It seems as of late that the overall amount of accepted questions has dramatically decreased.
It may be the tags I hang around in but am unsure.
I guess I'm curious if it seems that way to anyone else.
1
-
Hard to tell after they removed that metric from everyone's profile. But this would be neat to know.– BlackVegetableCommented Apr 27, 2013 at 18:54
Add a comment
|
2 Answers
The accept rate was removed on the 22nd of January 2013. To steal my query from this question the accept rate on a quarter-by-quarter basis has definitely been dropping:
Year Quarter Accepted Not Accepted Accept Rate ---- ------- -------- ------------ ----------- 2008 3 14235 3636 79.65% 2008 4 29138 10313 73.86% 2009 1 40448 13526 74.94% 2009 2 54915 20733 72.59% 2009 3 71176 27276 72.3% 2009 4 82231 30543 72.92% 2010 1 103838 37853 73.28% 2010 2 112158 44554 71.57% 2010 3 130723 52494 71.35% 2010 4 142305 58071 71.02% 2011 1 182440 76891 70.35% 2011 2 199785 86545 69.77% 2011 3 206725 90639 69.52% 2011 4 209429 89543 70.05% 2012 1 244217 110679 68.81% 2012 2 249610 120293 67.48% 2012 3 250879 129567 65.94% 2012 4 243888 136159 64.17% 2013 1 251704 168344 59.92% 2013 2 51163 45618 52.86%
This pre-supposes that people don't go back and accept answers, which they definitely do. I would expect the later numbers to rise.
However, we've not yet got a complete quarter so it'd be better to look at the more recent, monthly, results:
Year Month Accepted Not Accepted Accept Rate ---- ----- -------- ------------ ----------- 2012 1 76557 33547 69.53% 2012 2 81080 36696 68.84% 2012 3 86580 40436 68.16% 2012 4 82512 39198 67.79% 2012 5 85058 40981 67.49% 2012 6 82040 40114 67.16% 2012 7 87293 44283 66.34% 2012 8 86311 43929 66.27% 2012 9 77275 41355 65.14% 2012 10 85138 46253 64.8% 2012 11 83439 46769 64.08% 2012 12 75311 43137 63.58% 2013 1 85316 52202 62.04% 2013 2 79915 54079 59.64% 2013 3 86473 62063 58.22% 2013 4 51163 45618 52.86%
Once again, the rate is dropping but we expect it to be lower as people haven't gone back over their older questions and accepted them yet. Equally, the number of answered questions in each month will continue to rise as time goes on.
I honestly doubt that there'll be significantly consequential evidence, from the data, of the effect of the removal of the accept rate for at least another year.
I would guess that the rate of acceptance will drop. This appeared to be happening, to a tiny degree, prior to the change and it'll be interesting to see what happens in the future. I don't believe that much can be read into these numbers at the moment.
answered Apr 27, 2013 at 19:02
-
Interesting. Any idea how these figures measure up against community growth?– BartCommented Apr 27, 2013 at 19:07
-
2There's also the problem of causality. Perhaps we don't give acceptable answers to as large a percentage of questions as before. Commented Apr 27, 2013 at 19:08
-
To steal from another answer @bart, pretty good until 2013. There has been a huge increase in the number of users asking questions and the accept rate barely seems to have slipped... Commented Apr 27, 2013 at 19:15
-
1only time can (definitively) tell on the last few months though. I guess to a very large extent it depends on the number of 1 rep users who post once and never accept. If this group increases as a percentage of the population then the incidence of acceptance will plummet precipitously. Commented Apr 27, 2013 at 19:16
-
I think it makes more sense to look at either questions where an answer is accepted in a relatively short period of time (say, a week) or to just look at the ratio of accept votes to questions answered (ignoring what time period those votes are targeted at). FWIW, the former has held fairly steady for a long time now, while the latter did drop off noticeably (although hardly catastrophically) in January.– Shog9Commented Apr 27, 2013 at 19:18
-
I'm not sure I see your point about the second @Shog. I would expect it to drop or are you saying you have the historical view of this ratio (not relative to time)? It is similar to my approach, though both do seem a bit roundabout. If you have better data than is available in the explorer then don't let me stop you answering! Commented Apr 27, 2013 at 19:22
-
Yeah, I'll post something in a bit; gotta finish packing first.– Shog9Commented Apr 27, 2013 at 19:29
Just did a little research, triggered by this remark in Ben's answer:
Equally, the number of answered questions in each month will continue to rise as time goes on.
I wondered how much effect this "late acceptance" would have on the overall trend. So I compared the given numbers with numbers from a fresh query. Here's the effect of late acceptance on accept rates:
Year Quarter 2014 March 2013 April Increase
---- ------ ---------- ---------- --------
2009 0,1%
2010 0,2%
2011 0,4%
2012 1 69,5% 68,8% 0,7%
2012 2 68,3% 67,5% 0,8%
2012 3 66,8% 65,9% 0,9%
2012 4 65,2% 64,2% 1,1%
2013 1 61,9% 59,9% 2,0%
2013 2 60,2% 52,9% 7,4%
As expected, the more recent answers show the greatest increase in accept rate.
How does this affect the observed trend of decreasing accept rate? Well, it's still there. Here are the results until the end of 2013, measured today:
As said before, this is just an observation. There can be any explanation for this trend. Maybe there are less acceptable answers, as Emil Vikström suggested.
One thing is clear: not displaying accept rate didn't cause the decline, nor did it affect it clearly. That's something I'm happy about, because I think it has removed a lot of noise from the site and, as is apparent now, without adverse effects.
