MTurk and Qualtrics Guide

Amazon Mechanical Turk

Amazon Mechanical Turk (MTurk) is a useful tool for behavioral scientists to reach a large and cost-effective participant pool. Conducting experiments on mTurk is fairly straightforward, but I've put together a couple guides and files that I hope will help other researchers both learn how to use MTurk and improve their usage of it.

First, I have a PowerPoint presentation explaining how to get started on Qualtrics, a user friendly survey platform and how to connect your Qualtrics survey data to MTurk through the use of a random number generator. The general idea is that MTurk brings in participants, but it's better to run the actual experiment elsewhere. This guide was created to help MBA students.

Qualtrics startup guide:

A second issue that researchers run into is the tedious task of manually awarding bonuses to their participants based on actions they took in a study. The MTurkR package on R provides a super easy way to award bonuses to participants by directly interfacing with the mTurk API. Download the MTurkR package, then fill out this file per your study.

Bulk bonus:

Lastly, when running a multi-part study, it can be useful to exclude participants who have already taken some form of your study. There are of course a couple ways of doing this, like banning IP addresses at your survey site, or blocking repeat workerIDs, but these can sometimes make workers upset. A better way is to "soft block" them, by creating a qualification that you assign to workers who have already been in your study that prevents them from being shown the re-listed task.

Soft block:

If you're looking for some good research about mTurk to learn more about its viability as a research tool, see "Who are these people?” Evaluating the demographic characteristics and political preferences of MTurk survey respondents by Huff and Tingley (2015), available here.

Also see An Analysis of Data Quality: Professional Panels, Student Subject Pools, and Amazon's Mechanical Turk by Kees et al. (2017), available here.