15

Currently, community reviews of posts flagged with "not an answer" or "very low quality" flags can only dispute those flags, not decline them. Disputed flags don't count against the flagger, they are as if a flag was never cast. Declined flags eventually cause warnings and timed flagging bans.

"not an answer" and "very low quality" flags now enter the community review queue on sites like Stack Overflow, and flags there even are hidden from moderators for a short duration. This has allowed for some people to flag indiscriminately, because bad flags don't count against them and good flags get added to their helpful flag count.

This is causing the community's time to be wasted, and endangering good answers that were poorly flagged and now run a risk of being deleted due to a bad review. Moderators would have been able to decline these flags, educating the flagger first about what they shouldn't be flagging, then banning them from flagging if they continued.

I propose that reviews which unanimously reject a flag (all "Looks OK" on a review for something flagged as not an answer, etc.) should decline that flag, not dispute it. Mixed reviews (some for deletion, some against) that don't lead to a deletion of a post should simply dispute flags as usual. This would allow for the community to slow or stop bad flaggers in the way that moderators could before.

Also, we've had a problem with people editing posts that have been flagged as "very low quality" or "not an answer", which automatically approves these flags and removes the flagged item from review. A number of non-answers and spam have made it through review this way. I'm thinking that edits should not clear flags automatically, but I could hear arguments about the specifics on that.

This is a feature request extracted from a portion of this larger feature request.

5
  • 4
    I wrote this query because of the (now-deleted) answer: Find Low Quality review tasks completed unanimously by Looks Ok actions within the past 7 days - I gotta say, even looking through some of the ones completed by only one action, the unanimously completed reviews definitely deserve to be declined. Wish we could limit that query to only show tasks that were created because of a user's flag and not a system flag, but I don't really see that changing the results we're seeing in review.
    – animuson StaffMod
    Jan 29, 2015 at 19:48
  • I think one of the worst consequences of letting bad flags slip through too much is that flaggers can't learn. They are deprived of finding out what is helpful and what is not
    – gnat
    Jan 29, 2015 at 20:00
  • 2
  • 1
    @JoshCaswell - Cases of disagreement in review would still lead to disputed, not declined flags. These reviews are a vote by the community whether something should be deleted or not. If the community unanimously rejects multiple flags calling for the deletion of types of answers, that would seem to provide a strong signal as to what the community thinks policy should be. Even if an ironclad set of rules was laid down today on Meta, very few reviewers would ever know about it. Jan 29, 2015 at 20:34
  • I agree that the policy should be formed by consensus, but why should that consensus not be explicit -- and why should reviewers not be expected to understand it? How can an individual flagger even know what to flag if the flag is subject to the arbitrary oversight of three users chosen out of a hat? "Why was my flag declined?" "Because Tim, Jim, and Kim each lost their keys." Is that really going to lead to better flagging?
    – jscs
    Jan 29, 2015 at 20:50

1 Answer 1

2

I totally agree. That way, the outcome would be more technically accurate.

  • Positive outcome: If the whole review is complete and the final result was to close/delete, the flag is marked as helpful.
  • Neutral outcome: If the overall outcome was leave open/looks OK but some users have voted to close/delete, the flag is marked as disputed.

  • Negative outcome: If the review was unanimously leave open/looks OK, or if a mod rejects the flag via the mod queue, the flag is marked as declined.

If the above outcome is implemented, it would also match the three-level (positive/neutral/negative) outcome behavior of the close flags and suggested edits.

For close flags:

  • Positive outcome: If review completes with a result of the question getting closed, all remaining flags are resolved as helpful.

  • Neutral outcome: If a single reviewer casts a close vote for the same reason as you, the flag is resolved as helpful, even if the overall outcome of the review was to leave open. The review does not need to be completed and the final result will not change the flag status.

  • Negative outcome: If review completes with an unanimous result of leaving the question open, all remaining flags are resolved as declined.

For suggested edits:

  • Positive outcome: If the overall outcome of the review was to approve/improve edit, the edit is approved.

  • Neutral outcome: If the pending edit was invalidated/conflicted by a subsequent edit, or if it was invalidated due to a post lock, it is marked as rejected - however, these rejection reasons do NOT count towards the ban and/or trigger a warning. There was a feature request that such situations should mark the edit as simply dismissed instead of rejected.

  • Negative outcome: If the final result of the review was to reject/reject and edit, the edit is rejected. This will trigger a warning and will count towards the ban.

1
  • 2
    IMO, no need for the previous revision. (adding "I'll eventually edit this answer for more details") Apr 19, 2023 at 13:44

You must log in to answer this question.

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