10

We have several questions on SFF which are of the form:

What is the work/example that is "QUALITY-iest" for a given objective property.

For example, "What is the earliest SciFi book about X", or "What is the largest (engineered) physical object dreamed of in science fiction?"

While in an out of themselves they are good questions, they have an annoying quality:

  • They may have an absolutely correct answer

  • BUT, in the process of discovering such an answer, many intermediate "best found so far" answers are posted.

  • Many if not most of those intermediate answers are 100% invalidated by a later posted examples

For example, for an "earliest" type question, someone posts a work from 1970. Then, someone else finds a work from 1963. Finally, someone else finds a work from 1923.

QUESTION: What should be done about those earlier, currently-clearly-wrong, answers?

The options are:

  1. Leave them alone and let the correct "MOST-iest" answer be upvoted to the top.

  2. Downvote them. This seems somewhat unfair to the posters, since at the time of their posting, they were the best answers. Therefore, this should be done very discriminately - for example, if a better answer was a simple Google search and TVTropes/Wiki article away, downvote is OK. This should have an exception for good quality answers that have useful information.

  3. More drastically, prune them. Delete them completely.

  4. Same as #3 (delete), BUT post a collected union of "could have been a contender" information either as a separate "Community owned" answer; or as an addendum to the correct answer.

I don't have a strong preference for either of #1-4 (but I strongly prefer #4 instead of #3); but think that we should have a community standard policy that can be referred to by people who wish to do the cleanup.

Ideas?

  • 4
    Is there really need for special treatment if the accepted answer is the most correct answer? – Gorchestopher H Aug 27 '12 at 14:00
  • @GorchestopherH - the rest of the answers clutter up the question, somewhat. I don't think it's a major super problem - just an annoyance. Hence one of my options (#1 to boot) is "Do nothing". I was mostly responding to Kevin's chat message on the topic – DVK-on-Ahch-To Aug 27 '12 at 14:02
  • 4
    Since the accepted answer will be at the top, just below the question, and the other answers near the bottom, I think of it not as clutter, but instead as bonus info that you don't need to scroll past anyway. – Gorchestopher H Aug 27 '12 at 14:13
  • Forget about those - people have been posting new answers for objects that are orders of magnitude smaller than existing answers. Those are the concerning ones... – Izkata Aug 28 '12 at 0:50
6

I vote for option 1. Who's with me?

There's no clutter if the accepted answer is directly below the question. The other answers are just added info on the subject.

  • 2
    Option 1. Downvote only if the answer is posted later than a better one. – sjl Aug 27 '12 at 17:04
  • Would we just edit the accepted answer with a more correct answer if/when it shows up then? – ardent Aug 27 '12 at 23:41
  • @ardentsonata Ideally, the asker would still be around to change the accepted answer. Notifications are still given to the asker for new answers. – Izkata Aug 28 '12 at 0:51
  • @Izkata, then potentially the answer might just become community wiki'd, so why not go with option 4? – ardent Aug 28 '12 at 0:55
5

I think bitmask mostly has it in that we shouldn't be up-voting answers that can't prove they are correct to begin with, although I think it's a little more complicated than that. :)

All answers to these types of questions should be verifiable in principle: either X is the Y-iest Y that ever did Y or it's not. Backing up a claim made in an answer by providing evidence of the claim goes a long way in verifying the answer is, in fact, correct.

If an answer doesn't provide that, it's up to the answer to be obvious: no reasonable person, presented with the answer, would claim that the answer is wrong. This is high bar to reach: if a question is so trivial that there's no point in providing any reason why the correct answer is, in fact, correct, it would seem it'd fall under general reference.

If an answer is non-obvious and has not provided any reasons to believe it is correct or evidence to test its correctness, it should be down-voted not because it's potentially incorrect (it could just as well be correct), but because it's not useful.

But that leads to the question itself: if a "question" asks something that can only be answered by guessing, it's not a real question. If it's not at all possible to definitively answer a question the first try (never mind whether someone does), there's a problem with the question. We don't provide value by taking random stabs in the dark.

Insisting on proof and verification should minimize this issue, but I do agree there are potential cases where an answer is no longer correct even though they provided cogent reasons for believing in its correctness (a benign example would be X being the largest structure in a published work until Y comes out in 2013, which features Z that's 10 times larger than X).

In cases like these, there are only two options, which most closely resemble your first option:

  • Edit the top-voted or accepted answer to be correct (ask for CW conversion if it's necessary)
  • Add a new answer that explains why the accepted or top-voted answer is incorrect and explains what the correct answer is

Deleting answers merely because they're now incorrect is opening a can of worms that's discouraged on all sites except for I think Skeptics, which is entirely based around debunking spurious claims and has a rigorous "back it up" policy to begin with. Only diamond moderators have the ability to delete up-voted and accepted answers, and they shouldn't be in the business of deciding which otherwise community-approved answers are correct and which aren't.

Down-voting is a personal action, and the guidance is whether or not one thinks a post is useful, not necessarily whether one thinks the post is correct. If one thinks that a now-incorrect post is useful for whatever reason, they should vote it up. Otherwise, if they think it's not, they should vote it down.

However, in a lot of cases, particularly in the scenario of "this answer was once correct but is no longer", people would've already voted and their votes would be locked until the post is edited again.

1

If this site is about answering questions, especially those that have correct answers... why would we want to discourage them because it's often difficult to answer them correctly on the first try?

You're looking at the "best answer at the times" as if they are wrong. This is a poor perspective though. What they are is a foundation for a later correct answer. As a collective mentality, we're essentially narrowing down the possibilities.

This may make it difficult to decide who scores points, but then, are we here to score points or are we here to come up with good answers? I was under the impression we're doing the latter, in which case the difficulty of good scoring systems is moot.

0

This is a bit backwards. If I ask for the maximum of a collection and you say it's x, your answer is hogwash unless you prove it. Any unproven claim "answering" the question is useless and should be voted down.

It's as simple as that.

  • I think the question is regarding what to do with other answers once the most correct answer is chosen. Not how to choose which answer is best. – Gorchestopher H Aug 29 '12 at 15:08

You must log in to answer this question.

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