What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 130 Stack Exchange communities.

Suggestion:

Provide specific flagging examples to users new to /review:

  • Concrete (anonymous?) examples of recent valid flagging by other users for non spam/offensive flags at least. Probably just another tab on /review.
  • "Where are they now?" - longitudinal tracking of specific posts to see if other users flagged them. Basically a mechanism to say "I'm interested in learning what happens with this one", but without having to take a side in cases where you're not sure. Seeing how these borderline cases are handled will help users figure out whether or not to flag similar cases in the future. The implementation of this could be handled as a neutral flag, which causes it to be added to the flag weight summary page. It could also be used as a hint that the post ought to be weighted in order to be shown to more potential flaggers.

Rationale:

The feedback from flagging has improved a lot recently, but there still seems to be one aspect of it could be improved. Currently when using review there are four basic outcomes:

  1. Bad post, correctly flagged
  2. Good post, erroneously flagged (false positive)
  3. Good post, ignored or upvoted
  4. Bad post, ignored (false negative)

The current system seems to do well for all cases except for case 4. The current feedback on flagging encourages false negatives, particularly through the /review route (i.e. problem posts which don't get flagged.)

For case 1 you see what action the moderator took and get an increase in flag weight. Likewise for case 2, but with a decrease in weight.

Case 3 is the default, do-nothing behavior, the answer will probably get voted on by users and possibly accepted by the person who asked the question.

It seems though that the system as it stands positively encourages case 4 by penalizing flagging errors, which although reasonable to discourage excessive poor flagging encourages an "if in doubt don't flag" mentality.

share|improve this question
1  
My head is spinning. –  Robert Harvey Aug 23 '11 at 22:27
    
@Robert - I'm guessing that's not a good thing? –  Flexo Aug 23 '11 at 22:28
    
Simplicity is everything. I know my attention span is short, but still... –  Robert Harvey Aug 23 '11 at 22:29
3  
I fully agree on the problem analysis but I have my doubts that the listed suggestions will be of any use. Frankly, I don't see how they would be. –  Wladimir Palant Aug 24 '11 at 11:20
    
The other view I guess is to say that there are tools to do it right now, they're just not very will integrated. You can leave a comment on a possible problem post suggesting what's wrong and how to fix it. The problem is other users won't see it in /review and you won't know even if they did see it. Perhaps increasing the visibility and tracking on this with a direct "leave a comment" option on /review that leaves comments not only on the post in question, but also on the question in /review and is visible later, with flags from other users would help? (Perhaps that should have been an edit?) –  Flexo Aug 24 '11 at 13:20
2  
@bemace - thanks for making that much clearer. –  Flexo Aug 24 '11 at 15:19
    
I don't understand. Why is the current system using flag weight broken? People who consistently flag correctly are automatically given higher priority in the queue. Those who don't are not. People can learn by doing, which is far more effective than sitting here and pretending that they'll learn by example. When their flag weight goes down, that was a bad flag. When their flag weight goes up, that was a good flag. Rinse, lather, repeat. –  Cody Gray Aug 24 '11 at 17:02
2  
@Cody - that punishment for bad flags encourages people to get stuck in local optima and never try and escape that to search for the global optima. "Once bitten, twice shy". I was quite surprised when I learned yesterday that my flagging had been far too conservative. I'm pretty certain I wouldn't have got to that point if I'd seen a few more examples, or followed a the process of some. There were loads of bad things I'd ignored simply out of fear. I would be surprised if I was the only one. –  Flexo Aug 24 '11 at 17:21

You must log in to answer this question.

Browse other questions tagged .