What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 130 Stack Exchange communities.

I'm working with the SE data dump and am interested in voting patterns and answer contributions over time. I know it is established that the majority of questions are answered well (or at least to the questioner's satisfaction) in the first couple of days or even hours, but I note that about 2% of questions receive good answers and a lot of votes much later. I would be interested to hear your views as to why this is - is this an organic process or do new answers receive a boost somehow?

(full disclosure: this is for my PhD research - contributors to this Q will be acknowledged!)

update Here are a couple of examples (Still working on the graphs, so not yet perfect). Each color represents a different answer. Looks like some of the questions were almost before their time, the subject popularity only peaking later on?

Question 295579

enter image description here

Question 300855

enter image description here

share|improve this question
3  
So you're aware, there have been quite a few research questions recently--and the community doesn't always respond well to them. –  simchona Jul 24 '12 at 17:37
1  
@gnat Does the hot-questions tag really apply here? –  Bart Jul 24 '12 at 21:21
    
@Bart if 2% mentioned are observed regularly then sure - per my understanding that's how hot-questions are supposed to work: giving regular "popularity pushes" like this. Otherwise (if peaks happen at random), I'd be inclined to suggest "tags" like twitter or reddit –  gnat Jul 24 '12 at 21:35
    
@gnat I think the 2% refers to near-comatose-questions suddenly popping back to life, rather than very popular high-attention questions, which is how I have always understood "Hot". But my definition might be wrong. That's why I asked. –  Bart Jul 24 '12 at 21:39
    
@Bart I doubt this is the case here - near-comatose wouldn't be getting good answers and a lot of votes. Although... the question lacks details indeed - I'd downvote it for being too fuzzy but I am out of votes for today. Maybe tomorrow –  gnat Jul 24 '12 at 21:42
1  
@Bart et al points taken, I am adding some graphs that I've been working on. Looks like some of the questions were almost before their time, the subject popularity only peaking later on? Oops not enough rep to post images yet.. –  paulusm Jul 25 '12 at 21:46
2  
This is exactly why I'd like to have referrer information made available. –  Joel Coehoorn Jul 27 '12 at 13:57
1  
maybe you should note that these questions could be characterized as subjective, because there is no real answer to the questions. –  örs Jul 27 '12 at 14:02
1  
The second question isn't a real question at all, it's a Poll. –  user7116 Jul 27 '12 at 16:04

2 Answers 2

up vote 5 down vote accepted

There could be a lot of explanations for this pattern, and I don't think you will capture all of them using Data Explorer (though you might consider checking if any of these "late answers" had a bounty added to them after the initial rush of answers proved inadequate).

A couple that come to mind immediately:

  • The user posts to some other forum or twitter the link to the question and asks a wider field for help

  • A very knowledgeable user only visits certain tags, or the site in general, once a week

  • Someone who visits often might have missed the question when it was originally posted, but came across it later

  • A good answer might have actually taken a lot of time to put together, and/or the answerer got distracted by other things. He/she may even have had a draft of the answer open in their browser for days before believing their answer was ready to post

share|improve this answer
2  
Regarding the 3rd point, the automatic bump might factor in there perhaps? –  Bart Jul 24 '12 at 17:56
1  
There is also the newsletter. Once a week, an email goes out with links to something like 8 questions, some unanswered. I suspect that brings disproportionate attention sometimes. –  Ernest Friedman-Hill Jul 24 '12 at 18:01

One thing that can be attributed to a sudden increase in votes on a post could be the timezone consideration.

An "ok" post might be posted during a certain timezone's most active time, get an upvote or two and then get a big boost when the next timezone wakes up and starts checking their Stack Overflow accounts for any rep or badges they won over night :P


Even a small change to the posts title could suddenly bring in many more views and therefore possible voters. As we all know, editing a post also "bumps" it to the top of the active tab for the relevant tags. A few upvotes will make a post look even more attractive hence bringing in more views and voters. It's a rock slide effect - gaining momentum as the votes continue to come in.


One final consideration that might lead to a post getting a sudden increase in popularity is it getting attention from a high raking veteran user. Take this screen shot for example -

                The stories are real but the names have been changed  


I am speaking only for myself here - but sometimes when I see that a high rep user has given some post his attention, I'd also like to see what they have contributed - a little bit like the rock star effect. "If it interests them - it might interest me!". This is another thing that would reel in potential voters.

share|improve this answer
    
So it might be a high rep user contributing, then publicising their answer through blog / tweets etc? - as described here blog.echen.me/2011/09/07/… –  paulusm Jul 27 '12 at 15:04
    
I don't think that it matters who takes it out of the site- after all, your rep doesn't leave the site :P High rep users are considered skilled and as such their actions might be interesting to others. –  Lix Jul 27 '12 at 15:08

You must log in to answer this question.

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