03/23 - Fantastic Friday!

I could only do 4.

 

I'm pretty sure you'll be okay

 

Once you're past 3-5k you can pretty much jump into unrated stuff, you just have to be smart about it & learn how to analyze just how risky each instance is.

Surveys: risk doesn't matter, one rejection can't hurt you past 5k so just do w/e looks decent. The only "risk" here is wasted time. Never chase rejections, let someone else do it. Let it go, move on, focus on anything other than a rejection if it even happens.

Batches:
1) Communication is everything, fire off a quick message w/ any ol' banal question about the instructions or something. If you get a reply, and its in native English, you're good. Is it from a .edu domain? You're good. A real business? You're good. Is it from rrr66798@russia.ru? Later bro good luck w/ your work. You can include the instructions/format of the HIT in here. Instructions are the basic level of communication between Req/Worker, so if they're sparse or non-existent, welp you know what you're walking in to.

2) Sans communication, quals can still tell you almost everything you need to know. Masters? Presumably you don't have it so it doesn't matter anyway, but masters only w/ nothing else is as risky as work gets b/c only Requesters who have put next to no effort in learning the platform use it. They're going to get forked, and probably forked hard, so its actually scared money. Except the forced Masters templates, those are safe but you'll never see the Req again anyway.

Location quals? If it isn't limited to the US/UK/CA/NZ/AU - be careful if its anything that you'd think the Requester needs quality & a fundamental understanding of the English language for. Text parsing w/o a US qual? Just don't bother, the Req is not going to like what he gets & the risk of him having to just reject the entire thing & try again goes up exponentially vs properly qual'd batches. Risk a few if you want to show quality in case he quals it, IMO, but don't go ham.

Does the batch have location quals, >98% approved, & >5k HITs? Its safe. The req knows what they're doing & what to expect out of mturk. Unless you have a lobotomy halfway through the batch frankly you'll be fine. I wouldn't do 1k of any one batch but "betting" half of your cushion on them will snowball your #s and earnings pretty quickly, especially if you keep a good eye open for the super safe stuff as well.

Individual anecdotes of the above being "untrue" don't change the fundamental decision making in evaluating work on the platform. Sorry. Plus, frankly, people lie (to themselves mostly).

Small fringe/niche details can color the above, but IMO quals/communication are all that really matters.

Anything artsy/fashion related is NEVER safe. I joke about it a lot, but I'm not kidding here. Stay away from non-A9 fashion batches until you've got a full day of padding built up there's no point.

Crap instructions (or great instructions) can make up for qual discrepancies (or negate decent qual setups).

Always e-mail the Requester and ask. Most will let you know their rejection/block policy.

 

Yesterday there was one batch that had a 2 cap and then a second with a larger cap. You never know ahead of time with them. Current one is 4, which sucks. Let's hope another .10 batch follows.

 

I guess I shouldn't have accepted 15...

 

SmithOscar [AKL99MZHRN058] A survey on language - $0.85 | PANDA

[3V7AFCJQHCNAF55RXG0TZUZVLOK4AX]​

Fair

Unrated

Unrated

$7.91 / hour
00:06:27 / completion time

​

Pros:

Cons:

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

 

I accepted 29 and dumped a C-SAT. I mean, wish I held on to the C-SAT, but beyond that, on the the next!

 

Yeah, I'm capped at 4 today. Yesterday I think it was 25 for me.

 

Yueran Yang [AX4SFTZ3M5XJ2] People perception - $1.00 | PANDA

[3BJ81NIZBGH7RX39EJZ0BJK9SE36KX]​

Generous

Unrated

Unrated

$31.03 / hour
00:01:56 / completion time

​

Pros:

Cons:

To read Yueran Yang's full profile check out TurkerView!​

 

Title: Study on Decision Making(~ 15 minutes) [ $ | $ ] PANDA
Requester: Ben Schenck [A3DR0X7KXQOMCI]
TurkerView: [ $20.62 / hour Generous | Unrated | Unrated ]
Description: Study on decision making with a near-guaranteed bonus ($2.50+ in total payment). You will ACTUALLY be paired with another member of mTurk, and the study will require you to be patient and wait for the other person. Please only focus on this task. (15-20 minutes)
Duration: 01:00:00
Available: 4
Reward: $2.00
Qualifications:

• Total approved HITs GreaterThanOrEqualTo 100,
• Exc: [2543318-79810] DoesNotExist [ Request | How to Use ]
• HIT approval rate (%) GreaterThanOrEqualTo 90,
• Location In US,

[3AX4X2ATGI51F8NUY7PVAQRI9ZSYDV]​

 

Since the cap is only 4, I just went ahead and returned them all. I like to do those with headphones, and I didn't want to go get them for $.40.

 

You should've just done the 4, it takes less than 2 minutes to do that many. If you opened any of them, you're locked out of Endor for an hour.

 

Did anyone receive a rejection for the Mark Chang HIT done today?

 

Is there a way to quickly add a hit from HIT Finder to PC?

 

Jihwan [A1LOGBFMSAWF5S] 2 min consumer study (2 min; 20 cents) - $0.20 | PANDA

[33CCZ2QQZLLJ4W5GZBBEI50KRV19E2]​

Fair

Unrated

Unrated

$8.89 / hour
00:01:21 / completion time

​

Pros:
Simple and quick.
Cons:

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

 

Stanford Language and Cognition Lab [A1XUBU5CS5SFRV] Learning from animals - $0.80 | PANDA

[36DWM50OFIAGSGW2X4LHQZFV3EGEK2]​

Generous

Unrated

Unrated

$12.91 / hour
00:03:43 / completion time

​

Pros:

Cons:

To read Stanford Language and Cognition Lab's full profile check out TurkerView!​