What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

In episode 42 of the Stack Exchange podcast Joel Spolsky mentioned something along the lines that some sites (for example, Travel Answers) suffer a little from questions that don't always get useful answers - even when someone has good intentions.

So, for example, someone asks a question about getting a visa. It's tough and a rare visa, but nobody who has been online that day sees it AND knows the answer, so it gets no response.

However, Joe from Kansas sees it, remembers his cousin got a visa just like that, he thinks. He writes an answer that his cousin got a visa for that country, despite it being not quite the same, he thinks it'll be useful.

It's not necessarily useful, but now the question is 'answered', and vanishes into the ether. Obviously, the OP isn't going to accept the answer, as it's not really right.

We do have a few of these on Travel Answers, as well as many, many GOOD answers, to be clear - but what can be done about it?

  • Vote them down? This will make it still 'unanswered with no answers > 0' if we're lucky
  • Delete the answer? Odds are someone else coming along might see it and find it useful in a very similar situation
  • Change to a comment? Some of these are quite detailed (especially visa criteria and laws) that most of us wouldn't know to change it or not.

Any suggestions or solutions that other sites have found?

share|improve this question
Ummm, shouldn't this be something you ask the Travel.SE community? SO is vastly different in this regard, and asking SO users how you should handle scenarios that don't appear on SO is not very useful... –  animuson Feb 8 '13 at 7:00
@animuson: This isn't just a travel.se issue. I've noticed this to a smaller degree on Chem. –  Manishearth Feb 8 '13 at 7:24
It does happen on Stack Overflow, @Animuson. When it does, we flag the answers for not answering the question, and eventually delete those answers. –  Rob Kennedy Feb 8 '13 at 7:26
Just downvote it. Such answers are ultimately not useful and downvotes indicate that reasonably well. –  Bart Feb 8 '13 at 8:05
@animuson Mark isn't asking Stack Overflow users, he's asking Stack Exchange users. This site doubles as Meta.SO and Meta.SE. The issue isn't limited to Travel.SE, it comes up to various degrees on all sites. –  Gilles Feb 8 '13 at 10:59
The OP should comment on the answer indicating why it's inutile. Others, seeing the comment, should downvote it as inutile. –  msh210 Feb 8 '13 at 16:22
@msh210 it's not about individual answers, it's about the problem as a whole, given that Joel picked out travel.se as an example. –  Mark Mayo Feb 8 '13 at 17:13
Recipe from Workplace: downvote mediocre answers (see this discussion at WP meta). Recipe from Programmers: downvote low quality answers and fix the bug in "hotness formula" (see this discussion at Prog meta). –  gnat Feb 13 '13 at 21:25
I would say that down-voting at the workplace doesn't really work well because we tend to get users who "feel bad" for those down-voted answers and vote them back up even though they are worthless. The only exception is when users feel like they can pile on. –  Chad Feb 13 '13 at 22:25
Isn't this what bounties are for? To draw attention to questions that didn't get a satisfactory answer? –  Troyen Feb 14 '13 at 0:20

4 Answers 4

Obviously, the OP isn't going to accept the answer, as it's not really right.

This is not anywhere near as "obvious" in practice as this happens quite a bit.

On The Workplace we have what I consider a huge problem with "not really good answers being common." I wrote at length about this.

The summary of my meta.workplace post is the following quote

I propose that:

  • This site's core users [The Workplace] MUST much more frequently downvote answers even if they are "only" mediocre or we are going to face a consistent decrease in quality

because (each of these is expanded on at length below [here] ):

  1. All (or at least most) our questions are subjective
  2. Anyone can post answers on nearly all our questions
  3. Permissibility of mediocre answers invites more mediocre answers
  4. There is no real incentive to stop posting mediocre answers except DVing

I strongly believe the only solution is consistent downvoting (and comments explaining why) so the site culture does not "allow" this behavior.

The problem is there is a huge culture on many sites - The Workplace in particular - of "oh this answer adds some value, so even though it's not really a good answer or isn't totally answering the question I'll toss an upvote at it" or the drive by "gonna upvote all answers!!" type voting.

share|improve this answer
the problem of sympathy upvotes to low-quality answers is particularly amplified in hot-questions when bug in hotness formula forces collider to bring unnaturally high exposure of the question to wide SE community, making interesting and well presented problems look the same as non-constructive popularity contests: "There are just too many new visitors to keep things under local community control, and there are just too many new voters and commenters to get things going as designed" –  gnat Feb 14 '13 at 13:38
One issue is that downvotes cost rep. –  deworde Mar 25 '13 at 9:49
@deworde: reputation means nothing if you're not willing to lose some of it. –  user7116 Apr 2 '13 at 14:19
@sixlettervariables Arguable (I'd replace "lose" with "spend"), but it automatically discourages downvoting (which is the point). So that means that consistent downvoting is less likely than consistent upvoting. At the very least, if I'm 9 above the boundary that grants certain privileges, I'm not going to downvote 10 questions per day. I do that every day for a month, and it effectively cancels out asking one very good question. –  deworde Apr 2 '13 at 16:10
@deworde ... except downvotes on questions do NOT cause rep. –  enderland Apr 2 '13 at 16:14
@deworde: it only discourages people who are not productive members of the site. –  user7116 Apr 2 '13 at 16:19
@enderland Interesting, is it only downvotes on answers? I hadn't noticed. Makes sense, though. –  deworde Apr 2 '13 at 21:19

I think this is a fine job for the anonymous feedback feature that we already have in place. If the question is asking for information that is extremely hard to find, there's a high probability that anyone else looking for the same is going to find the answers that the question received. If enough anonymous visitors mark the answer(s) as not being very helpful, put the question back on the unanswered questions list until at least one or more up voted answers are posted.

There's also cases where there's clearly a disproportionate amount of views relative to up voted answers, based on other questions on the site. If at least one answer on an average question on the site receives 1.2 votes per 3 views on average, no votes on any answer to another question in 30 views is a signal that the question just isn't getting good answers, which means it's probably still unanswered. This assumes of course that the question itself is not the problem, other than asking for uncommon knowledge.

Obviously, answers that were left in the spirit of being helpful but don't actually answer the question should not be up voted, if they contain really nothing more than speculation then they often are at best a comment. I don't think there is a blanket rule for that, though - it would depend on the answer. If there's more than a trace of potentially useful information in that kind of answer, I'd generally be inclined to leave it.

This does have the down side of potentially putting questions that have a good answer back on the unanswered question list, but I think the idea could be tuned to deal with that.

share|improve this answer
Anonymous feedback is useful (generates sufficient data) only for sites with traffic comparable to SOFU... –  Lorem Ipsum Feb 14 '13 at 0:22
@LoremIpsum I thought about that, but it seemed to me like it could work on sites that get about 5k visits daily. Hard to say without actually seeing the feedback data across the network, though. –  Tim Post Feb 14 '13 at 1:45

Perhaps some of the answer could come in the form of considering a question "unanswered" (in the same way as 'unanswered with no answers > 0') If there has been no answer accepted and the user's accept rate (not counting that question) is above 80%. That would show us that even though there may be answers to the question, the asking user (who normally accepts answers) doesn't consider the question answered.

share|improve this answer

If the expert on Molvanîa comes along and wants to see what questions about that country need answering, all they need to do is do a search for [Molvanîa] hasaccepted:0.

Low-rep users providing well-intentioned but not-useful answers can happen. And it can be annoying. But I don't think they are to blame for better answers not coming along.

share|improve this answer
Very likely, the expert on Molvanîa won't have any idea of how to do that; if we really want SMEs to do this then it needs to be made a lot more discoverable. –  Aarobot Mar 22 '13 at 0:01
The expert from Molvanîa is currently on vacation in San Sombrèro. –  hippietrail Jan 28 at 13:55

You must log in to answer this question.

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