6

http://area51.stackexchange.com/proposals/4451/science-fiction-fantasy

One of the metrics that is tracked is 'Answers per Question' and it's one of the ones where we are simply 'Ok' currently (2.5 being the 'Good' level, we're at 2.4)

The problem I see is that given our current rules, extensive speculation is not encouraged (frankly, the amount of speculation I frequently do is likely borderline). This leads to questions with a single, valid answer being preferred.

This is GOOD. Real questions have answers, and this is not a discussion forum (if someone could make those links, that'd be great).

My question comes about because, well, these metrics and guidelines came from stackoverflow initially, where there are MANY valid answers to most questions. There's usually a different technique or language that can allow many different valid answers.

The same isn't true as often for SF&F - questions that allow for multiple correct and valid answers are often subjective or argumentative.

What do you think? Is this a valid metric for this site?

Notes: 1. I'm not suggesting this shouldn't be tracked or used in evaluating the site. 2. I know this is far from our worst metric - questions per day and visits per day are more problematic. 3. I'm asking this mostly as a way to guage what everyone else thinks this site SHOULD be moving towards, and to get some discussion of it going.

2 Answers 2

4

The Area 51 metrics are of course only guidelines. I doubt it really matters (in terms of launching) whether we are at 2, 2.4, 2.5, or higher.

I agree that this is a poor metric to have prominently displayed (for most sites, not just ours) - or is at least poorly described. There are several problems:

  • All of the other metrics are clearly "aim to be as high as possible". That's not the case here.
  • What's the way to get this number really large? Have a lot of list/recommendation/poll questions. That's completely not the direction that SE wants things to go.
  • If a question has a large number (e.g. 10+) of answers, that isn't IMO any indication of site quality. In most cases, it's an indication that there are a lot of low-quality answers - when was the last time you saw a question on any SE site that had 10+ high-quality answers? This is why we can 'protect' questions.
  • If a question has an answer that's good but missing something, the right thing to do isn't add another answer, it's edit the existing one (when Area 51 was created, not everyone could edit, of course). Another answer is only necessary when it's disagreeing or perhaps substantially more information. I suspect that this might be the root cause of this problem: when only a few users could edit, more answers meant more information; now more information is answers+edits.

IMO a good site should have more than one answer on average, because the description's correct that completing answers (and/or viewpoints) are good. However, it should have fewer than 5 answers on average per question, because this isn't the place to have discussions or arguments, and most questions (a) don't really have room for five competing answers, and (b) won't have more than five answers read.

Unless our value drops significantly below 2 (or rises significantly), I don't think we should be concerned about this at all. As you said, views and questions are much more important.

2
1

Too many answers is bad. Too few is bad too. A healthy question has some competition. Some impetus to improve your answer. If you are the only one who answers the question, there are less reasons to do better than simply providing the bare minimum correct answer. Great answers answer not only the asker's underlying question. But the questions that people 2 years from now are really asking when they manage to find the question on Google.

This is why a little more than 2 per question is the ideal place to be.

As Gnovice pointed out, there are more reasons to improve your answer. However, nothing about having more than one answer on a question prevents you from doing those things. And in the case of bumping the question, more answers means more bumps.

3
  • "If you are the only one who answers the question, there is no reason to do better than simply providing the bare minimum correct answer." This is an incorrect assumption. Editing your answer to improve it bumps the question, getting more views and more votes on your answer, which is one additional motivating factor. Also, some people just like to make their answers more complete, whether or not there's competition.
    – gnovice
    Sep 16, 2011 at 20:52
  • @gnovice, better?
    – DampeS8N
    Sep 16, 2011 at 20:53
  • 1
    If the Area 51 text just had another sentence about "too many answers is bad", I think it would be fine.
    – Tony Meyer
    Sep 17, 2011 at 2:03

You must log in to answer this question.

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