Note: clicking on the above banners and making ANY purchase returns a commission to Turker Nation.
If you can't see the ad, please click on Shop on Amazon instead. | Want to advertise here? PM Spamgirl to learn more!




Results 1 to 9 of 9

Thread: How to deal with foreign workers doing US-only HITs?

  1. #1

    How to deal with foreign workers doing US-only HITs?

    I set the location as US in mTurk. However, when I downloaded data from survey gizmo, there were two participants in India, one in Romania and one in Cairo. Any ideas on how this happens and what I can do to prevent it?

    Thanks
    Last edited by spamgirl; 09-14-2014 at 07:40 PM.

  2. #2
    Community Manager spamgirl's Avatar
    Join Date
    Aug 2011
    Gender
    Female
    Location
    Canada
    Posts
    12,494
    Thanks
    2,624
    Thanked 6,855 Times in 2,213 Posts

    People lie about their location when they sign up for mTurk so they can get payments and do HITs that they're not supposed to do. There is no way for mTurk to police it as these people also use false proxies to cover up their IP address location. There is nothing you can do to prevent it.

    Your best bet is to block those people from working on your HITs and report them to mTurk for lying about their location so they can be removed from mTurk and stop cheating requesters.
    It's all good.

  3. #3
    Requester
    Join Date
    Dec 2011
    Gender
    Male
    Location
    Provo, Utah
    Posts
    32
    Thanks
    8
    Thanked 5 Times in 5 Posts

    It sounds like you're doing a survey that needs only US participants, and that's what I do too, so I'll share with you how I handle that problem. I include "Location is United States" as a qualification, and I check the setting that requires them to be in the US to view the HIT. That filters out almost all non-US participants. Then, at the end of my questionnaire, I ask in a rather matter-of-fact way which country they live in. As part of my normal data-cleaning procedures, I filter out people who said they didn't live in the US. I admit, I paid them when they actually didn't qualify for the HIT, but it doesn't happen very often, so I don't think the lost money is a big deal. I also admit they could lie again (i.e., when they signed up and again when they complete my questionnaire), but believing people is inherent to survey research and I really don't know a better solution, so I just go with their answers. That's my $.02.

  4. #4
    Requester
    Join Date
    Dec 2012
    Gender
    Male
    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    I'm in a similar situation, but reversed. I have two one-per-worker survey HITs that are identical except one requires location US and the other location Canada (since, unfortunately, I can't set "location is US OR Canada"). Other qualifications are 90% approval and at least 50 HITs completed. Pay is $0.40 for 5-10 minutes; average completion time is about 8.5 minutes, but that includes vast numbers of dawdlers who accept, then respond later. The US survey is rolling along in a third batch of 500, but the Canada batch has been up for 20 hours and has exactly one respondent. Is this possibly because Canadians have strong incentives to pretend to be in the US to increase their eligibility for HITs? Or is there something I need to do to help Canadian Turkers find the HIT? I can't imagine I need to pay more, since the exchange rate today is exactly 1:1.

    SpamGirl, thanks for the quick help getting started on this forum yesterday. I see your profile says you're in Canada. Any insights?

  5. #5
    Community Manager spamgirl's Avatar
    Join Date
    Aug 2011
    Gender
    Female
    Location
    Canada
    Posts
    12,494
    Thanks
    2,624
    Thanked 6,855 Times in 2,213 Posts

    Did you know that minimum wage in Canada is $10.25 per hour?

    I personally wouldn't do your HIT as it only pays $4.80 an hour at best. Not worth my time when other requesters pay a fair wage. Very few Turkers will work for less than $6/hour. In fact, most of the "US" workers you're getting are probably Indian at that rate. They love to buy US accounts.
    It's all good.

  6. #6
    Requester
    Join Date
    Dec 2012
    Gender
    Male
    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Understood, Spamgirl. The observation on Canadian minimum wage is very helpful--that hadn't occurred to me.

    We clearly don't have all Indians, unless these Indians are fantastically skilled at sending email in current US slang and style. We've received feedback from many workers suggesting they found the pay to be fair. Of course, those may be the sort who are just looking for a few surveys to pay for an extra MP3 download from Amazon, not those trying to make a living.

    I just had a long, very friendly email exchange with a woman who used to have an apartment in Manhattan and now is trying to cope with very limited income. She thanked us for our approach to administering our work, but did mention that she normally doesn't look at anything under $0.50. As I told her, as a liberal arts school that doesn't do a lot of national-scale research, we are still grappling with the implications of moving from a long history of doing research where subjects are essentially all volunteers to one in which we have to regard them as, more or less, very-short-term employees. I have no intention of being cheap or stingy, and I want to recognize the equity issues. However, our current research clients are comparing mTurk to using students in the classroom, who are required to participate in a study or two each semester. We'd like to think of it more like sticking a few coins in the envelope to say thank you, rather than paying a living wage. We also face the counterintuitive problem that it's possible to pay too much for survey data, compromising the study.

    In any case, I'll definitely talk to our client about increasing the per-survey budget; considering that we're not quite on schedule, I suspect they'll be amenable. I'll also review the latest research output on pay rates. Thanks!

  7. #7
    Requester
    Join Date
    Dec 2012
    Gender
    Male
    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    We boosted the pay for our Canada HIT to 0.75 and then to 1.00. Still just a tiny, tiny trickle of submission, while the US data continues to pour in at a new higher rate of $0.50. So while I will to keep working on the fair pay issue, I don't think the problem with the Canada version HIT is mainly pay related; we ought to get a least few more responses as the pay goes up.

    That leaves the location qualification as the only possible explanation, since everything else is the same. Any other ideas why a location qualification for Canada might kill a HIT?

  8. #8
    Community Manager spamgirl's Avatar
    Join Date
    Aug 2011
    Gender
    Female
    Location
    Canada
    Posts
    12,494
    Thanks
    2,624
    Thanked 6,855 Times in 2,213 Posts

    I know others who have posted HITs for Canada only and haven't had a problem - in fact, I work on many of them. $1 is still only $6/hour, that's the only problem I could see. Canadians who are willing to Turk are likely to only work on higher paying HITs due to the higher pay here. I know many Canadians who Turk, and they won't work for less than $12/hour. My minimum is $20.
    It's all good.

  9. #9
    Requester
    Join Date
    Dec 2012
    Gender
    Male
    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    OK. We'll try a few cases again at $2.50 and see what happens to confirm the theory (large shares finish the survey in under 5 minutes, so that's almost $20/hour). I doubt our client will pay $2,000 for 1,000 survey responses, though. Thanks again for the advice.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •