What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 129 Stack Exchange communities.

I find reviewing suggested edits a great way to kill a few spare minutes, but I'm still quite apprehensive about whether I'm being too lenient when approving, or too anal when rejecting edits.

I occasionally look through my "reviews" activity, but clicking through each item to see the final judgement is a real pain. Furthermore, simply looking through a few entries may not be representative of the greater good since all it takes is 2 other approvals/rejections to overrule my judgement (or the 1 other to push it through).

It would therefore be nice to know my cumulative hit rate. Perhaps something along the lines of:

Table for edit stats

Clicking on the "40 overruled" would then filter down list of suggested edits to those that I should have rejected (or ignored) instead. It would then be down to me to determine which ones I'm doing wrong and which are just down to different opinions.


Alternatively, since the suggested edit page already display some reviewer stats ("So-and-so approved X edit suggestions, and rejected Y edit suggestions."), stats on hit rate could be inserted there to avoid making major changes to the UI.

Crude mockup:

Inserting hitrate within reviewer stats

share|improve this question

3 Answers 3

Whilst this may be an interesting stat, I'm not sure that it's really going to change anybody's behaviour. Ultimately, your perception of a good edit is a very personal thing and I don't think we really want to encourage people to always take the easy road and vote with the masses.

It's better that you reject an edit because you think there is something wrong with it, than accept it because you believe somebody else would have accepted it and there's not enough wrong with it...

To me, the suggestion is a bit like keeping track of what percentage of answers you've voted on have eventually become an accepted answer. Whilst it may be interesting, I don't think it's really useful, and it could encourage behaviour that I don't think would be in the best interests of the site.

share|improve this answer
    
Fair point. My concern is that while one's idea of right and wrong is a personal thing, it should also be subject to the site policy and acceptable norms. I do a search on MSO when in doubt, but what about cases where I am unaware that my opinion is different from the consensus view? –  Shawn Chin Nov 8 '11 at 12:49
    
You're already in a position where you can make your own, unreviewed edits, any edits your review are peer reviewed by somebody else as well, do you have to always match the consensus view? What you're evaluating is whether or not you think an editor has shown sufficient effort to earn any rep gain + has contributed to a post in some way that is beneficial. This is always going to be subjective. If you want to improve your reviews you will make the effort to do so (as your are), by evaluating your decisions. If you don't, the stats are just going to be a number... –  forsvarir Nov 8 '11 at 13:00
1  
Indeed if one has no intention to evaluate past decisions, then the stats are just numbers. But if one does, wouldn't the stats help? We're already being shown a full list of past approvals/rejections and the final decision of each (on click-through), just not in a manner that's easily consumable. –  Shawn Chin Nov 8 '11 at 13:26
    
Furthermore, being in a position to make my own edits doesn't necessarily mean that I can correctly judge the editing behaviour of others. I can tell if they're blantantly wrong, but if it's simply an edit that I normally wouldn't do (say converting a book title to an amazon link) I'd link to know I'm acting in accordance with the site policy. –  Shawn Chin Nov 8 '11 at 13:29
    
Something on the Reviews tab that indicated if a decision was contested, without drilling down may be useful an allowing you to identify 'interesting' reviews. If the purpose is to analyse your own decision making, I think it's more useful to know if anybody disagreed with you than if two people did. As far as the site policy goes, each person develops heir own interpretation of the rules based on their experience (what they do/read) on the site. To me, there's little difference between applying that to my own posts/edits, votes, approvals or flags. If it looks wrong, it probably is... –  forsvarir Nov 8 '11 at 14:07
    
I agree that it's only the contested decisions that are useful/interesting, hence my emphasis on overruled decisions in the mockups. It takes 2 others to disagree, and if that happens often enough on similar types of edit, I'll know exactly where I've gone off track. –  Shawn Chin Nov 8 '11 at 14:26

I wrote a small JavaScript code that when put into the browser console and executed, will bring up the summary of your review decisions on the current site.

It crawls the reviews in your activity, starting from the first page and count how many of each review "type" e.g. Approve, Reject, Edit etc you got, then continue to the next page.

OK, so the code is:

var _reviewMapping = {};
function ParseSinglePage(page) {
    var currentURL = "/users/current?tab=activity&sort=reviews&page=" + page;
    console.log("Currently processing page " + page + "...");
    $.get(currentURL, function(response) {
        var wrappedResult = $(response);
        var arrCells = wrappedResult.find(".history-table .reviewed-action");
        if (arrCells.length > 0) {
            arrCells.each(function() {
                var review = $(this).text();
                if (!_reviewMapping[review])
                    _reviewMapping[review] = 0;
                _reviewMapping[review]++;
            });

            window.setTimeout(function() {
                ParseSinglePage(page + 1);
            }, 1000);
        } else {
            //finished
            for (var review in _reviewMapping)
                console.log(review + " = " + _reviewMapping[review]);
        }
    });
}
ParseSinglePage(1);

Sample result: (my own current stats on Meta)

Approve = 258 
Leave Open = 30 
Edit = 20 
Close = 41 
Reject = 108 
Leave Closed = 15 
Reopen = 9 
Edit and Reopen = 2

The result will appear in the console as well when it's done. It samples one page every second, so even with lots of reviews this should not take more than a minute.

Tested on Chrome only at the moment, if someone want to make it compatible with more browsers, I'll be glad.

Enjoy your stats! :)

share|improve this answer

You may be interested in these SEDE queries:

Take this with a mountain of salt. There are a lot of bad decisions about suggested edits; going with the majority doesn't always make you right. Concentrate on approving the good ones, rejecting the bad ones, and improving the good-but-incomplete ones.

share|improve this answer
    
That's neat! Thanks :) –  Shawn Chin Sep 13 '12 at 8:20

You must log in to answer this question.

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