01/25 - Function Friday!

[nmelssx]

It is sad, and annoying and frustrating. But, rest assured, they're being dealt with as we speak. Luckily he got a hold of me and not some other requester though. Had to talk him off the ledge. We could have had another "scare". He thought it was all one guy with 30 accounts. It isn't. That said, @Metallica, your pet thread my get busier soon.


Grant money is easier to spend than "real" money.

What the hell is the qual on that HIT anyways?

Is it something like HIT approved over 100 and approved HITs over 50% or something? I can't imagine 30 workers with over 5,000 HITs completed and 99% approval rate doing something like this.

 

[Jerami]

"Dummy up, Edit'! Can't you see the turlette's busted?"

 

[ChristopherASA]

Good morning, Primus inter pares
[ @Jerami ]

 

[Metallica]

did they ever discuss why they hadn't put a stop to it sooner? that's ridiculous

Honestly, I don't believe they ever did.

I remember seeing that a lot of their funding came from Yahoo and a few other big companies, so I guess they had enough of that "fuck you" money to not care where it went... wonder if they ever managed to salvage anything from the batches for the project?

 

[Theta Carinae]

Better than I did. I was having an affair with beer yesterday. [drinking beer emoticon]

Bleh. It was brutal and dehumanizing. Some of the surveys I did... You definitely had the better day.

(also, is this what you were looking for?)

 

[xqyzt]

Man, why do workers have to make bad decisions? Just got through an hour of helping this requester. He got over 30 HITs scammed, in less than an hour, with random, wrongly labeled bounding boxes and nonsensical relationship data. I've been talking with him for a bit so, if you do get a rejection on this HIT, and you think/know that you did it correctly, let me know and I can probably be an MTC ambassador to Requestopia and I can see if I can broker an accord. If you know you scammed it, not that I suspect anyone, please don't ask me to fight to reverse it for you. It will probably end up worse, not better. Otherwise, let me know if I can help.

Gonna be an interesting day! Good morning MTC!

I totally agree with the motto you get what you pay for... HOWEVER, I think a lot of really high paying hits need to be more strict about the quals they put on them. Not slapping a masters on it, but maybe a higher hit count, I don't know. Just about anyone can do this hit; I really ride the fence on if that's okay or not. It gives people who do this more casually or newer people a chance to make a good buck, but it also leaves the requester vulnerable to super inexperienced workers who are not willing to try.

 

[Peachy]

Greetings from Amazon Mechanical Turk,

You've received a bonus from Stanford University for work related to 39RRBHZ0AVTKUL40NXNJ05F1AQSZV8.
The value of your bonus is: $10.00 USD

The Requester included this note:
I completely mistakenly rejected this hit here's the full payment really sorry!!

Thanks for being a Worker on Mechanical Turk!

the rejection remains from yesterday. took me 35m going at a decent clip

 

[ChristopherASA]

Bleh. It was brutal and dehumanizing. Some of the surveys I did... You definitely had the better day.

(also, is this what you were looking for?)

Yes, that is what I was looking for. Funny! and thanks!

 

[C to the J]

I totally agree with the motto you get what you pay for... HOWEVER, I think a lot of really high paying hits need to be more strict about the quals they put on them. Not slapping a masters on it, but maybe a higher hit count, I don't know. Just about anyone can do this hit; I really ride the fence on if that's okay or not. It gives people who do this more casually or newer people a chance to make a good buck, but it also leaves the requester vulnerable to super inexperienced workers who are not willing to try.

Qual tests would work best. Should be doing smaller versions of these and then qualing the people who do well.

 

[Theta Carinae]

yeesh, throwing this back. 5 minutes is too long

Title: Answer a survey about your opinions | PANDA Requester: Yoo yeon Sung [A1ECIJUHMZ6WEB] (Req TV): $10.11/hr/hr (TO): [Pay: N/A] [Fair: N/A] [Comm: N/A] [Fast: N/A] Description: keyword set recommendation Time: 1 hour(s) HITs Available: 1 Reward: $0.25 Qualifications: None;

didn't want to make cooking noises so early, but breakfast is seeming like the only option at this point.

 

[Gator]

@KYKITTY did you get this?
You've received a bonus from Leonardo Bursztyn for work related to 3DWGDA5POFUXDI8FYWAF2KUZHXE1VX. The value of your bonus is: $0.50 USD The Requester included this note: We sincerely apologize for accidentally rejecting your HIT. The mistake was due to the coding error, and we have spoken with the research assistant responsible to ensure that it does not happen again. If your rejection has not yet been reversed, it will be within the next few hours. We appreciate the work you do, and we hope that this small bonus compensates you for the trouble. Thank you to those of us who reached out to us to notify us of the error -- it is your diligence that makes MTurk such a valuable platform to us researchers. If you experience any further problems, please do not hesitate to contact us

 

[Peachy]

Greetings from Amazon Mechanical Turk,

You've received a bonus from Leonardo Bursztyn for work related to 3DWGDA5POFUXDI8FYWAF2KUZHXE1VX.
The value of your bonus is: $0.50 USD

The Requester included this note:
We sincerely apologize for accidentally rejecting your HIT. The mistake was due to the coding error, and we have spoken with the research assistant responsible to ensure that it does not happen again. If your rejection has not yet been reversed, it will be within the next few hours.


what a week

 

[4ourfuture]

Greetings from Amazon Mechanical Turk,

You've received a bonus from Leonardo Bursztyn for work related to 3DWGDA5POFUXDI8FYWAF2KUZHXE1VX.
The value of your bonus is: $0.50 USD

The Requester included this note:
We sincerely apologize for accidentally rejecting your HIT. The mistake was due to the coding error, and we have spoken with the research assistant responsible to ensure that it does not happen again. If your rejection has not yet been reversed, it will be within the next few hours.


what a week

Mine is still pending from yesterday morning. I'm still mad about the rejection from terence yesterday as well.

 

[Peachy]

Mine is still pending from yesterday morning. I'm still mad about the rejection from terence yesterday as well.

yeah mine is still pending. also got that same rejection from Terrence. at least Leo seems forgiving. Terrance told someone the data had been deleted and that was that. we’ll see. at least one person contacted his IRB about it

 

[Jerami]

I totally agree with the motto you get what you pay for... HOWEVER, I think a lot of really high paying hits need to be more strict about the quals they put on them. Not slapping a masters on it, but maybe a higher hit count, I don't know. Just about anyone can do this hit; I really ride the fence on if that's okay or not. It gives people who do this more casually or newer people a chance to make a good buck, but it also leaves the requester vulnerable to super inexperienced workers who are not willing to try.

It's not that it happened, so much as he didn't understand what was happening. Once he understood the situation, finding the culprits was easy. That said, scammers usually have ultra-high HIT counts. This wasn't inexperienced workers, at all. This was a complete disregard for any quality of data, non-native/expert English speakers trying to provide semantic data, and submitting a hit that should take around an hour +/- in an unreasonable short time. A bit of code was able to weed out most of it pretty quickly. Mostly, he wasn't sure how to deal with it. He thought it was one person, then when he realized it was more than one, and what was really happening(after I explained that one person doesn't probably doesn't have multiple accounts, and described the only reasonable inference from the data he was receiving) he was worried about rejecting people. I said reject and block 'em if they scammed. That's the only way to deal with that.

Bottom line though, Masters doesn't guarantee better data, always. Scammers have high HIT counts. Multiple semantic attention checks and closed quals are the only way to deal with it, imo. Just because a qual is "closed" doesn't mean it can't repost regularly. That's my story and sticking to it.

I Misread and didn't understand TV avg hourly. It had nothing to do with anything. MY FAULT ENTIRELY!! Just thought it was some interesting math based on useful data provided by a useful program.(which I've never used)

Also, if you did well, you may be granted a qual for further HITs in this project, to prevent further issues.

 

[Concise]

They remind me a few prior batches, by differently named requesters (couch, clock 3D descriptions). These have 5+ fields of entry though... good filler maybe. Some rejections on TO so work at your own risk -

Title: Label and describe a 3D car | Accept Requester: h2r [A2C6KDVCHUGUMC] Contact TV: [Hrly=$22.79] [Pay=4.40] [Fast=null] [Comm=null] [Rej=0] [ToS=0] [Blk=0] TO: [Pay=3.20] [Fast=2.75] [Comm=1.00] [Fair=3.00] [Reviews=6] [ToS=0] TO2: [Hrly=8.80] [Pen=3.00 days] [Res=0%] [Rec=0%] [Rej=1] [ToS=0] [Brk=0] Reward: 0.10 Duration: 4:00 Available: 4318 Description: Please rate the car based on the following properties, and give a one sentence description of the car Qualifications: Total approved HITs GreaterThan 100; HIT approval rate (%) GreaterThan 95; Location EqualTo US

 

[Trucker]

Good Morning All.

 

[4ourfuture]

This Barense Lab - List features of objects hit for $8.33 is far more difficult than I had imagined.

 

[C to the J]

They remind me a few prior batches, by differently named requesters (couch, clock 3D descriptions). These have 5+ fields of entry though... good filler maybe. Some rejections on TO so work at your own risk -

Title: Label and describe a 3D car | Accept Requester: h2r [A2C6KDVCHUGUMC] Contact TV: [Hrly=$22.79] [Pay=4.40] [Fast=null] [Comm=null] [Rej=0] [ToS=0] [Blk=0] TO: [Pay=3.20] [Fast=2.75] [Comm=1.00] [Fair=3.00] [Reviews=6] [ToS=0] TO2: [Hrly=8.80] [Pen=3.00 days] [Res=0%] [Rec=0%] [Rej=1] [ToS=0] [Brk=0] Reward: 0.10 Duration: 4:00 Available: 4318 Description: Please rate the car based on the following properties, and give a one sentence description of the car Qualifications: Total approved HITs GreaterThan 100; HIT approval rate (%) GreaterThan 95; Location EqualTo US

When they first showed up about 18 months ago they were .50 for one sentence.
It was beautiful.

 

[C to the J]

This Barense Lab - List features of objects hit for $8.33 is far more difficult than I had imagined.

The instructions on the HIT page say *up to 15*.
I took that to mean that all 15 didn't need to be filled out.
I did 8-10 each.