On IRBs, this is the language that Gabe and I have used for the MIT IRB applications. We also open all our surveys with an IRB approved consent form (feel free to e-mail me if you want a copy). From our IRB's perspective the key is that the survey is voluntary and, because of the way MT is structured, we have no identifying information. It's much more straightforward than a lab study in this regard.
"The survey will be posted to Amazon's Mechanical Turk website, where people will be able to choose whether or not to complete the survey for a small monetary compensation ($.25-$.75). Subjects will be those who willingly choose to complete this survey on Amazon's Mechanical Turk site. The PI will have no access to the names of the subjects or any other identifying information. "
hope this helps
adam
_________________________________
Adam J. Berinsky
Associate Professor
Department of Political Science
Massachusetts Institute of Technology
77 Massachusetts Avenue
Cambridge, MA 02139 E53-457
Tel: (617) 253-8190
Fax: (617) 258-6164
E-mail: [log in to unmask]
Web Page: http://web.mit.edu/berinsky/www/
________________________________________
From: Political Methodology Society [[log in to unmask]] On Behalf Of Mark Fredrickson [[log in to unmask]]
Sent: Tuesday, April 26, 2011 3:00 PM
To: [log in to unmask]
Subject: [POLMETH] Online subject pools (summary)
Thank you to everyone who wrote in with suggestions for online subject pools. Here is a brief summary of responses.
1. Amazon's Mechanical Turk was the lead suggestion. Responses split between warning about representativeness and suggesting it is no worse than student samples or other convenience samples. Several people sent in citations:
- Berinksy, Huber and Lenz: "Using Mechanical Turk as a Subject Recruitment Tool for Experimental Research." http://huber.research.yale.edu/writings.html#26
- Paolacci, Gabriele and Chandler, Jesse and Ipeirotis, Panagiotis G. "Running experiments on {Amazon Mechanical Turk}." (2010) 5:5
For a 10 minute survey, respondents estimated costs of $0.50 to $1.00 per completed survey.
I invite additional responses on selling the Turk to IRBs. One respondent has subjects paste in a code at the end of the study. They need not fill in any substantive questions during the experiment to get paid.
2. For more representative groups, one respondent suggested Survey Sampling, Inc. The respondent estimated the cost to be $3/survey.
3. Another suggestion is slight modification of student samples: get access to University email lists.
4. Erin Cassese (WVU) sent me a working paper on "Socially Mediated Internet Surveys," having others promote your experiment on, e.g., blogs and facebook. The paper compares these samples with other convenience methods and looks at the results of several designs that employed SMIS. For a copy, contact Cassese ([log in to unmask]).
There was also interest in the software used to conduct online experiments. For my work, I wrote my own software that is hosted on Google's App Engine service. I plan to open source it soon, after I make it more general. I would be interested to know what other software platforms are being employed for online experiments.
Thank you to all the respondents,
- Mark Fredrickson
[log in to unmask]
http://www.markmfredrickson.com
**********************************************************
Political Methodology E-Mail List
Editors: Diana O'Brien <[log in to unmask]>
Jon C. Rogowski <[log in to unmask]>
**********************************************************
Send messages to [log in to unmask]
To join the list, cancel your subscription, or modify
your subscription settings visit:
http://polmeth.wustl.edu/polmeth.php
**********************************************************
**********************************************************
Political Methodology E-Mail List
Editors: Diana O'Brien <[log in to unmask]>
Jon C. Rogowski <[log in to unmask]>
**********************************************************
Send messages to [log in to unmask]
To join the list, cancel your subscription, or modify
your subscription settings visit:
http://polmeth.wustl.edu/polmeth.php
**********************************************************
|
