On Stack Overflow, I am noticing a trend of there being almost 10 answers to a question, with an accepted answer, and no one voting (not even the asker not voting for the accepted answer). What's happening?
-
3I agree. I've been trying to only post good questions and useful answers, but it seems quite difficult to get someone to vote for them.– ZootCommented Jan 7, 2011 at 20:01
-
1This is likely a duplicate, but I can't find a suitable one with simple searching.– PollyannaCommented Jan 7, 2011 at 20:01
-
5I've had a few answers go with no upvotes or only very few, and they were just as good or better than other answers. The voting is very subjective and really should not be taken in any way as an offense if your answer does not score as high as you feel it should, but I must agree that if you take the time to answer thoroughly and accurately, why is it so hard for people to click the up arrow?– WillCommented Jan 7, 2011 at 20:15
-
6Maybe everyone is going for the Electorate badge?– LittleBobbyTables - Au RevoirCommented Jan 7, 2011 at 20:19
-
Has it gone down since the /interesting tab was made the default view?– ChrisF ModCommented Jan 7, 2011 at 20:26
-
@LittleBobbyTables, or folks want others to get the Tenacious badge (which requires at least 5 accepted answers with a 0 score, and 10 for Unsung Hero). So far I've got 2 on SO, so it looks like I need 3 more for Tenacious.– TangurenaCommented Jan 7, 2011 at 20:47
-
The question is about to receive a deluge of upvotes, I have already given it my guilty upvote– abelCommented Jan 7, 2011 at 20:50
-
4people are lazy. simple as that.– studiohackCommented Jan 7, 2011 at 21:20
-
7The only way to really look at this is to examine the votes:user ratio. You can't look at votes:question/answer, because if the number of questions is rising faster than the number of active users, of course votes:question/answer is going to drop. Now, if votes:user is dropping, then there might be an issue.– aleCommented Jan 7, 2011 at 21:21
-
4This does seem fairly anecdotal. I'd like to see a statistical analysis of this, like was done here: meta.stackexchange.com/questions/64631/…– Brad LarsonCommented Jan 7, 2011 at 22:10
-
2@Brad I don't know who made this query but (unsurprisingly) it seems to confirm the obvious -- the more questions you have, the less visibility each can get at the top of the homepage.– badpCommented Jan 7, 2011 at 22:32
-
@radp wow, that's scary.– systempuntooutCommented Jan 7, 2011 at 22:51
-
@radp - However, that is looking at questions as of this moment. Older questions will of course have more views and answers on average because they've been around longer (personally, I've seen a long tail on my questions). If there was a way of querying for only the views and answers a question receives in its first month, that might be a more accurate measure.– Brad LarsonCommented Jan 7, 2011 at 23:32
-
@Brad Fair point. One could do that by integrating the dumps with data from timelines... but it's a far too expensive query for us to run remotely.– badpCommented Jan 7, 2011 at 23:41
Add a comment
|
9 Answers
answered Jan 8, 2011 at 20:40
-
12
-
That's some interesting data. If you look at the other graphs, you notice that almost all of them show downard trends of one sort or another. For example, questions with accepted answers has been downward trending as well. Interesting, closed questions has skyrocketed. I wonder if there's a correlation. Commented Jan 8, 2011 at 21:24
-
That's great info. I think it would be helpful for this question to show the amount of questions per week, answers per week to contrast the vote counts shown here. Perhaps count of active users too. Commented Jan 8, 2011 at 23:36
-
@Mystere: accepts by week trends up, but so does questions by week - so it shouldn't be that surprising that the number of questions with accepted answers decreases. I suspect there is a correlation with closed questions - even assuming a fixed % of crap (off-topic, unclear and plain ol' poorly asked) questions, the total number will be increasing.– Shog9Commented Jan 9, 2011 at 0:37
-
@Jeff: I think both of those graphs can be found by following the link at the top... However, if there's something you don't see you can access the same data Greg has via the data dump.– Shog9Commented Jan 9, 2011 at 0:38
-
@Shog: Oh I know, I'm just saying it would make the answer (as shown here) more informative to see that as well as the vote counts. In other words, include those pictures in the answer. ;) Commented Jan 9, 2011 at 0:44
-
1Oh damn the internet. It is killing the art of statistics. Will statistics become an actual science now?– abelCommented Jan 13, 2011 at 8:30
-
1Nice data to illustrate the question, but I don't think it says anything substantive about why this trend is occurring. Commented Feb 9, 2011 at 8:32
Why aren't people voting as much as they use to?
I think it is because the questions are not getting the views that they did back then and there's a strong correlation of views to votes. (Let's face it) Stack Overflow doesn't have the same amount of users as it used to. (This may not be much of a factor anymore as Mr. Atwood points out, it might just be that it is harder to get views now) Looking at all the older questions, it was generally easier to break the 1k view count just for being out there. With all those views, it is very much likely that there would be voters. Nowadays, questions (at least the ones I regularly look at) seem to get around 100-200 views at most and stops there, unless it's a really good question would get another 100-200 views. That's not to say it will never get views anymore. In all likelihood, it will get a more views as time passes, just not in the volumes like it did when Stack Overflow first started. It seems the only way to get the kind of attention quickly is to ask a crazy question (good or bad) and throw out links to it elsewhere (such as reddit) or get your blog minions in on the action.
e.g.,
Why does this go into an infinite loop? (can't find the corresponding external link)
Where can I learn more about the Google search "did you mean" algorithm? (external)
What are the Windows A: and B: drives used for? (ivo, bit.ly)
Personally, with the above in mind, I know that climbing the SO ladder would take a bit of work providing great questions and answers. Although I'm familiar with many topics here, it's hard to compete with other well established users for getting those answers out there in the hopes that there will be enough people around to vote up without the fear of being accused of posting duplicates. Back in the day, that used to be common practice in many cases, now you're likely to get downvoted instead. As far as questions go, it's hard to ask a good/popular question that isn't a duplicate nowadays because many of them were already asked.
As a side-effect, I don't really browse Stack Overflow just to read questions and answers most of the time, I'm generally looking for unanswered questions hoping to be the first to answer. That means questions that already have answers might not get a view from me, a potential answerer or voter. Part of that is my fault admittedly but you can probably understand the dilemma. If I genuinely see a faster (or better) answer that I would have written myself, then I'd upvote it. Perhaps offer come comments to supplement the material if there were some important points not mentioned. But these answers doesn't get the votes unfortunately since there simply isn't enough people reading them.
answered Jan 7, 2011 at 20:44
-
Maybe it would be a good thing if we granted points for reading? Commented Jan 7, 2011 at 20:50
-
@Kees: I agree that there should be some incentives for simply reading questions and answers, but granting rep for doing so wouldn't be the way to do it. Commented Jan 7, 2011 at 20:59
-
6Jeff, what you're likely seeing is the radical changes we made to the homepage, see blog.stackoverflow.com/2010/11/stack-overflow-homepage-changes Commented Jan 7, 2011 at 21:15
-
@JeffAtwood: While I understand that the changes to the homepage serve as a factor to this, I don't think it is for that reason alone. Times has changed and so has the standards here. The homepage is what brought people here and the questions that kept everyone around. I think it's ultimately the tightening of Stack Overflow's policy on questions and the influx of viewers of the then new site eventually receding to what it is now. Commented Jan 7, 2011 at 22:02
-
I gotta wonder, reading stuff like this... How many up-votes do you expect to get for an answer?– Shog9Commented Jan 7, 2011 at 22:42
-
2@jeff traffic is continually up, way up. We just suppress things that have "sufficient" votes and answers on the homepage, so you may see dilution -- this is a good and necessary effect of letting ALL questions get a chance to get answered. Fairness and all that. Commented Jan 7, 2011 at 22:48
-
@Shog: Are you suggesting that I'm not entitled to have an opinion on this discussion question? I could have gone the other way and tried to be funny but I chose not to (and probably wouldn't be able to get away with it like someone with a high rep would). :P Commented Jan 7, 2011 at 22:55
-
@JeffAtwood: Ok fair enough, I don't have access to see those metrics that you can (and rightfully so). Commented Jan 7, 2011 at 22:58
-
1@Jeff: you're entitled to your opinion... I don't necessarily even disagree. But when you talk of promoting questions on blogs/reddit just to get the view-count up, I'm a bit confused: plenty of questions get <100 views, < 10 votes but still get answered satisfactorily - surely that's the key metric! When you talk of browsing the site for unanswered questions, that sounds like a win for the site to me... If spreading the votes around more encourages that, then it's a win for everyone!– Shog9Commented Jan 7, 2011 at 23:00
-
@Shog: My answer was on the basis that the likelihood of getting votes was strongly related to the amount of views that the question receives. Sure we're getting our answers but (as far as I could see) not getting the views like before. The "question" on the google code only got the votes that it has due to the attention it got from external sites. There was even a question here on meta about that. I'd say that highly influenced my answer here. Commented Jan 7, 2011 at 23:10
-
-
4@Jeff: And I think you're probably correct, as far as that goes... I'm just not convinced that raw views are all that important in the grand scheme of things. I recently got an up-vote on an answer that I'd posted almost a year ago... The question still has under 200 views, with < 10 votes on the question and answer combined. So chances are, that up-vote came from someone else with the same (rather obscure) question, not from someone who just happened upon it while browsing. I feel a whole lot better about that one up-vote than I do about the dozens of votes cast on my more popular answers.– Shog9Commented Jan 7, 2011 at 23:37
-
1@Mystere Man: If you're an expert on, say, MAKEFILEs... and you spend your day watching for [makefile] questions to show up on the front page... You're gonna have a really lousy experience no matter what. But you wouldn't do that, you'd follow the [makefile] tag. The front page has (and had) a nasty tendency to act as a feedback loop, granting ever more attention to questions that were already receiving a lot of attention. That's what has been adjusted - there's nothing stopping you from finding questions you're interested in, regardless of how popular they might be already.– Shog9Commented Jan 8, 2011 at 19:29
-
1@Mystere: probably 80% of the time I spend browsing SO - that is, browsing just to kill a bit of time - I spend on the "Newest Questions" list. I don't have "Interesting" tags, I don't have "Ignored" tags... I see the raw list of questions as they come in. That's always been possible, and hasn't changed. But that's never been what the front page provides. If you thought it was, and that now it's not, you were wrong then and you're still wrong - the idea behind the front page has always been to provide exposure to questions that needed exposure coupled with opportunity for quick amusement.– Shog9Commented Jan 8, 2011 at 20:16
-
1@Mystere: furthermore, you're misinterpreting my MAKEFILE example. I'm saying if you're specifically interested in MAKEFILE questions, you wouldn't watch the front page for them... But you may still see MAKEFILE questions if you're watching the front page. If you're telling SO that you don't want to see [makefile] questions and DO want to see [java] questions, expecting SO to show you MAKEFILE questions is silly - if you want random, don't give it specific interests.– Shog9Commented Jan 8, 2011 at 20:22
Because of the solstice lunar eclipse.
Or just because a lot of folks take vacation around the end of the year.
Or maybe because of complicated sunspot patterns...
Or maybe it's all a matter of perception, and folks vote as much as they ever did, but just not on your posts...
Naw, gotta be the eclipse.
Or... HYPNOTOAD!!!
-
10
-
2This is the only valid answer as long as there is no actual statistical data.– PekkaCommented Jan 8, 2011 at 12:14
Anecdotally I have noticed this too, over the last 6 to 8 months, and i don't think it is related solely to the volume of questions. I think Jeff M has the answer:
I'm generally looking for unanswered questions hoping to be the first to answer. That means questions that already have answers might not get a view from me, a potential answerer or voter.
People have changed their usage pattern. As the easy questions have all been asked a billion times and the questions have got harder and more specific, it is harder to build rep on the site these days, so users are browsing for questions where they have a chance of being singled out as the answer. I would wager that even Jon Skeet isn't getting as many votes on his new answers as he was 9 months ago, simply because users are browsing past the questions.
Maybe the solution is not another badge, but to award a small amount of rep (say 10 or 20) for using up all your votes. I'm sure the guys here can come up with some clever algorithmn for awarding the rep, like the user has to use the votes over a 6 to 12 hour time frame (eliminating from contention all those who use their votes willy nilly in their first hour on the site), and the votes have to be spread across a range of answerers and tags (once again avoiding those who just splurge the votes to get the rep).
-
2Awarding reputation for using all votes per day is imo bad idea. Commented Jan 8, 2011 at 20:51
-
1
-
1@Shog, @Ladi, i agree it is a slightly different approach, but it should only be a small amount, and i think it will promote participation in the site, and more votes being cast will promote more answers being given. Regardless of what form it takes, something needs to be done to change the current pattern.– slugsterCommented Jan 8, 2011 at 21:56
Sometimes I invest a lot of time to make good answer - the complex one which describes the problem and possible solution deeply. Even such answers are in my opinion voted very poorly. I think problems are:
- Less visited tags
- Not too much visits for questions
- Perhaps an answer is at the end too complex and too long
- My writting style
- My English :(
Getting Nice answer is probably the best achievement I will ever get on SO. But I'm still happy to answer questions and helping people - and I guess it is the purpose of the site.
Also I think that reputation is making even more reputation.
-
Don't worry about your English, I'm sure everyone appreciates your answers and it will significantly improve with practise!– TomCommented Jan 13, 2011 at 10:42
Maybe we need a nice badge for people who initiate enough 'close by exact duplicate' rituals.
I've noticed the trend too (anecdotal evidence, I know). A related problem is that a lot of questions which deserve to be closed stay at 3 or 4 close votes / idem for questions which ought to be migrated.
I'm wondering if the sheer number of sites that scrape SO data and repost it could be affecting the equation. When I do a google search on something, frequently a "clone" site like "efreedom" shows the answer before the stackoverflow answer in the pagerank.
How many people are just not getting to SO because they stumble across SO information elsewhere?
The hot questions get a lot of views.
With that out of the way, the rest of the questions get hits mostly from diligent SOers who scrounge through the questions or through Google.
Due to the specific nature of questions that get asked(a lot of common questions may have been covered by now on SO) most new questions aren't likely to be widely searched. This lends itself to two problems.
1) new questions may not get enough answers
2) Answered questions may not get enough views.
A question after being asked, stays on the front page for some time(On SO this time can be really short, due to volume of questions being asked or answered), which is enough for the moderators and diligent SOers to answer the question. The answering of the questions brings it back to the top of the homepage. Now if the question is interesting enough or if the answer given has something which can spark a debate, the question tends to get more answers and continues to stay on the homepage. If this goes on for long it gets into the hot questions list(?). Otherwise once the question is answered the tendency to re-answer decreases and the question disappears from the homepage. The same thing happens if no one answers the question.
The traffic which comes from Google is mostly looking for a solution. This user mostly does not have a SO account. The users that come to an unanswered question generally use the back button on the browser. (Rarely they may answer.) Since the user is not likely to have a SO account, the user is unable to upvote the question or any answers(if they land at an answered question).
Some of this Google traffic gets converted to SO users. A lot of questions thus get asked by first time SO users who may or may not register. Since they do not have any reputation, they cannot upvote answers even if they are satisfied.
I guess, that is the reason for fewer upvotes now that SO is maturing. However that is certainly not all bad. Users which find SO through Google and return often to the site, start asking more questions and thus earn reputation and the ability to upvote, which will balance the upvotes issue and other issues in the long run.
In the initial phase SO may have grown very quickly due to common questions. The next growth phase will be slower, but given SO's hold over most programming related search results, this phase will only strengthen it.
(Personal view.Off the top of my head, no reasearch, no sources.)
The problem may be with the sheer volume of answers.
(waits for other answers equally as valuable to get posted)
answered Jan 7, 2011 at 22:35
