first 4 answers all said basically the same thing. The highly-voted answers were posted 5, 6 and 8 (not so far from the line) and the people who posted past 10 got very little benefit from doing so, so an auto-CW cap would have been largely irrelevant to them. –  pdr Jun 10 '12 at 20:57
    
The first 4 answers being the same thing is part of what we want to avoid, and this is another tool for that. We want to discourage blindly answering all questions with low quality answers. –  Rarity Jun 10 '12 at 23:34
2  
@Rarity: Ok, I buy what you're trying to achieve. I'm not disagreeing that it is a problem. But is there any evidence that auto-CW will do that? You know what would be a very effective deterrent? Downvoting or deleting duplicates. Maybe auto-lock the question after 10 messages and flag it up to a moderator who could delete duplicates and decide whether to unlock it. –  pdr Jun 10 '12 at 23:54
    
We look over all questions with that much activity, but the problem with deleting is unless it's a complete non-answer (it's a comment or irrelevant) we don't have much grounds to delete, and more generally we really can't read every answer ever. I've tried to get some sort of policy on duplicate/useless answers but so far we really don't have anything actionable, and downvotes alone don't work well –  Rarity Jun 11 '12 at 2:49
2  
@Rarity: But, if you locked a question at 10 answers and people desperately wanted to add something, they might start reacting to earlier dupes. More likely they just won't bother answering, which is what you were trying to achieve. And I still contend that auto-CW has no effect on that problem. –  pdr Jun 11 '12 at 8:19
    
Locking would prevent people from adding valuable content like workplace.stackexchange.com/a/1760/42. It just removes the reputation reason to add answer after answer, not the actual ability to do so. Plus we'd probably get tons of angry hate flags and edits if they were auto-locked, new users get mad enough when a popular question is Protected. Plus if it's locked, you can't do anything to the answers anyway, the whole set of Q/A is totally frozen except to mods –  Rarity Jun 11 '12 at 16:28
1  
@Rarity: Hey, it was Yannis who suggested that anything after the 10th answer was useless ;P. But seriously, that was a 5-second idea, I'm sure it can be improved. The point is that auto-CW ONLY stops the earlier responses from gaining points, I think you'd be hard-pushed to find a response after #10 with more than 5 points. So how is it ever going to be successful in your stated goal: "to create high quality content"? –  pdr Jun 11 '12 at 16:59
1  
It does'nt stop them from gaining points, just future points. For the most part when this happens the question has already gone crazy popular (like your example) so the top answers have already gotten a few hundred rep out of it--and often that rep is mostly a function of popularity, not quality of the post. Basically rep is already broken in the other way. –  Rarity Jun 11 '12 at 17:09
    
@Rarity: Not true. Those votes are usually capped at 200 points anyway, and rightly so. Your argument is exactly why I say that it's points gained long after the response that give me satisfaction, not the ones gained on the day, because it means that it's still helping people. And those are the points you're blocking. And still, that would be fair enough, if it had the effect you are claiming it would have. But I don't see how it can do. –  pdr Jun 11 '12 at 17:18
    
@pdr So how is it ever going to be successful in your stated goal: "to create high quality content"? Removes the incentive for those of us who are only here to rep whore, and encourages downvotes on answers (they are free on CW posts). –  Yannis Jun 21 '12 at 4:47
1  
@Rarity: On the flip side, it also makes it more tolerable to post an XKCD or Dilbert link, offer nothing new to the debate, and not worry about the downvotes. –  pdr Jun 29 '12 at 16:30
show 3 more comments

You must log in to answer this question.

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