TurkSuite MTurk Tutorial

Please cite as: Alper, Morris (2014). "TurkSuite MTurk Tutorial". Available online at http://mturk.mit.edu/tutorial/

Contents:

0. Introduction
1. The MTurk User Interface
2. Simple Surveys
3. More Advanced Surveys
4. Analyzing Data

0. Introduction

Amazon Mechanical Turk (MTurk) is a useful tool for conducting research in the social sciences. Research commonly consists of the following stages:

  1. formulate a hypothesis
  2. solicit test subjects
  3. administer experiment
  4. collect and analyze results

MTurk makes steps 2–4 vastly easier. An experimenter can tell MTurk to find 1,000 random people in the world and pay them for answering his/her linguistic survey. MTurk will automatically administer the experiment, and return the results as a database file. This file can in turn be easily analyzed electronically. Clearly MTurk can save one a great deal of time if it is used correctly.

This tutorial will explain the basics of how to use MTurk for research in linguistics, and the information covered will also be useful to those in other fields which make use of human surveys. Section 1 discusses the MTurk user interface, including uploading experiments and downloading results. Section 2 covers making simple custom surveys with basic HTML. Section 3 covers making more advanced surveys using the TurkSuite Template Generator, and section 4 explains how to use the data resulting from your experiment once you have downloaded it from MTurk.

This tutorial uses a couple of terms which are specific to MTurk. MTurk users may be either requesters or workers. A requester is someone like you who creates experiments and pay subjects to take them. A worker makes money by taking these experiments. The experiments themselves are known as HITs — Human Intelligence Tasks.