01/04 - Totally Snowy Tuesday

Jocelyn Belanger [A1FRNYP4JBDAB6] A survey on life habits and politics - max 20 min(~ 20 minutes) - $2.00 | PANDA

Unrated

Unrated

Unrated

$14.81 / hour
00:08:06 / completion time

​

Pros:

Cons:

Catchability: Okay (HIT Last Seen: 5 minutes ago)
Qualifications: TP Panel: -290823352 N >= 100; Exc: [339558-338969] DoesNotExist ; CR Research Group #1 >= 100; Location In US;

To read Jocelyn Belanger's full profile check out TurkerView!​

 

JH [A1S945OCUC30XR] (3-4min) Rating texts on five personality dimensions - $0.50 | PANDA

Unrated

Unrated

Unrated

$7.63 / hour
00:03:56 / completion time

Rejected 1/1 HITs!
Requester Feedback: ​

Pros:

Cons:
Rejected: Responses too random or meaningless for intentional human response

I passed all attention checks and provided thoughtful feedback. Bogus rejection and adding this requester to my BLOCK list

Catchability: Impossible (HIT Last Seen: 4 days ago)
Qualifications: HIT approval rate (%) > 90; Location EqualTo US;

To read JH's full profile check out TurkerView!​

 

SAI Team [AD1FJ5BDJ24AP] Academic Study of Psychological Foundations of Text, Memory (multi-phase, $0.50 + potential $9.25)(~ 90 minutes) - $0.50 $9.25 bonus | PANDA

Low

Unrated

Unrated

$9.74 / hour
01:00:03 / completion time

​

Pros:

Cons:
low pay still, despite the bonus.

Catchability: Okay (HIT Last Seen: 68 minutes ago)
Qualifications: TP Panel: 1412002035 N >= 100; Exc: [-1395117064-338878] DoesNotExist ; Total approved HITs >= 500; HIT approval rate (%) >= 98; Location In US;

To read SAI Team's full profile check out TurkerView!​

 

group_id makes more sense! This proposes to make both Approval and Rejection (since they're just mirrors of one another) by group_id, then, instead of per HIT?

 

As it was explained to me yes.

Though I'm kinda on board w/ $ amount now lol. Eating a rejection on a $5 andy gets real thunderdome status that way

 

Playing the devil's advocate here, @Sherry

Wouldn't the most likely result be, if Amazon did prevent mass rejections, would be to limit everyone from doing over a set threshold/certain amount, (which ironically would hurt the most proficient workers who have the most "padding") like Crowdflower and Upwork did, back in the day?

Additionally, if requesters had to pay for bad work, would that not discourage them from the platform, affecting all workers negatively in the long run?

With how few and far between mass rejections are, and how little impact they have on the top workers who do a lion's share of the HITS, shouldn't Masters/older accounts be against this petition?
Mass rejections are actually good for seasoned people, since they have little impact directly on older accounts with millions of HITs, and reduce the competition if a new person blows up their account.

-

I did notice several things from working as a "golden" class reviewer for certain requesters recently, group chats, etc.

There are most definitely people who abuse the system-There were a huge amount of people who abused the Noah Turk description Hits, just entering "a" "123" etc, then submitting. The amount of non-English-speaking people is also astounding some of the writing HITs.

Preventing them from mass rejections would just enable them to spam more, before they are caught. Limited consequences definitely would encourage some of those serial abusers to spam more, since they wouldn't blow up their accounts to an unacceptable approval rate. What type of consequences would you suggest for people like this?

-

The social system seems pretty decent now for identifying mass-rejecting requestors. They are quickly identified by addons, forums, discords, or hit approval in the default UI. Good and bad requesters alike get a reputation, and in this time of digital sharing, are often identified very quickly, sometimes in minutes.

-The risk/reward seems fairly balanced. If you do a new requester's HIT or batch, then you don't have as many addons grabbing for it, and you can get them more easily, but you face rejection chance. You don't know about them, they don't know about you. It seems fair, and properly adjusted in the long run for a freelance work site.


-All these items aside, suppose we did get Amazon to listen, and they addressed the issue. Are there not other issues that impact people more?

-Giving consistently high performance people Masters, setting benchmarks for it.
-Addressing VPNs and US-location-faking software
-Updating Amazon Requester data to be responsive (Some people have been 18-25 forever!)
-Standardizing CR#1, 2 or removing it, transparency on it.
-Improving Requester tools/Sandbox/interface to attract more Requesters
-Adding full or limited support for requesters, to attract more work.
-Reducing fees/becoming more competitive with other platforms to bring in more work.


Curious what you think. It seems like there are other issues that impact people more, surprised that mass rejections was the top issue!

 

I feel like experienced turkers won't see any real difference with this method, but it seems like it would stop new accounts from being instantly torpedoed bc they don't know better with working for bad/new requestors.

Experienced turkers will still do test batches, but might reduce some of the impacts of a requestor who suddenly starts mass rejecting despite not having mass rejected before).

 

THIS makes sense and was the language my brain needed. Thank you!

 

Jocelyn Belanger [A1FRNYP4JBDAB6] A survey on life habits and politics - max 20 min(~ 20 minutes) - $2.00 | PANDA

Generous

Unrated

Unrated

$21.49 / hour
00:05:35 / completion time

​

Pros:

Cons:

Catchability: Okay (HIT Last Seen: 8 minutes ago)
Qualifications: TP Panel: -290823352 N >= 100; Exc: [339558-338969] DoesNotExist ; CR Research Group #1 >= 100; Location In US;

To read Jocelyn Belanger's full profile check out TurkerView!​

 

Correct me if I'm interpreting incorrectly, but I believe the work will still be rejected after the threshold, and thus the requester would still not pay for it. However, any rejections above the threshold would not be used to calculate approval rating, and in that calculation only. They will still appear elsewhere in the worker's account.

 

Assuming they don't have motive like grandfathered incentives/discounts/data, if they mass reject, and get to 0% approval, as a company or institution, they could create another requester account as some requesters did, while workers are limited to 1 account only, no?

 

This is why I stay away from projects like this

Me after watching Spartacus for 1 week pitching to Bezos: Hey, what if we just murdered 1 out of 10 workers just so the others worked harder???

 

That was my thinking about the issue. I wouldn't see any difference in this because I have a large buffer and can afford a couple of thousand rejections before I am crying like a baby.

But then I could say, "back in my day, we had to be in the weeds with rejections. We had no help lol"

 

Said what boss?

 

In this pitch are you pitching that you get to choose who gets the axe or no?

 

@Sherry This! I have CR #1 and an over 99.9% approval rating, but I live in fear of getting the boot from 80%(?) of surveys by losing the qual with zero recourse or even understanding of why I lost it. Do y'all have any plans to address this?

 

I'm sorry, some of you might die, but that's a price I'm willing to pay.

No, I called dibs on this one.

 

Also,