Timeline for Do we use the downvote button more than other sites?
Current License: CC BY-SA 4.0
16 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 31, 2021 at 17:37 | comment | added | Glorfindel | The problem is that the number of views is only known for questions and it includes anonymous users and users without voting privileges. | |
| Jan 31, 2021 at 17:33 | comment | added | YCor | I agree that downvotes/post is of little relevance to compare with other sites, and should be divided with the number of views by people who have the right to vote (dividing by the number of views would be a first good approximation). | |
| Jan 24, 2021 at 12:42 | comment | added | kjetil b halvorsen | In addition to downvotes per post, downvotes per view could be interesting, and MO have rather high view numbers (at least compared to Cross Violated, which I frequent more ...) | |
| Jan 19, 2021 at 16:24 | comment | added | Martin Sleziak | @MatthewDaws Perhaps queries looking only at posts which were not deleted would - at least to some extent - solve the problem that this includes some posts which definitely deserved to be downvoted (such as spam and off-topic posts). I have added one such query into my answer - probably there are many other meaningful ways how to restrict the questions from which we're taken the average. | |
| Jan 19, 2021 at 15:37 | comment | added | მამუკა ჯიბლაძე | Despite all this questionability, it is very interesting. Seems like in terms of (downvotes per post)/(upvotes per post), the least friendly are sites like Islam, Android, SO.Spanish, Health, SE.Meta, Christianity, Biblical Hermeneutics, ... while the most friendly ones are ... SE.Korean, Mathematica, Latin, Emacs, and TeX | |
| Jan 19, 2021 at 15:27 | comment | added | Matthew Daws | Finally, I suppose the problem is that we don't have the data analysis tools to really be able to say any more...? | |
| Jan 19, 2021 at 15:26 | comment | added | Matthew Daws | An alternative explanantion is of course that a small number of posts have a large number of close votes. We do get spam (blantant advertising or otherwise nonsense) and some "please do my Calculus 101 homework for me" questions, which I guess are downvoted, closed and deleted in short order. Would these affect the average? From the perspective of being "unwelcoming", I am not in the least concerned about this class of question. | |
| Jan 19, 2021 at 15:24 | comment | added | Matthew Daws | An unfortunate problem with average (=mean) downvote counts is that it's hard to know what these mean. With about 0.5 downvotes a post, this could mean that roughly 1/2 of all posts have one downvote. I know that I sample questions in a very biased way (I tend only to read question with certain tags which I follow) but this seems way off the mark: it's very rare for me to see any question with a downvote. | |
| Jan 19, 2021 at 13:35 | comment | added | Martin Sleziak | I will add that the table Votes still contains the date, but not the exact time. But since we cannot look at them on a finer scale than days, this is unlikely to give some interesting data, since the most votes come on the first few days: data.stackexchange.com/mathoverflow/query/1236796/… data.stackexchange.com/mathoverflow/query/1236798/… | |
| Jan 19, 2021 at 12:56 | vote | accept | Alec Rhea | ||
| Jan 19, 2021 at 12:18 | comment | added | Glorfindel | It turns out I made a mistake in the query and now managed to create a correct one, see the updated answer. | |
| Jan 19, 2021 at 12:18 | history | edited | Glorfindel | CC BY-SA 4.0 |
deleted 63 characters in body
|
| Jan 19, 2021 at 12:01 | comment | added | Alec Rhea | Ah, I just saw that you attempted this and it timed out, my bad for not reading and thank you for trying. | |
| Jan 19, 2021 at 11:41 | comment | added | Alec Rhea | Running a few quick queries, it appears that one downvote every 5-4 posts is fairly standard, while MO is rocking double that at almost one downvote every other post. Is there any way to view all sites in one window? | |
| Jan 19, 2021 at 11:34 | comment | added | Alec Rhea | Thank you Glorfindel; I agree that anonymity in voting is important, but I don’t think (say) hour-by-hour timestamps on votes would be enough to reconstruct anything personal, and they could still be edifying to see how many posts receive ‘first hour’ downvotes here versus other sites. I think this may be an important metric in understanding our perceived unfriendliness. | |
| Jan 19, 2021 at 11:26 | history | answered | Glorfindel | CC BY-SA 4.0 |