Taking the Plunge: Online Data Collection Using MTurk

Mechanical Turk, an online crowdsourcing service created in 2005 by Amazon.com, is gaining in popularity among academics as a source of subjects for survey research. The initial attraction is based on access to a very large pool of participants who can be accessed online producing rapid response at low cost with minimal effort on the part of the researcher. For the uninitiated, however, the task of designing and running a research project on MTurk is rather daunting.

This paper begins with an introduction to MTurk. We review the language of MTurk and explain how a researcher goes about setting up an online survey. Next we talke about several of the decisions that have to be made by the researcher. Decisions relate to the compensation rate to be paid, qualifications for respondents, timing decisions, and aspects of project design to aid in filtering out bad data. Such decisions are especially complicated when it comes to lengthy surveys.