How do I find the user that has been downvoted the most (both in terms of answers or questions) on a Stack Exchange website?


We can use the following script


In the case of Stack Overflow, we can see that "god" Jon Skeet has been downvoted a lot, even though HE is a "god".

We can also choose another SE community. For example, in the case of meta, "god" Jeff Atwood has been downvoted the most (https://data.stackexchange.com/meta.stackexchange/csv/70204), even though he is also a "god" (and he's unemployed now at SE, so he's technically no more a god).


This possibly proves that downvotes are often used as a revenge (or, in general, they are cast on a whim), but I need to test this hypothesis using a statistical test.

  • 1
    At best it proves that sometimes "god" gets it wrong as well ... – rene Jun 11 '19 at 15:48
  • @rene Skeet has an answer with more than 10000 upvotes (this answer surely doesn't relatively deserve so many upvotes) and about 30 downvotes. Why do you think this is the case? 30 genius against 10000 undergifted ones? – nbro Jun 11 '19 at 15:50
  • 1
    30 users found the answer not useful, for whatever reason they personally dreamed up and voted accordingly. "God" can't please everyone. – rene Jun 11 '19 at 15:52
  • @rene God indeed does not please everyone and the downvotes have indeed been used as a revenge. Thanks for confirming my hypothesis. I still need to reject the null hypothesis though using a statiscal test :) – nbro Jun 11 '19 at 15:53
  • 2
    You are the one that is on the revenge hypothesis. I've never used that word. Please stop demonizing me as down voter. It gets a bit tiresome . – rene Jun 11 '19 at 15:56
  • @rene Yes, I am the one. Just ignore this comment and it will not be tiresome anymore. – nbro Jun 11 '19 at 15:57
  • 1
    Here is a network-wide query to further prove a point. – rene Jun 11 '19 at 16:23
  • @rene Which point does it prove? – nbro Jun 11 '19 at 16:23
  • I have no idea. You are collecting stats per site, I happen to know how to do that across all sites in the network (I excluded per site meta's). I assumed you might find that helpful. If not then I still had an awesome 30 minutes. – rene Jun 11 '19 at 16:29
  • @rene Ok, thanks. I will maybe later use your query. – nbro Jun 11 '19 at 16:30
  • 1
    -1 I was actually going to upvote this as an objective answer to the question until I read the last paragraph—which is both impossible to prove and baseless. – Jason Bassford Supports Monica Jun 11 '19 at 18:41
  • @JasonBassford Thanks for your feedback. I don't want to rigorously prove it (and this is not usually the case in statistics), and indeed it is impossible to prove such as thing. I will just perform some statistical tests. – nbro Jun 11 '19 at 18:43
  • 1
    You've yet to prove your hypothesis, nbro. You've failed to consider other alternative hypotheses, e.g., please go back to the query and view the column pctDownVotes (percent of downvotes wrt number of posts) and you'll see that, e.g., Jon Skeet has answered so many questions, that when comparing the number of downvotes he's received, divided by the number of his posts, is less than $0.005$, or less than 0.5 percent. Lesson here, you need to think beyond mere raw numbers. – amWhy Jun 11 '19 at 21:41
  • @Namaste Thanks for pointing this out! – nbro Jun 11 '19 at 22:11
  • On technical site such as SO, a good answer can still be "not useful" to the reader when it explains how/why a problem is almost impossible to be solved, but providing no workaround... because, well, they only want the solution to their problem! (no revenge voting at all, just a disappointment on the answer). On Meta SE, SE employees tend to post announcement or reasoning that the community just didn't like it. (again, no revenge downvoting, just a disappointment on the post) – Meta Andrew T. Jun 12 '19 at 2:49

You must log in to answer this question.

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