A Data-Driven Analysis of Workers' Earnings on Amazon Mechanical Turk

Abstract

A growing number of people are working as part of on-line crowd work. Crowd work is often thought to be low wage work. However, we know little about the wage distribution in practice and what causes low/high earnings in this setting. We recorded 2,676 workers performing 3.8 million tasks on Amazon Mechanical Turk. Our task-level analysis revealed that workers earned a median hourly wage of only ~$2/h, and only 4% earned more than $7.25/h. While the average requester pays more than $11/h, lower-paying requesters post much more work. Our wage calculations are influenced by how unpaid work is accounted for, e.g., time spent searching for tasks, working on tasks that are rejected, and working on tasks that are ultimately not submitted. We further explore the characteristics of tasks and working patterns that yield higher hourly wages. Our analysis informs platform design and worker tools to create a more positive future for crowd work. Figure 12. (a) Hourly wage distributions of seven HIT categories provided by Gadiraju et al. [25] (with an additional category Research). (b) Strip plots showing median hourly wages of HITs associated with the topical keywords in Table 4.

When Beliefs Lead to (Im)Moral Action: How Believing in Torture's Effectiveness Shapes the Endorsement of Its Use

“…Other linguistic differences also emerged, but due to this article's focus on how effectiveness beliefs shape overall conceptualizations of torture, discussion of other tangentially related linguistic 2 This amount is reflective of the average compensation for MTurk workers (Hara et al, 2018). 3 This research utilized the LIWC 2015 system, which contains an extensive dictionary and contemporary software system that analyzes language files on a large number of linguistic dimensions.…”

The PhotoBook Dataset: Building Common Ground through Visually-Grounded Dialogue

“…In order to render reference grounding as clean as possible and facilitate automatic processing of the resulting dialogue data, participants are asked to try to identify the common and different images as quickly as possible, only describe a single image per message, and directly select an image's label when they agree on it. The compensation scheme is based on an average wage of 10 USD per hour (Hara et al, 2018). See Appendix B for a full account of the instructions and further details on participant payment.…”

Self‐Priming in Production: Evidence for a Hybrid Model of Syntactic Priming

“…Individuals who took longer than 25 min for any of the experiments were excluded from the analysis. In addition, we know that workers on Mechanical Turk are generally paid significantly below the American minimum wage on average (Hara et al, 2018). Thus, participants are incentivized to complete studies as quickly as possible, so that they can move on to the next paid task.…”