Which accounts have more questions than answers?

Question

There has been some discussion about certain users that ask tons of questions, but otherwise do not actively participate in a productive manner. I do not know if this is in fact true, I just want some proof one way or the other. I would assume that at the very least these particular users have a habit of making enemies. I would like to see further discussion about this, which is the reason for question's existence.

---

• They usually have less than half as many answers as questions, or possibly none at all.

• They will often have given an uncharacteristically high number of down-votes compared to how many up-votes they have given. One in particular has 4 down-votes with only 2 up-votes.

• If they have participated at all on Meta Stack Overflow, it has been minimal.

• They probably have had their accounts locked at some point.

  ---

  That is not to say that there aren't accounts that might appear to be similar, but that aren't actively anti-productive.

I would like to have more information about the accounts that follow this trend, whether or not they are productive members of Stack Overflow.

I don't really need to know who they are, just some of the specific details:

---

• up/down votes given
• up/down votes received on questions
• up/down votes received on answers
  
  
• Ratio of good questions to bad
• Ratio of good answers to bad
• Ratio of questions to answers
  
  
• Whether they have been put in the penalty box, and if possible how many times, or even the reason they were put there.
• Their general reputation range ( 0-50, 100, 500, 1k, 2k, 3k, 4k, 5k, 10k, ... )

• Whether they are active on any of the other sites

• Whether you think they are useful members, or not, based on what you've seen. This will have to be a judgment call.

• If you have any further information, go ahead and add that as well.

If you have something added to this list, just post a comment, and someone will add it, if it is deemed useful.

---

Basically what I want to know is, to what extent there is such a problem, or is there even a problem.

As I said earlier I also want information on all accounts that generally fit this profile, no matter what.

I'm trying to coax this information out in the hopes that it will be useful, to reduce the problems that may occur from these users, without hurting other users that may have similar attributes.

Answer 1

If you really want to know, why don't you download the database dump and examine it. You'll probably find all you want to know and more.

For those without the means or inclination to do so, I'll share my findings. I wrote a query that would scan order users by their question:answer ratio. So as to not point fingers, I haven't included the user IDs, but seeing as this data is publicly available, it's not exactly a big secret who these people are. (note that this is the data from the last dump, which is now over a month old)

ratio   rep     #qs    #as
139.0   3556    278     2   
137.0   1041    137     1
111.0   566     111     1
78.0    491     78      1
60.66   886     182     3
36.0    1106    36      1
35.0    237     35      1
35.0    830     35      1
34.0    461     34      1
30.0    340     30      1
29.0    1102    29      1
29.0    293     29      1
28.0    242     28      1
27.0    426     27      1
26.0    118     26      1

If someone has no answers, their question:answer ratio will be infinite, so I had to query them separately:

#qs   rep
134   1169
98    678
66    292
59    509
57    874
49    958
47    558
44    463
40    461
39    548
39    193
34    173
30    79
30    141
29    1

Looking at this data, I think that this 'problem' is way overblown. The maximum rep that appears here is 3556, which is not that high (all things considered).

Answer 2

Even if they never post a single answer, somebody asking (reasonable, original) questions is hugely productive to the site... otherwise we'd be sat their looking silly while we chew the cud here on meta.

Re "locked" accounts; actually, many of the (few) people who have been suspended at some point are regular, usually-constructive users. In the case of total trolling/spamming, there are more... "immediate" fixes than suspension.

I don't consider involvement on "meta" a critical measure of their benefit on SO. I fully expect the number of regular meta users to be a tiny fraction of the SO users; that doesn't in any way mean that the people who just stay on SO are less important. For example, I wouldn't expect to see Eric Lippert on meta (although he does have a parked account), but on SO he is a huge asset to the C# fraternity (i.e. pretty much by definition his answers on C# are correct).

Answer 3

Given the statistics you request, what would you do? I don't think anything could be done to change or improve things.. There's always going to be people that participate in a site less than others...

The more important fact is lots of people are hugely active in positive ways - I think that far outweighs the list of people you're looking for!

Answer 4

If you really want to know, why don't you download the database dump and examine it. You'll probably find all you want to know and more.

Either way, I don't understand why you care. I believe that the bad feelings caused by all this petty accounting outweigh any benefit there might be to finding all the "freeloaders" who only ask questions.

I'd like to remind you that those "freeloaders" view ads, as do the people who view their questions. If they happen to click through they're doing their part by helping to fund the site. Isn't that enough?

Answer 5

Excellent post. I think there are many metrics that would be useful to show on a user's profile page, such as:

• ratio of questions/answers
• upvotes/downvotes
• number of total upvotes and downvotes received from questions and answers
• average score of all questions and answers
• number of questions closed
• number of questions deleted

Perhaps all of these stats could be combined together into a magical meatgrinder algorithm to provide an overall ranking of the "quality" of the user. It would be interesting to see how this value graphed against total reputation.

EDIT: I'm also coming around to the earlier suggestion that upvotes on questions be worth only half of their worth on answers.

Answer 6

Why does it matter? If somone is asking lots of questions the good questions get up voted the bad ones get down voted. If it's a bad question they wont get any rep, unless it's about your favourite webcomic/computer joke, but I think we are over that now.

It's not like we as users are footing the bill for all the Guff questions in the database, they are there taking up a miniscule number of megabytes on a server, while all the usefull questions are there to help us do what we need to do.

Some people ask a lot of questions, some people answer a lot of questions, some people vote, some don't. So long as at least some people do all of the above I'm happy.

The number of people answering questions is always going to be the minority, because most questions require domain knowledge, where asking a question implies a lack of domain knowledge. Given the number of domains there are in programming and how often programmers have to tackle different problems it can be assumed that people are going to get stumped more than they can answer questions.

TBH I think that stack overflow could possibly survive if only John Skeet was answering questions full time. The man is a machine.

Although arguably I think many people seem to turn to SO before even doing a simple Google search on their problem, this might be because forming a human readable question is easier than formulating a google readable query that will come up with the results you want.