05/24 - The Great Owl Tuesday!

[Squatty]

Title: Study on how people make decisions. $0.25 - $2.00. ~20/25 minutes | PANDA Requester: Empiricist1740 [A365K67FGB6U8B] (TO) TO Ratings: ★★★★★ 2.33 Communicativity ★★★★★ 2.50 Generosity ★★★★★ 5.00 Fairness ★★★★★ 3.25 Promptness Number of Reviews: 8 | TOS Flags: 0 Submit a new TO review Description: Complete a series of tasks. $0.25 for participation + bonus payment of up to $1.75. ~20/25 minutes Time: 50 minutes HITs Available: 1 Reward: $0.25 Qualifications: Location is US

 

[hyway]

I don't have the tenure of a lot of people here but I believe that people will be a lot happier if they stopped thinking Amazon Mechanical Turk exists for workers to make money. MTurk exists because Amazon already had the platform set up for their own use within their company. When they saw a market to sell it to other requesters without them having to do hardly any work to implement it, they jumped on it. EVERYTHING Amazon does with MTurk is based on getting money from requesters. They don't give a damn about the workers except where it impacts the flow of money from the requesters.

 

[Krikket]

Title: Tell us what you think and make some money! (~10 minutes) | PANDA Requester: Cocolab at UC Berkeley [A1N1YFYVJGJSBR] (TO) TO Ratings: ★★★★★ 4.84 Communicativity ★★★★★ 3.83 Generosity ★★★★★ 4.71 Fairness ★★★★★ 4.81 Promptness Number of Reviews: 132 | TOS Flags: 1 Submit a new TO review Description: Tell us what you think Time: 4 hours HITs Available: 222 Reward: $1.00 Qualifications: HIT approval rate (%) is greater than 95;Location is US

 

[Squatty]

^^^

 

[Maria Tortilla]

I don't have the tenure of a lot of people here but I believe that people will be a lot happier if they stopped thinking Amazon Mechanical Turk exists for workers to make money. MTurk exists because Amazon already had the platform set up for their own use within their company. When they saw a market to sell it to other requesters without them having to do hardly any work to implement it, they jumped on it. EVERYTHING Amazon does with MTurk is based on getting money from requesters. They don't give a damn about the workers except where it impacts the flow of money from the requesters.

 

[Eric Farmer]

It's a lot better on your sanity if you pretend masters doesn't exist.

I just focus on doing the best when I do HITs and I am trying to get more efficient. Thanks for the advice though.

 

[coffeeowl]

Title: Study on how people make decisions. $0.25 - $2.00. ~20/25 minutes | PANDA Requester: Empiricist1740 [A365K67FGB6U8B] (TO) TO Ratings: ★★★★★ 2.33 Communicativity ★★★★★ 2.50 Generosity ★★★★★ 5.00 Fairness ★★★★★ 3.25 Promptness Number of Reviews: 8 | TOS Flags: 0 Submit a new TO review Description: Complete a series of tasks. $0.25 for participation + bonus payment of up to $1.75. ~20/25 minutes Time: 50 minutes HITs Available: 1 Reward: $0.25 Qualifications: Location is US

So many typos on the preview page!

 

[hyway]

you think what i said was asinine?

 

[Hotlamp]

Got an email from a requester saying I qualified for a survey. Say the survey it said "Females only".

I'm a dude.

A girly dude?...

 

[vickiJ]

I don't have the tenure of a lot of people here but I believe that people will be a lot happier if they stopped thinking Amazon Mechanical Turk exists for workers to make money. MTurk exists because Amazon already had the platform set up for their own use within their company. When they saw a market to sell it to other requesters without them having to do hardly any work to implement it, they jumped on it. EVERYTHING Amazon does with MTurk is based on getting money from requesters. They don't give a damn about the workers except where it impacts the flow of money from the requesters.

So true! All of the "improvements" they made benefited requesters and not workers. Requesters got sick of workers emailing them and asking for quals, so Amazon said, great we'll hide those hits so if the worker doesn't see it, they don't know about it and they'll quit emailing and begging.

Requesters complained because workers complained about having to return so many hits due to screeners and other reasons. Amazon said, fine, we'll take that off the dashboard, so workers don't know how many they're returning. The thing they forgot to do and didn't care to do was to take away the qual that users like Dodd still use, that I can't do, because I've returned too many hits.

Requesters wanted a state qual for every worker. They got that. What Amazon won't do is change the qual if you move.


I could go on.

 

[Spindle]

I don't have the tenure of a lot of people here but I believe that people will be a lot happier if they stopped thinking Amazon Mechanical Turk exists for workers to make money. MTurk exists because Amazon already had the platform set up for their own use within their company. When they saw a market to sell it to other requesters without them having to do hardly any work to implement it, they jumped on it. EVERYTHING Amazon does with MTurk is based on getting money from requesters. They don't give a damn about the workers except where it impacts the flow of money from the requesters.

Oh, I don't disagree with you. It still doesn't excuse the fact that they pretty much lie to requesters about how masters are picked. What they're doing is fraud and I wouldn't be surprised if someday a few unhappy requesters don't go and open a lawsuit against them.

 

[Tribune]

So many typos on the preview page!

Could the requester have a "misspecified mental model"?

Perhaps Dirk Martignoni and Mechanical Turk aren't a good "cognitive fit." :whistle:

 

[aveline]

Oh, I don't disagree with you. It still doesn't excuse the fact that they pretty much lie to requesters about how masters are picked. What they're doing is fraud and I wouldn't be surprised if someday a few unhappy requesters don't go and open a lawsuit against them.

We don't know that, though. Nobody knows how masters is assigned except for Amazon. Sure, we all say it seems random, but that's still just speculation. How do we know they don't actually watch a few random batches or something and check for accuracy? Every time masters is handed out, we all compare stats to try to find out why one person got it and another didn't. The only way we can even attempt to compare accuracy/quality of work is by comparing how many rejections we each have. "Oh, two people have the same stats and same amount of rejections, but only one got masters. It must be random." Everyone just assumes that if they're not getting rejected on something then their work must be perfect. But with mturk, that's not always the case, and we've all seen that. They could very well be looking at stuff that we'll never be able to see. Just because we don't understand something, it doesn't mean that it's random or that it's somehow fraud.

 

[coffeeowl]

Could the requester have a "misspecified mental model"?

Perhaps Dirk Martignoni and Mechanical Turk aren't a good "cognitive fit." :whistle:

Glad I nope'd out!

 

[Spindle]

We don't know that, though. Nobody knows how masters is assigned except for Amazon. Sure, we all say it seems random, but that's still just speculation. How do we know they don't actually watch a few random batches or something and check for accuracy? Every time masters is handed out, we all try to compare stats to try to find out why one person got it and another didn't. The only way we can even attempt to compare accuracy/quality of work is by comparing how many rejections we each have. "Oh, two people have the same stats and same amount of rejections, but only one got masters. It must be random." Everyone just assumes that if they're not getting rejected on something then their work must be perfect. But with mturk, that's not always the case, and we've all seen that. They could very well be looking at stuff that we'll never be able to see. Just because we don't understand something, it doesn't mean that it's random or that it's somehow fraud.

Until you start looking at times when people get it when suspended. Also, people with multiple blocks on their accounts have gotten it, which should also affect if you are one of the "best" workers, etc.

Even though we don't know completely how it's done, there's been many pieces of proof that show that it's most likely just random.

 

[Bode "Emo Barry" Miller]

@kryssyy you're like 40 minutes late to your typical spamming hits to change the topic thing

 

[electrolyte]

@kryssyy you're like 40 minutes late to your typical spamming hits to change the topic thing

Discussing masters is relevant and doesn't really need the topic changed, though?

 

[Bode "Emo Barry" Miller]

Discussing masters is relevant and doesn't really need the topic changed, though?

help i was hard blocked by survey research and can't decide how many expletives to use in my email

 

[Jaded]

Until you start looking at times when people get it when suspended. Also, people with multiple blocks on their accounts have gotten it, which should also affect if you are one of the "best" workers, etc.

Even though we don't know completely how it's done, there's been many pieces of proof that show that it's most likely just random.

Blocks aren't a fair measure of someone being a good or bad turker though given that they are recommended by amazon for preventing retakes. People getting it while suspended makes sense if they are checking the work of a "golden" batch the person did /before/ they were suspended.

I'm not saying it's not random, but like @aveline said, we don't have all the information so just because it looks random doesn't mean it is.

 

[aveline]

Until you start looking at times when people get it when suspended. Also, people with multiple blocks on their accounts have gotten it, which should also affect if you are one of the "best" workers, etc.

I've always been skeptical of the people that claim to get masters while their account is suspended. People aren't always truthful on the internet. But even if it is true, maybe whatever system they're using to pick masters (because it's probably some sort of algorithm, I doubt any actual people are picking by hand) just doesn't take something like that into account. Also, blocks don't always necessarily mean that someone does poor work.

Even though we don't know completely how it's done, there's been many pieces of proof that show that it's most likely just random.

There's been no proof of anything.