Let's wait just a bit
I left a comment with some data in it yesterday, and I'd like to expand on that a bit. The text of the comment in case it gets deleted at some point:
My first inclination was to say that we rely too much on the mods to delete things for us, so I went looking for data to back that up. What I found is that since the start of October, we have had 16 question hit HNQ. On those questions, we've had a total of 119 answers, with nearly 40% (47 total answers) deleted. Of those 47 deleted answers, just 36% (17 answers) were deleted by someone with a diamond. Note that I did not count answers where one of the mods cast the 3rd delete vote as being deleted by a mod.
In my eyes, the most important factor in choosing to add more question to HNQ is how well equipped the community is to moderate those questions and answers. We have some excellent moderators here, but the bulk of the moderation burden needs to fall on people without diamonds.
As of today (Nov 20th) we have been back on HNQ for exactly 6 months. As a gauge of how the community is doing at moderating, I took a look at some deletion stats for our first month-ish (5-6 weeks and 16 questions in total) back on HNQ and our most recent month-ish (again 5-6 weeks and 16 questions in total). So here are the stats
- 16 HNQ questions
- 104 total answers on HNQ questions
- 43 answers deleted on HNQ questions
- 5 answers on HNQ questions deleted by someone with a diamond
- 16 HNQ questions
- 119 total answers on HNQ questions
- 47 answers deleted on HNQ questions
- 17 answers on HNQ questions deleted by someone with a diamond
Note: deleted by diamond does not include cases where a moderator cast a delete vote in a situation where a single additional delete vote from a non-moderator would have also deleted the answer
Numbers Are Fun, What Do They Mean?
Great, so what does this mean? The first thing that jumps out in this statistic is the fact that in the first month we deleted ~40% of the posted answers, but only 11% of all the answers deleted were deleted by mods, but in the most recent month, while still deleting ~40% of the answers, 36% of those answers deleted were done by mods.
There are a few reasons why a mod might unilateraly delete an answer
- It is obviously spam or rude (a single spam/rude flag from a mod will delete the answer)
- The mod has gone rogue (spoiler alert: none of our mods have gone rogue)
- The answer has a lot of upvotes, so it can't be deleted by the community
- The mod cast a delete vote because the answer didn't meet our quality standards
I can say based on several conversations I've had with mods in chat that they are all quite judicious with their votes. The 4th scenario is unlikely to happen because they try to avoid casting a delete vote unless the answer is upvoted and they've received a sufficient number of flags pointing out the quality issues it has. Spam and rude posts have no place on this site, and mods nuking them on sight is a good thing. I already mentioned that our mods have not gone rogue, which just leaves the third bullet point.
In my opinion, the amount of posts that get lots of upvotes and then subsequently must be deleted by mods is the key metric. When I gathered my data, I didn't count up what kinds of posts where being deleted (spam, rude, lots of upvotes but no backup, etc...), so I can't say for sure how much of an uptick there was in this metric from the first month to the last month, but I don't think we've deleted that many more spam/rude posts over the last month than we did the first month. That means that we likely have more answers collecting upvotes and then requiring a diamond to be deleted. That (to me) indicates that we're not doing as good of a job as a community of staying on top of HNQ questions in order to ensure that answers meet our standards.
One last point and a recommendation
Most of the answers that we delete on HNQ questions are deleted because they lack backup. There's been a some discussion on Meta recently about how we apply that policy to older posts. This has led to the moderators deciding that we should try some new strategies for applying this policy. I don't know that a decision has been reached yet as to what those changes will be. Given that we appear to be slipping in our ability to moderate HNQ and there are discussions in the works to improve how we handle answers without backup, I think that we should hold off on increasing our number of concurrent HNQ questions for the time being.