What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 131 Stack Exchange communities.

I used the PostHistory table and ran a Python diff ratio using the first and last edits for both questions and answers. This is supposed to give 1 if the posts are identical and 0 if there are no similarities found.

The mean diff for questions was 0.92, whereas for answers it was 0.94 (they are both close to 1 as many posts are not edited). Moreover 35% of questions have been edited compared to only 20% of answers.

The shape of editing is different between the two, however, with questions getting more smaller edits than answers (questions in red, answers blue dashed):

density plots of questions (red) and answer (blue dashed) diffs below 1

I'm interested in the differences in edits between questions and answers

Disclosure: This is part of my PhD research — answers or useful comments will be acknowledged.

share|improve this question
13  
Questions proceed from ignorance; answers from knowledge. –  trashgod Sep 17 '12 at 20:05
    
@trashgod: Nice! Is that your own? –  paulusm Sep 17 '12 at 20:51
    
Yes; more here. –  trashgod Sep 19 '12 at 1:10
    
Could you show a version of that graph that was normalized on the Y axis? –  Servy Sep 24 '12 at 19:38

5 Answers 5

I believe one of the primary factors that goes into this is that the asker of a question is looking to find a solution to their problem. Generally they will have trouble formulating a concise question and it is often the case that such questions get edited by peers who genuinely want the user to find an answer to their question. Usually the edits correct misspellings but often they also help with better formatting or clarification of terms or even ideas (without changing the meaning).

The same kind of quality issues can occur for an answer, but there are various reasons why they are less often the target of peer edits. One is simply selfishness: the people reading the answers are likely people who are submitting competing answers. Another is that I would bet the number of folks answering questions who are also sloppy or post incorrect answers that are worthy of edits is quite small - much more likely that bad answers aren't edited by peers but rather down-voted and deleted.

I think in general it is easier to improve a question than to try to make someone else's answer more palatable or more correct. The range of use cases where you can do the latter without changing the answerer's intention is probably quite narrow.

This is not true for every single question and answer, of course, but I bet it captures a good portion of the trend you are seeing.

share|improve this answer

I'm sure that if you are able to find such metrics you'll see that there tend to be lots of questions from new users, whereas answers tend to come from people who have been around for a while. (People come to the site once or twice to ask a question and then leave; when someone starts answering questions they tend to stick around for a while, not just answer one or two and then go.) Really what you're more likely seeing is that posts from newer users tend to be edited more heavily than posts from people who have been around for a while (that this correlates to questions/answers is simple coincidence).

Newer users tend to be much less familiar with the site, they aren't familiar with the markdown/formatting options, etc.

share|improve this answer
7  
...and often we set out to edit just the formatting or just the grammar and we end up finding a lot more of both than we originally thought. :-) –  Aaron Bertrand Sep 17 '12 at 20:09

I can't speak for the community as a whole, but I can offer my perspective, as someone who's been around on StackExchange for a fair while.

I edit questions quite a lot, but only rarely do I edit answers. For me it's to do with ownership. I feel that questions are much more loosely tied to an individual than answers are. When you ask your question, there's an intrinsic appeal to the community. You're asking for help, and one of the ways we help is by fixing your question, requesting clarifications, and so on. Also, the question itself is the first point of contact with the post, and the thing that future readers will be seeing and responding to ("that's my problem!"). So a question is really something that is "given" to the site, as an offering. And I improve that when I can.

When I read an answer, I associate that response strongly with the answerer. They have provided the answer, their avatar is associated with the answer, and their reputation will stand or fall on the basis of its quality. By providing an answer, they are making a kind of implicit claim for their own expertise. I don't copy and write a similar answer, because that would be a breach of etiquette, and feels like stealing. I don't edit to modify or add information, because that would be putting words in that person's mouth, and they would reap the consequences (which could easily include downvotes). If an answerer particularly wanted this communal interference, they could use community wiki or a comment as a way to signal that. And I don't edit answers for grammar or spelling or style, because that feels too intimate. Just about the only thing I edit answers for is broken links. I think this is because they clearly need fixing, and absolutely nothing is added that wasn't originally intended.

In general, I notice the same behaviour from others on my own answers. The only time I really see people try to edit my answers is if they are very new (I usually reject the edit), or if it is a meta question like this one (when people do it to make a point or be humorous). Comments are the accepted convention, and much more respectful of the original answerer. I am protective of my answers, and I don't like having them edited directly by anyone else.

When I do feel an answer could do with significant editing, I usually say so in a comment, and downvote if necessary. I often provide the content in the comment, but leave it up to the answerer to add it to their answer if they see fit. If they do, I usually delete the comment and reverse my vote.

share|improve this answer

Beyond the reasons that others have stated here, I should point out that retagging a question counts as an edit. I know that I've made dozens of edits in a single day simply by cleaning up tags.

There are no tags on answers, so this isn't an operation you can perform on them. This would help to weigh edit numbers in favor of questions.

share|improve this answer
    
Kinda sounds like he's just diffing the text though. –  Shog9 Sep 18 '12 at 2:49
    
@Shog9 Right. A change in the tags only would result in a 1.0 on the chart, according to how I read the question. –  Gaffi Sep 18 '12 at 3:40
    
@Shog9 - Yeah, I wasn't sure if tags were part of the post history being diffed against here, but I thought I'd throw this out there in case they were. –  Brad Larson Sep 18 '12 at 4:21

Also consider a form of sample bias peculiar to this site: down-voted answers are more likely to be deleted, rather than edited. Once deleted, they become invisible except to moderators and higher reputation users (>10K).

Addendum: Questions proceed from ignorance; answers from knowledge.

Yes, I wrote that. A question inevitably reveals one's ignorance; a good question reveals a lot of ignorance. I suspect that people's reluctance to expose their ignorance contributes to many ambiguous, elliptical questions that require considerable refinement to answer.

share|improve this answer

You must log in to answer this question.

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