The use of crowdsourcing in addiction science research: Amazon Mechanical Turk

Abstract

Crowdsourcing, the use of the Internet to outsource work to a large number of people, has witnessed a dramatic growth over the past decade. One popular crowdsourcing option, Amazon Mechanical Turk (MTurk), is now commonly used to sample participants for psychological research. Addiction science is positioned to benefit greatly from crowdsourced sampling due to the ability to efficiently and effectively tap into populations with specific behavioral and health histories. The primary objective of this review is to describe the utility of crowdsourcing, broadly, and MTurk, specifically, for conducting research relevant to substance use and misuse. Studies in psychological and other health science have supported the reliability and validity of data gathered using crowdsourced samples. Promising research relevant to addiction science has also been conducted, including studies using cross-sectional designs and those for measure development purposes. Preliminary work using longitudinal methods and for interventions development has also revealed the potential of MTurk for studying alcohol and other drug use through these designs. Additional studies are needed to better understand the benefits, as well as the limits and constraints, of research conducted through crowdsourced online platforms. Crowdsourcing, such as on MTurk, can ultimately provide an important complement to existing methods used in human laboratory, clinical trial, community intervention, and epidemiological research. The combinations of these methodological approaches could help improve the rigor, reproducibility, and overall scope of research conducted in addiction science. (PsycINFO Database Record (c) 2019 APA, all rights reserved).