Title: Impressions Requester: Dr. Nathan Heflick [A1O4VDJWZOFGM1] (TO) TO Ratings: ☢☢☢☢☢ 5.00 Communicativity ☢☢☢☢☢ 3.38 Generosity ☢☢☢☢☢ 5.00 Fairness ☢☢☢☢☢ 4.11 Promptness Number of Reviews: 16 (Submit a new TO rating for this requester) Description: 15 minute study on how people form impressions Time: 60 minutes Reward: $0.35 Qualifications: Total approved HITs is not less than 100, Location is not IN, HIT approval rate (%) is not less than 90, Location is US 5 minutes
Came across a band on these songkicks, thought they seemed familiar. A simple youtube search proves that I know one of their songs, listened to it. Instant bad mood ensues. I really hate the effect music can have on me.
The TO you linked was not an example of a worker lying or us pulling their eyeApps Master Qual -- it was just us pulling their *regular* eyeApps qual ("quick demographics survey"). We do THAT fairly often, and it's very odd that they gave us a 1-star review, because we paid them for 100% of their hits! (It looks like they were mad we didn't e-mail them with an explanation quick enough.) Based largely on feedback from this forum, we've moved to soft-blocking suspicious workers or workers who aren't giving us helpful data (i.e. users who hate 98% of our jokes, users who repeatedly fail attention checks, etc.) rather than rejecting their work outright (again, except in the most extreme cases, for < 1% of workers). That particular user was giving us really bad results, but not so egregiously bad that we felt justified in not paying them (you have to reallllllly @#$% up to fall in that category). In fact, we soft-block, while still approving 100% of hits, somewhat often. This in no way hurts a worker's reputation or rating on AMT, and allows us to get reliable ratings from good workers without pissing anyone off (well, except that oddly angry worker!) For those curious, we soft-block under 10% of our workers (while still paying them for all of their hits), and we only soft-block + reject hits from < 1% of our workers. Also, as our internal metrics improve, we're able to move more suspicious/bad workers into the "soft-block and approve all work" category, which further reduces the number of people whose work we have to reject outright. Also, one last note: I actually don't think many TO workers are outright lying. Definitely a couple of them have been, but most of the negative reviews are from people we rejected (most of which were quite a while ago, when we weren't as good at soft-blocking preemptively before someone submitted thousands of crap ratings). Keep in mind, of course, that someone who gets rejected is vastly more likely to submit a TO review than someone whose work gets approved, which makes the TO a highly biased sampling of our more than 6,000 workers. That said, I still appreciate TO and encourage everyone to make up their own mind! If someone is still afraid of us, despite our > 99% worker approval rate, so be it. And we always encourage feedback -- over the last half a year, we've made a number of positive changes based on your guys' suggestions here. So, thanks
I'd just like to suggest you stick to pulling qualifications. We found out recently that Amazon doesn't seem to see the difference between hard and soft blocks.
the most needed thing you must do ASAP is to review the hit's done by workers quicker than it is of now.
what i meant is: now it takes them about5-7 days to approve and pay the hits. i'm suggesting to shrink that duration.
I've suggested that before. Newer people who could do them to have more HITS to their account are afraid to touch them because of the rumors of rejection, so they'll do a few and then have to wait five days to do more.
Oh, your referring to auto approve.....yeah its 5 days. Pretty standard for alot of batches. But it Would always be nice if it was faster
Title: Short 10 to 15 minute survey Requester: Christina Thompson [A2EFZ1D0VZYQ9J] (TO) TO Ratings: ☢☢☢☢☢ 0.00 Communicativity ☢☢☢☢☢ 3.33 Generosity ☢☢☢☢☢ 5.00 Fairness ☢☢☢☢☢ 5.00 Promptness Number of Reviews: 4 (Submit a new TO rating for this requester) Description: Help us understand how problem solving affects thinking processes. Answer a few word puzzles and then provide us some feedback. Time: 30 minutes Reward: $0.50 Qualifications: HIT approval rate (%) is not less than 80, Location is US