42

So I've recently learned my lesson about "No Action Needed" in the First Posts and Late answers queue. This has been discussed to death:

Obviously, it's important to have the No Action Button in the first place. But it's been suggested before in an answer to a buried, low view question that we should have more than one person needed to click "No Action Needed" to clear it from the queue

We know that lots of people are going to click "No Action Needed" when they should do more, it's certainly better to have it than upvote it, but if it clears it from the queue, then no one can review it properly, because it's gone. Most of the time when I review First Posts or Late Answers, I spend so much time editing the post that by the time I get back to the queue, all the other items in it are cleared, most of them by "No Action Needed".

Here's 10 most recent "No Action Needed" on Stack Overflow in First Posts:

Now that I've been educated (through a much-deserved review ban), it's super extra clear to me that "No Action Needed" is very often wrong. So, let's require more than one of these to clear the queue! This won't solve the problem, but it should help it somewhat.

Update: This "No Action Needed" was so egregious I had to edit it in. Yes, sample size of one, I know, but still: https://stackoverflow.com/review/late-answers/9460408

What exactly went wrong with your changing of the contact.php? Can you find and paste the part of the source code which is showing the Google Maps, or just paste all of the file here so that we may be able to help you better?

If you want to create a Wordpress website from scratch - you may want to follow our awesome checklist: How to create a complete Wordpress website for beginners


With the help of @Ixrec, here are the last 5 No Action Needed on Programmers.SE for First Posts:

My conclusion is that I think this is much less of a problem on smaller graduated sites, and I'm on the fence about whether this change should be rolled out to all sites or just the largest ones.


Other relevant links:

  • 7
  • 1
    @animuson Ahhh I searched but I didn't find that one. I think this is just marginally different enough to not be a dupe - because I don't want more reviews for all cases, just no action needed. Since you're an employee, are you able to share if this is still under discussion? – durron597 Sep 8 '15 at 18:22
  • 2
  • Note: I'm a careful reviewer, addicted to editing, commenting, and flagging, and not unwilling to vote, but I hit NAN about 20% of the time on SO (43 out of my last 200 FP reviews), or 28% on ELL (57 out of 200). "Almost always wrong" is an overstatement. – Nathan Tuggy Sep 8 '15 at 19:33
  • @NathanTuggy The actual frequency of NAN in the history is much higher than 20%. – durron597 Sep 8 '15 at 19:34
  • @durron597: So say that it's "usually" wrong, or that it is "grossly prone to overuse", or that "3/4 of the time you hit it you shouldn't". Hyperbole should be handled carefully if you're not certain your audience is uniform in composition. – Nathan Tuggy Sep 8 '15 at 19:35
  • 1
    @NathanTuggy I've edited. I still think 20% is higher than what it should be, in my experience (I have the Silver badges only for those queues on SO) but you're right "almost always" is too strong. – durron597 Sep 8 '15 at 19:48
  • 2
    @NathanTuggy there are some reviewers who have an anomalously high percentages of no action required reviews. And while those who are in the 20% range may be right, there are more than a few reviewers on sites who are in the 80+% range of no action needed (or IMHO, worse - "Welcome to Some.StackExchange." as the entirety of the review, but that's out of scope of this feature request). – user213963 Sep 8 '15 at 19:56
  • Related: Review completed by one person in First Posts queue at SU. – kenorb Sep 21 '15 at 10:08
  • +1: It's an ongoing concern for me. I'm not exactly comfortable with being able to unilaterally declare that a post needs no action. A lot of new user posts need to be checked by another set of eyes. – bwDraco Sep 21 '15 at 21:13
22

There's a very important reason to require more than one "No Action Needed" review to clear items from the First Posts or Late Answers review queue: spam.

Right now, all it takes is a single bad "No Action Needed" review to approve spam in First Posts or Late Answers. It then is allowed to live on the site until someone happens to come across it and takes the time to flag it. I've seen spam live for months like this due to a single careless review.

New reviewers often try to game the review badges by spamming "No Action Needed" clicks for as long as they can until they hit an audit, then they either stop or tend to review more carefully going forward. In the time before they hit those audits, they can cause a lot of damage by allowing spam posts or non-answers to slip through. Even more careful reviewers can sometimes make mistakes on these posts.

It has been argued in the past that adding this extra review requirement would slow down the review queues, but these queues are handled very quickly even on a large site like Stack Overflow. It has also been argued that voting may potentially be skewed on these posts because they are artificially exposed to more people. I haven't seen that in practice from the First Posts or Late Answer reviews, but maybe there's something I've missed.

  • I do know the queues were changed because the voting was considered to be out of control. However, since a single vote would take the review out, multiple opportunities to vote are a lot less of a problem; the probability of getting a single vote goes up somewhat, but that's all. – Nathan Tuggy Sep 9 '15 at 17:19
  • See latest edit which includes this review – durron597 Sep 10 '15 at 13:45
  • @durron597 - I've handled that answer, but I believe that was an honest mistake by the reviewer. They had tracked down a number of other non-answers and spammy posts before that, so they were reviewing well outside of this. It does show that even good reviewers can make mistakes and allow bad posts to live, though. – Brad Larson Sep 10 '15 at 14:28
  • 1
    @BradLarson I agree that it was an honest mistake by the reviewer, but if I hadn't been going through /tools/new-answers-to-old-questions it never would have been found. Anyway, my point was not to punish the reviewer but to emphasize the need for more eyeballs on these posts; it's easy to imagine one person making that mistake but not three. – durron597 Sep 10 '15 at 14:32
4

There are a significant minority of users who take either a positive action or no action on an unusually high number of reviews.

Having yet to dig into the data.se to come up with firm numbers, glancing at the first post history on Programmers.SE to back up my claim I see users who have a breakdown of:

reviewed | no action needed
  22     | 22
   3     | 33
  12     | 0
   6     | 0
   1     | 0
   5     | 4
   2     | 0
   2     | 0
  49     | 6
   3     | 0
   0     | 2
   0     | 1
   5     | 2

Unfortunately, I can't separate out the 'up vote' reviewed from the 'down vote' reviewed or the 'comment' reviewed easily. I have verified however that for the individual with 33 no action needed reviews, at least one of those 3 review reviews was an up vote.

It's not the 20%, or even the 50% of no action required. But when someone is reviewing at 91% no action required there is a significant chance that new users are missing out on guidance that the review can provide, letting spam get through (less of an issue on small sites because the front page is good at killing spam too), or helping set incorrect expectations on the part of the new user.

Additional eyeballs are needed on these no action needed reviews.

I would go further to suggest that instead the review should be two of the same outcome (positive, negative, or neutral) to complete the review.

A positive review:

  • up voting

A negative review:

  • down voting
  • flagging
  • vote to close
  • vote to delete (IIRC, a negatively scored answer can be VtD from the queue)

A neutral review

  • no positive or negative review
  • a positive and negative review (up voting and flagging)
  • comment without a vote
  • voting on a comment
  • no action needed
  • edit (this is so that the next one of the neutral actions completes the review)

Consider that on all the sites other than Stack Overflow, FP and LA review queues are the only thing that comes close to what the Triage queue can do on Stack Overflow. Furthermore, consider what the Triage queue would be like if one review completed it.

  • I appreciate the sentiment, and I've upvoted, but I'm not sure I entirely agree. The vast majority of no action needed reviews, the proper action is edit, which you don't discuss here. I want more eyeballs to increase the chance of these posts being edited. – durron597 Sep 10 '15 at 1:26
  • @durron597 Indeed, and I've added it in as a neutral review (in theory - another neutral review after an edit is confirmation of it). – user213963 Sep 10 '15 at 1:28
  • Also, what about when a user takes more than one action? – durron597 Sep 10 '15 at 1:31
  • @durron597 if they are all of the same category, then it is of that category. If they are of different categories, then it is a neutral review. – user213963 Sep 10 '15 at 1:32
4

Another reason why it makes sense to get a consensus for these queues is that they are the first review queues made available to users with 500 reputation. (I assume this can vary, but that is what I've seen.)

  • 1
    Either 500 rep or 350 rep, yes. Triage comes online at the same time on SO, but nothing that I've ever seen comes earlier, and in fact the privilege is named something like "Gain access to review queues". – Nathan Tuggy Sep 11 '15 at 7:39

You must log in to answer this question.

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