Yeah, literally 45 of my rejects were due to that. I emailed him saying that I put the time in and did the work properly and need paid for it and my rejections needs reversed. Ew.
I didn't do any of those hits, but I suggest you all email him very nicely, explaining that Mturk isn't ran that way. That rejections are used for surveys and spammers. Throw in there the effects that rejections have on your livelyhood because you take turking very seriously like a 9 to 5 job. Then I would send him the links to mturks TOS and links on how to reverse. Just like it is not ethical for a requester to reject you based upon a mistake they made in their own hit, it's not ethical to reject everyone because they do not have a common answer. If he wants certain answers only, he could have done the work himself instead of depending on hundreds of people to answer things the same way. Here's some links that might be useful to him: https://www.mturk.com/mturk/conditionsofuse https://requester.mturk.com/help/faq http://docs.aws.amazon.com/AWSMechTurk/latest/RequesterUI/ReversingRejectedAssignment.html http://mechanicalturk.typepad.com/b...ignments-in-the-requester-user-interface.html sure it might be an earfull of bullshit, but if you word it right, it sounds like good bullshit. This is how I've managed to get all my rejections from the past year or so reversed. Unless they just do not respond to email at all, they always cave and reverse them.
Luckily I got tired so I only did 13, 4 rejected, but that's still 30% rejection rate from him 1 was for no majority, 3 for not matching. Deleted him from Turkmaster, there are so much better requesters to spend our time on. Edit: he reversed the 1 for no majority
Probably because they were working just fine when he first started loading them, but when he got up to 160K live hits they broke.
No I mean I get that. What he stated is a very specific issue. It makes sense too because the last time the batch was up it crapped out around after three hours. Whatever SS is doing calls to for the labels is what is likely expiring. I'm just curious the sourcing of that info. Sounds like he might have an insider on the case
Dudes....my rejections just went down from 68 to 39... He's changing all the ones that had no majority to approved. Leaving the ones that didn't match the majority. *sigh*
If you can get into TN they might have more insight since the Sergey people actually communicate with them.
It's at least a start. Hoping he reverses all the rejections and decides to try another method of approval altogether.
On the 'red' place somebody posted what they found was the problem and how they somehow got the topics to either show up correctly or just access them.
Prashant Majmudar can kiss my ass. Out of 55 hits, 7 rejected because I "Didn't match the majority". Guess I won't be doing his hits anymore.
Well, I sent him a nice long message (capped out the character limit in the message and had to edit) about the various ways to have better quality work for less money. Mainly doing quals or inserting quality control hits to verify workers efforts. The problem with his hits were, there were so many that I got that could go into multiple categories. Who is to say the majority was right? We will see how he handles the message and go from there. I did see that the TO dropped quite a bit though. Good job team on that one.
Exactly this. Just because someone doesn't agree with the majority doesn't mean they're wrong. I'm pretty sure a public defender's webpage for instance could go into legal or civil services.
6 min Title: Answer a survey about your behaviors, thoughts, and emotions Requester: Robert S Jobe-Shields [A1IZEVBDYMQOMJ] (TO) TO Ratings: No TO (Submit a new TO rating for this requester) Description: This is a psychology course survey about behavior, thoughts, and emotions Time: 2 hours Hits Available: 1 Reward: $2.00 Qualifications: Total approved HITs is not less than 100, HIT approval rate (%) is not less than 90, Location is US