19

Especially in light of the recent changes to reputation points for questions, it seems like this site encourages (via positive reputation) quantity over quality. Especially when one could receive a net positive reputation score with a downvote-to-upvote ratio of 5:1, it seems like we're encouraged to give as many answers as we can spit out, and ask as many questions as we can think up, regardless of quality.

Just to test my theory, in the last day or so, I wrote a lot of bad answers. Not wrong answers, mind you, but bad answers, in the sense that they are unsourced, simple, and not written well (but not incorrect, from what I understand). The type of answer that I would usually put in a comment, but typed in the text field below the one I usually use.

Lo and behold, from the 8 "bad" answers (only three of them have sources!), I have (so far) 16 upvotes and 5 downvotes, for a net gain of 150 reputation points. While in the rest of the this month (so far), I have two answers which (I believe) were well written (at least, relative to the answers given in the last day or so) that got a combined 7 upvotes and 0 downvotes, for a net gain of 70 reputation points (and it probably took me longer to put together those two answers than it took me to write all 8 others combined). So if fake internet points are what the site uses to encourage asking and answering, it seems like the site encourages mediocre questions and answers over fully fleshed out questions and answers.

Does the setup of Stack Exchange encourage quantity over quality, and want us to just throw as much as we can on to the site and see what sticks? Is there something that can be done to prevent "bad" questions and "bad" answers and encourage "good" questions and "good" answers?

  • I suppose it's not a problem that questions get 10 points for an upvote, as the quality of questions is definitely getting better. The problem is this 5:1 ratio, which you mention. – Kazi bácsi Nov 27 at 17:47
8

All 8 of your recent answers are to questions from the same one user. That user has nearly 20,000 reputation, yet his average answer score is negative and his average question score is barely over 1. This does indicate that one can post mediocre or worse content (as defined by post score) and still wind up with lots of reputation by churning out enough posts.

However, this is presumably not so common – many users would probably give up in the face of so much negative feedback. Indeed, a moderator suggested this to that user.

For my part, I think my voting habits have definitely changed since the value of question upvotes was changed. A "good" question and a "good" answer can mean very different things. A question can be good merely by being well-written, noting any assumptions/premises going in, and stating any prior research that was done. An answer can generally only be good if it answers a question. A good question doesn't necessarily contribute any inherent value – it's just a setup for an answer. It therefore feels wrong (to me) to reward a question in the same way as an answer. In the past I would upvote a question merely for fulfilling the above mentioned criteria. Lately I have only upvoted questions that demonstrated value on its own (e.g. a question that itself shows expertise, or that someone can learn from).

Take this question as an example:

"Am I Jewish​?"

There's nothing wrong with the question. In fact it is a very important question. It is a fundamental question about Judaism, and it deserves to have a good answer on this site. But the question itself doesn't contribute value. There's nothing special in it. There's nothing that teaches you something you didn't already know. Anyone could have written the question.

In the past I might have upvoted a question like that. Such a question deserves a high score, and deserves a prominent place on the site, due to its importance. But rewarding the question – which required no effort or expertise – equivalently to the answer which presumably required at least one if not both, seems inappropriate. It would be nice if there was another way to recognize good/important questions, but that does not seem to be the case.

See also Today's change has retroactively changed the meaning of our past votes. Should we be able to change them? for related discussion.

Is there anything to do about it? On a structural level, I don't think so. Stack Exchange has been very firm in its desire to reward question askers, and to make things better for new users in general. It is unlikely that upvote/downvote value ratios will change again such that you stand to lose more than you gain by posting below average content. However, as a specific community, and as individual people, we can certainly do our part to deal with "bad" content. Downvote when appropriate. Leave comments to help improve quality. Post a better answer if the existing one is low-quality. Etc.

Note that this post is solely my own opinion, and does not (necessarily) represent the views of anyone else, or the site as a whole. I expect others will disagree. This is also not intended to put down any individual or group.

  • 2
    For the record, I wasn't intending to single out any specific user, it just happened to work out that way for my test. It might be better to remove the first two paragraphs; the rest of your post still gives over the same meaning without singling any one person out. – Salmononius2 Nov 25 at 1:03
  • 4
    I upvoted all of your mediocre answers precisely because they were the only appropriate answers to the questions asked as asked. – Josh K Nov 25 at 5:48
  • 1
    I think the votes are not so important. Everyone has his criterions. For me to do the best. – kouty Nov 25 at 11:36
5

Reputation is gained by participation in the site, and to a lesser extent, by quality of posts. The reputation system gives more reputation for better answers. But if a significant minority who choose to vote find a user's questions to be good (>1 upvoter for every 5 downvoters), there will be a net gain in reputation, and it will only increase with volume of posts. This was true even before questions were changed to give 10 reputation; it only changed the ratio of the significant minority. From the perspective of the reputation system, writing answers that a significant minority will like and a vast majority will hate is a positive investment. If excellency is encouraged, mediocrity is still rewarded.

If the purpose of reputation were to gauge the quality of a user's contributions, then it could have taken into account metrics such as the average post score. But it isn't. The purpose of reputation, by design, is to reward participation, and to encourage further participation.

This is only a problem if you take reputation to be something other than what it is. Upvotes and downvotes are simultaneously a way of rating a post, and a way to give or take points (and privileges) away from the poster. For voting as a way to rank content, reputation is irrelevant. For reputation as a score, the score is in a game of participation, and therefore users who contribute many mediocre answers are rewarded more than people who contribute a few good ones.

4

I think that this change was made for an other type of site, not for Mi Yodeya. I can easily understand that in a professional site as StackOverflow, a good question is easily defined. For a community of developers, a good question is not somewhat obvious, not somewhat based on a lack of knowledge, somewhat based on false premises. A good question can push to progress experimented people. So a good question is a benefice for the community.

On the other hand, Mi Yodeya is open for a very diverse audience. people who don't know basic topics are welcome and to open a question that asks about fundamentals is good, if it's not a duplicate. 'Is it permitted to eat chicken with milk?" is absolutely correct. For some people the question is not relevant but they can think that it's right to ask it and upvote. A very technical question about rules of Teruma or Tumeat Met may be impossible to decifer for a large part of the community, and is less likely to receive a large audience. A hard Kashia of Rabbi Akiva Eiger will perhaps remain unnoticed. The answer as well may to go unnoticed. And a great pshat of the Beur Hagra may as well be an unnoticed answer. And a stupid question or a stupid answer can compete with them from voting point of view.

So what? Everyone need to do his best and to understand how it works. Mi Yodeya is heteroclite and the classment is not a classment of Chochma. For StackOverflow if someone has a reputation of 600,000 he surely knows programming. For Mi Yodeya it's a bit more complicated than that.

But someone who learns Tora needs to do his best. And even for his contribution in Mi Yodeya, someone needs to learnd honestly and to try to be clear in his mind and his words. The score may help him to understand that he is not right or that he is not clear, and the score may be often a false indicator. Many tests have false positives and false negatives. It's inevitable. So what?

You must log in to answer this question.

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