19
\$\begingroup\$

These days I come to the site in the morning and I usually see 10, 15, even 25 posts in the mod-queue with flags waiting to be handled. When I started (two+ years ago) it was usually more like 2 or 3 posts.

I wondered aloud how much of an uptick there'd been in flagging recently, and SE staff were kind enough to rip off a quick visualization:

timeline of flags thrown on RPG.SE [click for larger image]

That's count of flags per month on the vertical axis, month on the horizontal, and a legend of the stacked lines below. Even without clicking in you can see that flags are up quite a bit recently.

The huge spike toward the end is because of a spam wave we had at the time.

I don't think there's anything wrong here. But this uptick in flagging does lead to a commensurate increase in moderation activity, and most of that activity occurs in a behind-the-scenes way that you-all can't really see or know about. (Except inasmuch as it affects your posts or flags.) So I wanted to get it out there publicly.

Some unordered observations of my own:

  • First, note that the number which caught my attention is actually posts with active flags, not just flags. So while commenters are almost certainly right that some increase in absolute number of flags (as graphed above) would be due to people flagging whole threads, there's also just more posts receiving flags.
  • Off the top of my head I'd say ~90% of the flags I see are valid/helpful. So this is (in my opinion) good work the community's doing.
  • We know the userbase is growing, basically any way you measure it. Some of this increase in flagging is to be expected, then.
  • This increase, though, does seem to outpace userbase growth. We've something like tripled our flag-rate in the last year or two, whereas users and posts are up on the order of 10% year over year.
  • That thick reddish bar that's ~half of the flags from 2013-mid2017, then disappears... remember they renamed some of the comment flagging options? Those red ones are "obsolete"; the teal bar at the very top these days is "no longer needed." I.e. they're the same flag: comments that can be cleaned up having served their purpose.
  • The flag I most often decline is Not An Answer, generally because someone's flagging a poor answer as not an answer. My understanding (derived from this meta.SE post and from my early declined NAA flags being explained to me by mxy/SSD) is that NAA is used for posts that don't attempt to address the question: gibberish, new questions, riffs on a theme raised elsewhere, &c.
  • The other flag I decline a lot is when some comments are flagged No Longer Needed (NLN) but they've only been around for ten or fifteen minutes. Sometimes even those are obviously trashable. But when there's good commentary that appears to be helping posters and readers I'm loath to trash it too quickly. After a few days of inactivity, definitely. After a few hours, often not.
\$\endgroup\$
  • 5
    \$\begingroup\$ I wonder if you'll see fewer flags now that @Rubiksmoose and I are mods and can handle stuff ourselves... :P \$\endgroup\$ – V2Blast Jun 13 at 3:15
  • 3
    \$\begingroup\$ I wonder if your post on flagging comments has anything to do with it recently? It certainly increased my flagging. \$\endgroup\$ – linksassin Jun 13 at 3:36
  • 2
    \$\begingroup\$ @linksassin Mine too, normally i would have only flagged one, now i flag whole threads if they are NLN, solely because of that meta answer. \$\endgroup\$ – Liam Morris Jun 13 at 9:55
  • 5
    \$\begingroup\$ @linksassin and LiamMorris--that's a good point and, I think, probably cause for some of the uptick. It's also made me realize that I left out a subtlety in the moderator dashboard which probably means that can't be all of it. Editing now. \$\endgroup\$ – nitsua60 Jun 13 at 11:49
  • \$\begingroup\$ I'm sure nitsua60 will @V2Blast, but it's more because now you folks are busy modding you'll have less time to generate stuff we keep needing to flag. :D ;) \$\endgroup\$ – RyanfaeScotland Jun 13 at 11:59
  • 3
    \$\begingroup\$ Here's a thought: the increased flagging activity is related to tomorrow, 14 June, being Flag Day in the USofA. 8^D \$\endgroup\$ – KorvinStarmast Jun 13 at 14:24
  • \$\begingroup\$ I wish this number used to be higher, then I wouldn't stare in dread at so many old posts that I read. If I wasn't actively requested to not improve older posts by flagging them, I would use all of my flags - on each of my active days. \$\endgroup\$ – Akixkisu Jun 13 at 15:04
  • 1
    \$\begingroup\$ The post on flagging comments made me flag a lot more recently, and the majority of these flags, of course, are targeted on comments. \$\endgroup\$ – Kuerten Jun 13 at 22:17
  • \$\begingroup\$ Just wanted to say thanks for all the hard work! \$\endgroup\$ – C. Ross Jun 15 at 5:00
  • 1
    \$\begingroup\$ @C.Ross and thank you! We wouldn't have this great resource and community available to us if you hadn't done so much ground work! \$\endgroup\$ – nitsua60 Jun 15 at 21:20
  • \$\begingroup\$ I'd like to point out that because new users who don't know how the site works are likely the major source of flagged posts, the relevant metric is not necessarily userbase growth rate as a percentage of our total userbase but instead the growth rate as an absolute value. If we used to be getting 100 new users per year and now we're getting 300, then it would make sense for our number of flagged posts to triple. In particular, as total userbase grows, a given absolute growth value will translate to lower and lower percentages of the total, making the percentage metric less useful. \$\endgroup\$ – Oblivious Sage Jun 16 at 22:54
  • \$\begingroup\$ Having said that, a jump in the percentage growth rate would indicate an especially large jump in the absolute numbers, which would obviously mean a lot more flagged posts. How does that 10% compare to our historical growth rate (assuming that data is available)? \$\endgroup\$ – Oblivious Sage Jun 16 at 22:55
3
\$\begingroup\$

I would say there are many reasons but the most likely is purely a numbers game. This looks like non-linear growth with step changes.

The non-linear nature I would put down to number of users reviewing * number of questions * number answers, all of which should be steadily increasing on a healthy SE site.

I would put the steps down to singular events, such as: a user who has a passion for flagging joining the site, or a post on flagging highlighting it to more users.

I agree with your main point there does not seem to be anything wrong here. I am also going to ask the very helpful SE staff if they can run me one of these off for Homebrew forum; you have piqued my interest.

\$\endgroup\$
  • 1
    \$\begingroup\$ While the active user group grows slower, the kind of users that are not filtered out by the QA format and heavily enforced guidelines may be more inclined to moderate themselves. This environment actively filters and selects for this user type. \$\endgroup\$ – Akixkisu Jun 14 at 5:21

You must log in to answer this question.

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