30

I recently complained over at Meta.Programmers that they were upvoting poor quality and often times very short answers. But I also keep an eye out for SU's hottest questions and I see exactly the same problem. Answers being upvoted that aren't worth their length in rep. At the bottom you can see a sample I took from our hottest questions of the past month.

Basically what I'm saying is: If your answer fits in a tweet, it's not really an answer

My problem with all these examples is that they merely show how its done, but they don't teach you how to solve it. When someone comes to Super User asking for help, we shouldn't just point them to the solution, we should make the Internet a better place and teach others how their computer works. Like the Chinese proverb goes:

Give a man a fish and you feed him for a day. Teach a man to fish and you feed him for a lifetime.

So in the future, I hope we could stop massively upvoting poor quality answers and instead leave a comment to ask for more information instead.

With Super User's second anniversary contest coming up, I hope some of you will use this as an excuse to edit these posts into more awesomeness. To make sure that our hottest content is actually worth sharing with your friends and family!


What are some options for transfering large files without using the Internet?

15 upvotes: Have you explored satellite and wireless mesh options - maybe there's one in your area if the local xDSL services are not too hot.

Right, so how does this explain Kyle how to effectively get this setup work? It doesn't even bother to explain what kind of satellite it would be (one in space or on your roof) and expects everyone to know what xDSL is.

34 upvotes: Get reimbursed for the gas, or report the mileage on your taxes. OR Are any "consumer" level internet connections available? They tend to be highly asymmetric in terms of transfer rates, so it might be worthwhile to have an inexpensive line for downloads.

I get that getting a second line is probably a great idea, but as the comments point out there's a lot more useful things you could tell. Like downscaling the very expensive account in favor of the cheap one. As another comment explains, this does add complexity of which the answer mentions nothing at all.


How can I change to the previous directory instead of going up?

A billion upvotes for: cd - should perform the swap you need.

Great, but what if I know need two directories? Do I need to go back and ask yet another question or would it have been helpful to point to the documentation for any other similar commands or just go the extra mile and explain them why the system works that way.


Will my computer slow down if I install multiple operating systems?

72 upvotes: No. Only one operating system can run at a time, unless you use virtualization.

Great that you've answered the Yes or No part of the question, but how about explaining why you can't run two systems at the same time without virtualization. The top answer at least takes the trouble of explaining that dual boot is simply something that only affects your hard drive.


Can Screwdrivers With Magnetic Tips Cause Damage To Electronics?

Its fine if all the answers are nothing more than anecdotal and simply state that its not likely to happen. Only Tom's answer tries to explain what might happen and why, yet he's stuck at the bottom with 3 upvotes, while the 4 answers above him got a grand total of 85 upvotes!


How can I permanently customize the DOS prompt (a.k.a. "command prompt") in Windows 7?

25 upvotes: At the command prompt type: setx Prompt $p$g$_$f Then reopen the command prompt.

Great, but would you mind explaining what setx does and why it was resetting itself each time?

9
  • 5
    I suspect that many people vote for the Accepted answer (just because it is the Accepted answer) without even reading it, and do so without bothering to vote for other answers that are helpful, provide screen shots (as my answer in that question certainly did), are insightful, or add value in other ways. Aug 19, 2011 at 23:43
  • 5
    I also think once people see it has lots of votes, they just throw another on. Aug 19, 2011 at 23:49
  • 2
    Some people might be providing short Answers so that theirs will be first since I've noticed that a lot of folks feel that they should Accept the first Answer to their Question despite there being another Answer that is the same but was posted a few minutes later (probably because it took longer to write -- it could have been started earlier for all we know). It's almost as if there's a feeling of "this is the ethical thing to do" (loyalty isn't necessarily ethical) mixed in there somehow. Aug 20, 2011 at 0:11
  • 6
    Any suggestions for how we can disincentivise such behavior? Personally, I'm starting to think that downvotes on answers should be free as well, so users will think twice before posting just anything to be the fastest.
    – Ivo Flipse
    Aug 20, 2011 at 0:19
  • @Ivo Filpse: Short posts (e.g., under 3 lines or less than 500 characters) could be placed in a temporary moderation queue for users with less than 5,000 reputation points, then if after a certain amount of time (calculated based on reputation points where fewer points equates to a longer time-period) the answer could be automatically permitted if no decision is made by a moderator. (Perhaps users with more than 15,000 reputation points could also be included with those moderators for these purposes.) Aug 20, 2011 at 3:12
  • Downvotes aren't free? I have no idea what that means, I just noticed that downvotes are cast from time-to-time, but that the reasons aren't always clear. If you're going to do something that permits/encourages more downvoting, then I suspect that we'll also see a lot more people asking "Why are people downvoting my Question/Answer?" It seems to be a problem already where people often ask, and many people don't seem to bother to provide any explanation (although I've noticed that downvotes get retracted on occasion in the face of such questions, which I also find somewhat suspicious). Aug 20, 2011 at 3:15
  • 1
    @RandolfRichardson Downvotes on questions used to result in a -1 rep to you too, but not anymore.
    – Sathyajith Bhat Mod
    Aug 20, 2011 at 3:40
  • To me, the issue is when a user, hungry for rep, spots a relatively simple question, they see the opportunity to get some quick upvotes, but that can only really work if they get there first. These users want to be helpful, but they also want to get credit for their assistance. If they spend the time crafting a high-quality answer, by the time they submit it, there could already be three other answers, and they've become lost in the shuffle.
    – acjay
    Sep 16, 2011 at 17:08
  • Randolf's suggestion sounds like a plausible solution, but I would amend that by making the auto-publish delay relative to the time when the question was posted, not the answer. That way, if someone sees a relatively simple question right when it's posted, they know that they have a little bit of time to craft a high-quality response before the floodgates open. This, of course, has the drawback of delayed gratification for the question-asker.
    – acjay
    Sep 16, 2011 at 17:09

5 Answers 5

16

Welcome to Stack Exchange! As you've noticed, voting is not always a good indicator of quality. In fact, trivial (or rather trite) questions tend to attract a lot of visits and thus a lot of votes — this is called the bikeshed effect. Everyone understands a simple answer to a simple question, so these get a lot of views and a lot of votes.

The dominant attitude on Stack Overflow is to encourage the quick and dirty answer. If you want to encourage quality answers, you have to work against the dominant culture. Reputation is detrimental here, because it strongly disincentivises long, thoughtful answers: they take a long time to write, and by that time everyone's upvoted the one-liner posted 30 seconds after the question and moved on. Reputation efficiency rewards the short-winded. To encourage upvoting quality answers, you need to have quality answers, which means people who take the time to write them; these people can't be bribed with reputation, so you need to find some other motivation for answerers.

The other part of the problem is voter attitude. That can be trained, to some extent. Whenever you see a one-liner answer, comment on it. Don't be afraid of being unfriendly: shame the answerers who were clearly trying to gain some easy reputation rather than provide a helpful answer. That won't make you any friends. And, when you see this kind of poor answer, write your own answer. Don't be afraid of providing an answer with the same technical content, but a better explanation. The explanation is part of the answer. If someone complained that you “stole” their answer (I get these from time to time), ignore them. Sometimes you'll even gain reputation for that, I've had more than one asker shift the accepted answer to mine simply because I provided better explanations. Of course, sometimes your answer will be ignored; see above regarding motivation for answerers. And don't be afraid of downvoting answers that may technically solve the problem but don't explain what's going on.

Speaking of downvoting: downvoting costs 1 reputation point. Having a post of yours upvoted gives you back the reputation from 10 downvotes. It's not a big cost. Downvoting is almost free. The bigger problem with downvoting is that it's not much of a disincentive for the downvoted party — it takes 5 downvotes to compensate a single upvote.

Another tool in the toolbox is tougher moderation. If an answer fits in the comment length limit, don't be afraid of hitting that “convert to comment” button! Not always, of course, there are question that genuinely deserve short answers. I would support this policy (which, unfortunately, no SE site is close to following except for Skeptics, which has very particular requirements due to its subject matter)¹:

If an answer lacks explanations, and it's clear that the asker wanted explanations, or that the asker will struggle to implement the solution without explanations, then then answer is not an answer but a method to obtain an answer. A method to obtain an answer should be a comment, not an answer.

Again, such a policy will not make you popular. Ignore the whiners. And don't be afraid of deleting a highly-voted post; if anything a lot of upvotes are a sign that bikeshedding is going on and moderation is in order.

Now, looking at the examples you give:

So, aggressively comment, downvote, flag, delete answers that aren't helpful. I wonder if people are still reading my rant, er, I mean my long answer? Recruit other people to help you, like you've done here. Keep it up. And up. And up. I do think it pays — I've been writing long answers to Unix & Linux questions (and others) for a year, and I think I've helped foster a culture that a long answer is better than a short one. It's not that one-liner answers don't come up and get upvoted, but I've acquired “followers”, who also try to go beyond the immediate question and make their answer a “definitive” answer. I do think it drives the site's quality up.

Good luck!

¹ CSTheory doesn't have such a policy but naturally doesn't get upvoted one-liners. This is an exceptional case, due to that site's high entry barrier.

11

As a user who strives to earn reputation through good questions and answers, seeing this happen actually bothers me. We let perfectly good questions and answers that took users time and effort to create and put together, just slip by, almost ignored.

Let's take a look at "Alternative to Windows Explorer for long path names". It was a well formed question, and then the user josh3736 put in a good detailed answer. How many votes did he get? 6.

Now lets look at "How can I change to the previous directory instead of going up?". I hate to use names, but, Bryan here got 173 upvotes for writing "cd -". The rest of his answer was added by another user.

Bryan earned 600 rep for 3 characters, and a space. Josh earned 60 rep, for putting his time and effort in, and writing out a huge paragraph on the topic.

So what is this issue here? Users who put in more effort are not being rewarded properly. We should be focusing on giving out upvotes to users who put time, effort, and hard work into putting out the best and fullest answer they could, as opposed to giving and undetailed answer so much attention.

Super User needs to focus on rewarding those who deserve it.

As Jeff Atwood said in the comments, we should be looking and reviewing First Question and First Answers, in order to help keep new users earn reputation for good answers.

8
  • 2
    To whoever downvoted this, I would appreciate if you said why. Aug 19, 2011 at 23:51
  • 1
    well, use superuser.com/review to find new awesome answers by new users and vote them up! Aug 20, 2011 at 2:46
  • @Jeff Fantastic idea. That definitely helps keep users here too. Aug 20, 2011 at 3:25
  • 1
    I agree with your points, which I think are excellent and logical (+1). However, I don't see why there's a problem in the first case because it appears that josh3736's Answer has exactly the same number of upvotes as the Question (currently 7) -- when I vote, I can only vote once (I'm pretty sure this is how it works for everyone, except perhaps moderators). What am I missing on this point? Aug 20, 2011 at 3:32
  • 1
    @Randolf i do see what you mean, but look at the cd question. 100 upvotes on the question, 173 on the answer. I believe the numbers aren't related. Just because there was a good answer, doesn't mean the question was good. (In general, not for this case since I'm outlining cd - as our bad thing) Aug 20, 2011 at 3:40
  • The Q100:A173 ratio does seem unfair to me, especially because it looks like the Questioner put effort into their writing (and even touched on a "memory" solution of some sort that isn't currently working for them). My guess is that people read the question, and then the answer, and then vote only for the answer. I don't know if this is the case though because I've noticed on that when I click on the upvote arrow that it doesn't always work (rare); when I come back to the question a few minutes later, I see that my vote didn't count, so I click again and then it stays. Aug 20, 2011 at 3:47
  • 1
    @Randolf I felt it was more people just piling votes on simply because it got votes fast, and then the votes on the question were an after thought Aug 20, 2011 at 3:50
  • I bet there's a combination of all these factors and lots of other stuff we haven't thought of, which should prove to make this a challenging matter to resolve. Aug 20, 2011 at 5:30
9

This is all so hypothetical. Let's make it clear:

  • You can't tell people not to vote up something they see.
  • You can't tell people what to vote for (i.e. the "usefulness", the "lack of context")

Thus, if you want to stop bad or poor posts from being created in the first place, you need a stronger reinforcement.

What kind of reinforcement do we have?

  • Commenting
  • Downvoting
  • Flagging as "not an answer" or "low quality"

What if an answer is bad, but (partly) useful?

If an answer is bad (i.e. short, lacking context, lacking background), but probably correct or partly useful, it won't ever receive a downvote in our system, because it's not "not useful" per se (as indicated by the downvote arrow). Downvote it, leave a comment, and you'll be soon experiencing the rage of somebody who just wanted to help and doesn't understand why they got a downvote.

Thus: If you want people to write better answers, be obnoxious unpleasant and tell them.

It would be great to expand your answer – it is currently lacking context [add other points here]. Please look at our short How to Answer guide to see how you can make your answer better.

This means telling them repeatedly. This also means not immediately making friends, but that's not what we're here for in the first place. Super User is not a social network.

I haven't seen these types of comment very often and it'd probably be about time to make use of them.

We should particularly watch new users who aren't used to the concept of our site and gently remind them how they should write answers (example here and here). Edit their posts, show them what it could look like and leave a positive example.


What if an answer is bad and not useful?

This is such a big problem. We hardly see downvotes on answers ever. Why is that?

Users don't want to lose reputation because of someone else's fault.

So I see two ways:

  • We flag those answers (but they'd have to be really bad)
  • We make downvotes on answers free (just like we did with questions, where I saw a positive outcome from the change)
  • We try to care less about our reputation and downvote more often.

And this is where I'm stuck. I don't know if "free downvotes on answers" has been proposed already, but if we want to tackle this problem, this is where I suppose we should start to rethink. If (long-time) users were more "afraid" of downvotes, they'd write better answers from the start.

0
-2

It's a bit intimidating to answer a thread on short answers....

One thing I notice is that there is nothing in the FAQ regarding how to answer, how to vote, etc. From what I see its all emmergent behavior. Perhaps the first step is to begin developing some desired behaviors and document them.

Since I've been on the S.Ex sites I've been corrected on my use of a signature but never been asked to give more thorough answers.

As for my opinion on the topic - this is a Q and A site. Not a Q and Essay site. Posters should be encouraged to read more on there own. So I'm fine with an answer that says "you need to do X and if you want to learn more about the topic take a look at this".

Also, if a poster asks a short "how do I do X" question - are they entitled to an essay in response? Whereas questions where the poster has invested their time the responses may be more enlightening.

-4

I agree with not upvoting crap, and I agree that some extra explaining can be useful in many cases. But several of your examples are just people giving a good straightforward answer where no extra stuff is needed or wanted, and I see absolutely nothing wrong with that.

I don't know about anyone else, but I'm already a super user. I don't need to be taught why things work the way they do, and I don't need to be taught how to fish. When I ask a question on SE, it's because I didn't previously know and currently can't find the answer. All I want is the answer. I'll upvote people who help me get the answer, just like I imagine people upvoted the cd - guy because he gave the answer without wasting their time.

I can guarantee you that somebody who thinks installing another OS might slow down his computer doesn't care what virtualization is. He just doesn't want a slow computer.

Do I think cd - deserves so many votes? No. But it answers the question perfectly, and the next answer has extra details on how to manage more than 2 paths. All the info you need related to that question is right there. No one comes away from that page without their problem solved unless it's a different problem. How is that a bad thing?

13
  • 3
    Most of our viewers aren't even registered on the site, a LOT of traffic comes in from Google. You need to account for those who are looking for a detailed answer, and it comes up on Google. Aug 19, 2011 at 23:33
  • 4
    You would have had a point, if 80% of our visitors were also super users. But a lot of people keep making stupid mistakes, exactly because nobody ever bothered to explain them why things were the way they do. If we do the same, we're just another tech site and not making the internet a better place.
    – Ivo Flipse
    Aug 19, 2011 at 23:34
  • Besides if answer to a question would only be cd -, then perhaps its not a great question to begin with. Surely not something worth hundreds of upvotes.
    – Ivo Flipse
    Aug 19, 2011 at 23:35
  • 2
    @Ivo Don't blame the answerer for a bad question. Simon, you also need to account for those who are looking for a simple answer.
    – user59659
    Aug 19, 2011 at 23:44
  • 4
    Just because the question was bad, doesn't mean you get to post a bad answer as well. But what troubles me most is that some of these answers got 5% of my total rep with one answer. That's just plain wrong.
    – Ivo Flipse
    Aug 19, 2011 at 23:54
  • 1
    @IvoFlipse: In the current system, the question isn't bad as far as I can tell. The only reason you could close it is for the general reference close reason; however, this shouldn't be applied in a literal way. We can't really expect users to have fully read the manual of every command they use, this isn't really a case where we just get to say RTFM and thus this question is not bad at all. Aug 20, 2011 at 0:07
  • 1
    @MatthewRead: On the other hand, the answer is bad and there is no reason to blame the question for being bad as explained in my previous comment. Let me explain why it is bad: Simple, by just saying cd - you are giving the man a fish. He wanted that fish so badly, now he got it. But, where did he got it from? If you don't explain that, the man can't fish for himself. He won't know that you can find such detailed information in the man page... Aug 20, 2011 at 0:09
  • 4
    @IvoFlipse side note, the reputation cap does keep them from earning unchecked from a single popular answer. That aside, I fully agree that the voting is skewed. Unfortunately, 99% of users won't ever see this discussion, because they don't visit meta. We need to somehow get the word out about this - how?
    – nhinkle
    Aug 20, 2011 at 0:25
  • @nhinkle Well we were discussing it in chat and I'm wondering if this should be a step up to free downvotes on answers as well. That way users should think twice about posting poor quality answers, for fear of getting downvoted. The problem users wouldn't read this question and else wouldn't agree either. With more downvotes, they don't get to disagree.
    – Ivo Flipse
    Aug 20, 2011 at 0:36
  • 2
    If all they came for was a simple answer, they'd be better off asking in chat, or else it just makes the site look poor if we're full of one word answers Aug 20, 2011 at 4:16
  • 2
    I'm kinda new around here, but right in the about page it says this is a Q and A site. I know I'd be awfully excited to find out this annoying problem I'm facing can be fixed with one 4-character command. While a lot of info can be valuable to people that want it, especially in this internet age there is GREAT value in having exactly the answer you wanted as fast as possible. Which, I'd just like to point out, is exactly what the original question answer wanted. Why should the answer not be rewarded when it was exactly what was needed?
    – apathos
    Aug 21, 2011 at 19:59
  • 3
    @apathos The difference is: Yes, it should be rewarded, but generally such answers should 1) not be rewarded as much as a more elaborate answer and 2) not be encouraged. If users continue to score with one-line answers, we wouldn't get the great answers. Consider this one here. Moab nailed it in a one-line comment, but the real answer has 1500 words and is awesome. (And thanks for taking an interest!)
    – slhck
    Aug 21, 2011 at 20:22
  • 2
    @slhck I completely agree with you - that is an awesome question and answer. I didn't know that myself, and enjoyed the read. But that question was looking for a long-form answer. The asker wanted to understand the process. But I can also see a great need at times for brevity and clarity. I guess it's just from my perspective that I seem to spend more and more time (on the internet at large) not in finding the information I want, but filtering it out of the mass of results. So when I see a short, direct answer, I find it beautiful. :)
    – apathos
    Aug 22, 2011 at 6:19

You must log in to answer this question.

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