Well the majority of people have experienced a rejection at a time or another. It would be interested in seeing what the average amount of rejections the majority of the turkers have so we can compare numbers just for future reference. I'm more interested in hearing your rejection total and your total amount of hits submitted!
5046 Submitted 4967 Approved 6 Rejected(3 from Imsportz Admin, 2 from Claritrans, 1 unknown.) 73 Pending(All from Imsportz, so they should all be approved after 30 days.)
Impressive! You have avoided mass rejections completely! Congratulations. Do you ever monitor your abandoned/returned rate, or do you not pay attention to this? I'm working on getting my returned / abandoned rate to both less then 1% also! Some say it don't matter, but it doesn't hurt to try and improve it the best you can I suppose?
HITs Submitted 5186 — ... Approved 5157 99.9% ... Rejected 5 0.1% ....Returned 159 3.0% ... Abandoned 27 0.5% I avoid batch HITS unless I'm sure they don't reject like Coman Petre, Cognitive Development etc... I return alot of HITS as I heard it doesn't matter too much. I prefer to return than get rejection any day.
I'm trying to avoid returning anything anymore. I am trying to use Oscar to get to 5,000 approvals. I am also very, very careful to the point that I avoid almost everything but the penny hits. I've done a few ClariTrans, but since I can hear them before I accept them, I generally don't have to return them. I think I've only returned one or two that I just couldn't convince myself that I had it right. I understand the 1st rejection from them. It was completely deserved and expected. The 2nd rejection from them I am pretty sure must have been unfair, but since I can't see them after they are submitted, I don't remember what it was. If I am doing Oscar, I never return them anymore unless I am unable to see a picture at all, and refreshing doesn't work. It won't accept a click on those, so I have no other choice. If called away, I will put the person's name and submit. Then at least I will be paid. Didn't use to do that, until I decided that I should. My current rate is: 5732 Accepted 5046 Submitted 88.0% 640 Returned 11.2% 46 Abandoned 0.8%
16k Approved 50 Rejected Honestly I don't even recall who the rejections were from now, they happened quite a while back. It was some sort of .02 [I think] HIT choosing categories for food items. The guy basically rejected 10% of what ever your total HIT count was.. across the board for everyone who did them.
... Approved 8077 100.0% ... Rejected 3 0.0% I wasn't turking for a for a while, but I'm getting back into it. Oddly enough, I'm not all too picky about who I work for, so I guess I just haven't had any bad luck so far.
Approved 15099 Rejected 75 About 50 of those are from when NetMSI started to be really strict. The rest are usually from some fault of my own or an error on the requester's side that they never reversed.
HITs You Have Submitted Value Rate HITs Submitted 114275 — ... Approved 113948 99.8% ... Rejected 285 0.2% ... Pending 42 — Majority are from doing about 80k classify this. I hardly get rejects from them now though.
Approved 11681 99.7% Rejected 35 0.3% My most recent rejection was from Crowdsurf, which I thought was weird because they've got very good Turkopticon ratings and I thought I was doing a very conscientious job working their transcription HITs. I've requested an explanation from them to see where I screwed up, if I did. My rejection before that was from Ignite Media Solutions, when I mis-transcribed someone's last name as "Gross" when it was "Rose".
You would have been auto approved if that happened on a Redwood. I don't think they reject anything, at least when I was reviewing transcriptions.
IMO, rejections (along with hit totals) are among the most meaningless statistics (and I'm not saying that out of bitterness; I don't think I've ever been below 99% approval). Yes, 'newbies' have to tread carefully until they build up a protective cushion. But beyond that, high approval is more of a reflection of discretion in choosing work rather than brilliant performance. The best requestors approve all honest work, so theoretically, any conscientious worker should hover close to 100. Hit volume isn't an accurate reflection of anything but persistence in working on low-paying hits (again, that's not a slight -- I also racked up big numbers with that strategy early on)
I can't help but disagree with you on this one. Just because HIT's don't take long doesn't make them low-paying. I would rather do 600 $0.01 HITs that can be completed (Including load time) in 5-6 seconds than a single $6.00 HIT over an hour.
I should have been clearer -- I meant low-paying per hit. And I agree; when quick and easy hits are available, they can return a great hourly rate. That was my point -- racking up hundreds of thousands of penny hits isn't hard to do, and if you steer clear of bad requestors, you build up an impressive and nearly rejection-proof statistic. Meanwhile, someone else who consistently works on high dollar hits takes a lot more risk. An analogy: The way the system works, you get as much credit (statistic-wise) for turning in a one-word report as the kid who writes a whole book.