Hello,

I am currently doing research on how passwords are created all over the world. I created HITs on MTurk where I targeted different countries. So, for example, I made a HIT that could only be answered if the participant was in Ireland, and I did that for a lot of countries that MTurk says it has workers in. I made 45 different HITs for the different countries, but I only got 6 total responses from all the HITs that targeted a country outside of the United States. I was wondering if anyone had any thoughts as to why I wouldn't have gotten any responses from outside of the US.

Thanks.

 

Are your HITs currently up?

 

No, they were up a couple weeks ago during the middle of the week (4/13).
Is there some information you want about the HITs?

 

I am not an expert, but was going to look at one and see if I saw something amiss that might help.

 

Ah, I see, thank you for the offer. Is there a way I could share information about the HIT with you without having to republish it?
(Also, to be clear, I did publish it in the US and it did fine, it just didn't really get responses outside of the US).

 

Again, not an expert, but if all of your HITs mirrored the US version, then perhaps there was nothing wrong with the HIT itself and/or the Amazon qualifications you selected. So removing "Amazon" from the equation, was there any added features, such as VPN detection, that would prevent someone from accessing the survey? With only 6 responders, the chances are very very low that anyone of those 6 would have have left a review, but you could also check https://turkerview.com/requesters/ to see what people wrote about your HITs. Sometimes people will share comments such as "couldn't get this to work in Chrome, but works great in Firefox." If the HITs were still active, we could read the return reviews, which would have better data and point to potential problems. I don't know if you can see these after the fact, but that would be a @CT question.

Another thing to consider is that many markets outside of the US are considerably smaller. A good portion of work is geared towards the US labor force, so I would think that plays against you. I don't know any of this as fact, but lets say hypothetically Amazon tells you that there are 20,000 Irish turkers, but maybe in reality only 5% are active to semi-active because they can't find work. I would hate to steer you to CloudResearch, but if Amazon isn't giving you great numbers, CloudResearch arguably has loads of data that they would love to charge you for. You could also try other platforms like Prolific to see what they have to offer.

 

I didn't have any added features, just the one qualification for country. Thank you for linking me to that website. I didn't know it existed.
I've considered Prolific before, and I'll take a look at CloudResearch too. Thanks for the suggestions.

 

Panos Ipeirotis consistently publishes hits to track mturk demographics. https://demographics.mturk-tracker.com/#/countries/all
The other countries in the chart is probably 50% Canada. Prolific has a lot of participants from the UK because that is where they are based. CloudResearch uses Mturk so you will probably not get better responses from them.
This is a link to the main site which may be helpful. https://ipeirotis.org/project/mechanical-turk-tracker/

Whatever Amazon says about demographics, which is very little, you can't trust.