1
$\begingroup$

I'm sure we've all noticed that there is a big advantage in being first to answer a question, it gives an opportunity for upvoting to occur before other (potentially better) answers are added. This means (I'm now assuming a situation where a good answer comes, gets upvoted, and then an amazing answer comes) that the best answer does not get to the top of the pile and the good-but-not-perfect answer sits as the most upvoted. The most read answer will likely be the one at the top while those further down might get ignored. Further the OP might just accept the most upvoted answer without considering the others properly.

Would it be a good idea to mask the votes on answers for an initial period (1 day, 7 days, what would be good? and after the question or first answer is posted?) so that the order answers appear in remains random for a while allowing a fair opportunity for all answers to be seen first?

$\endgroup$
6
  • $\begingroup$ This is probably not a bad idea. The question for me is if this is really relevant. Most questions get one good answer (I don't know if there is no one to answer them or no one wants to answer them after the first answer), many never get any. $\endgroup$
    – Chris Mod
    Oct 8 '14 at 12:20
  • 1
    $\begingroup$ Reddit does this, and it's great. I find that this is a great idea, but this is probably a discussion for StackExchange Meta. I doubt they'd apply this for only Bio.SE. $\endgroup$ Oct 8 '14 at 21:19
  • 2
    $\begingroup$ You're right, @LanceLafontaine - this would be a better fit for MSE. It would be applied network-wide, or certainly on more than just Biology, if it were implemented. It's been discussed before, too. $\endgroup$
    – hairboat
    Oct 8 '14 at 21:56
  • 1
    $\begingroup$ This isn't always the case. On GL, I often end up answering later, and still get upvoted/accepted. The most contrasting one is this, answered over 2 years later: gardening.stackexchange.com/questions/2932/…. $\endgroup$
    – J. Musser
    Oct 23 '14 at 21:27
  • 1
    $\begingroup$ I personally am glad I can view the vote count, especially as it gives me a good idea of what the community thinks of a certain answer. I guess this may be more flawed and less accurate on Biology, possibly. $\endgroup$
    – J. Musser
    Oct 23 '14 at 21:29
  • 1
    $\begingroup$ I guess it depends on the community. I usually upvote other answers on questions I've answered (hence my sportsmanship badge :P), and still often get the higher score. If the community is voting without looking at the answer quality, that's different. $\endgroup$
    – J. Musser
    Oct 23 '14 at 21:41
4
$\begingroup$

What I have done to prevent myself from falling prey to this effect is to view answers sorted by "active", not by "votes". Of course, this is not currently a viable solution for the larger problem, because it would require most people to make the switch themselves.

There is a UX conflict in choosing the default answer ordering. If the reader is going to vote, you want to show them the answers in a random order, or, if he has already seen the question before, with not-yet-read ones on top, so all answers get the chance to be seen and voted upon. But if the reader is someone who only wants to see a solution, and is not going to vote, it is in their best interest to see the answers ordered by usefulness. StackExchange is known for tailoring the interface for the thousands of one-time readers who come to the site from a search engine and never register, so I guess this is why the "best solution first" order was chosen as the default.

Maybe the best solution would be to distinguish between these two cases, making a guess at when somebody is likely to be interested in voting, and when he wants to grab a solution and go away, and show a different default order in both cases. A good way to guess would be having an accepted answer: anything coming afterwards needs less to be upvoted.

So my proposal is: - if a registered user who has enough reputation to vote views a question which does not yet have an accepted answer, show the answers ordered by "active" by default - in any other situation, show the answers ordered by "votes" by default

This is of course more complicated than it looks on the surface. It throws up the problem: how to show answers to a user who has chosen a different setting than the default one? I guess the team will have to think that one through before they know if it is implementable, it depends partly on how they save the users' default preferences in the database.

$\endgroup$
2
$\begingroup$

This issue is not as serious as you think, given that I asked a highly relevant question last year on Meta.SE.

To quote from the accepted answer:

"What is the proportion of late answers that overtake earlier answers?", it is anywhere between 5% and 35% depending on which factors you choose.

If a late answer is truly superior, it can and will overtake the early, worse answer. This statistic is evidence that the SE voting model works well enough.

Also see the (rather ironic) top-voted answer to a similar question to yours on Meta SE.

This problem was solved 80 years ago. (See here for XKCD-author Randall's explanation).

Of course, this answer is late, so it will never get upvoted :)

$\endgroup$
1
$\begingroup$

If an answer is truly amazing, it will be able to stand on its own. When a new answer is posted, the OP has the opportunity to change their check mark if they so please.

A check mark on a post means that post helped the OP gain the greatest understanding which doesn't necessarily mean that the post is the most eloquent or educational. It is entirely up to the OP to decide which post is the best in their eyes, and if they community thinks otherwise, then so be it. There are many post across all SE networks where an accepted answer has less upvote said then another answer. Additionally, their are post that occurred later then the first but have more up votes then the older post as well.

However, if you think a post is remarkable and is not receiving it's just desserts, then you can take a stand by offering a bounty. You can even specify what post is so great in the bounty remarks. By using a bounty, you will draw attention to the post for 7 days (if you wait the full period before rewarding the bounty) and other users will see this and read that post. From the added attention, this answer is bound to get some votes and then a bounty of your pleasing which could be substantial if you want to shell out 500 of your rep.

Therefore, I don't think masking is needed since we have a bounty system available to us here. I have seen on TeX a user award 1000+ points to an answer by repeated 500 bounties since they thought the answer was that great. With that kind generosity, it will take 100s of up votes to overcome and wouldn't really matter who posted first.

$\endgroup$

You must log in to answer this question.

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