Mechanical Turk Demographic Trends During the Year of COVID-19

Aaron Moss, PhD

published on April 16, 2021

Last month, people in the U.S. marked the one year anniversary of the COVID-19 pandemic. For most of 2020, circumstances on the ground were changing faster than many forms of data could catch up with, and there was fear this may be true with the demographic composition of people on Amazon Mechanical Turk. So, after one year, we decided to look at whether there were any major changes to the demographics of people on MTurk as a result of the pandemic.

In a recent blog, we summarized data that CloudResearch published in a letter in the June issue of Trends in Cognitive Sciences. In that letter, we reported that the demographics of people on MTurk had not shifted much during the first three months of the pandemic. Here, we take a broader view and report on the demographic trends we examined across all of 2020.


129,428 Workers Completed Studies through CloudResearch

In 2020, just under 130,000 people completed a CloudResearch-hosted study on MTurk. This number represents nearly a 7% increase from 2019 and almost a 20% increase from 2018. For the last five years, we have seen consistent growth in the number of people completing CloudResearch-hosted studies.

Despite significant upheaval to people’s lives during 2020, there was a similar pattern to the number of people completing studies each month. As the red line in the figure below shows, at no point in 2020 was there a sudden change in the number of participants completing studies.

Line graph comparing the number of MTurk participants per month during 2019 (blue line) and 2020 (red line), showing similar patterns throughout the year with no sudden changes despite the COVID-19 pandemic

Figure 1: Participants per month during 2019 and 2020


Overall Participant Demographics Didn’t Change Much

CloudResearch regularly gathers demographic data from workers on MTurk. Over the course of 2020, we did not see large deviations from 2019 data in terms of participant gender, race, ethnicity (Hispanic or not-Hispanic), age, or household income. 

The stability in household income is noteworthy because some people suspected that a portion of the tens of millions of people who lost jobs during the year may turn to platforms like MTurk for money. The data we have do not suggest this happened.  

Four pie charts showing the stable demographics of MTurk participants across 2019 and 2020, displaying distributions of gender (roughly 50/50 male/female), political party (more Democrats than Republicans), race (predominantly White), and ethnicity (mostly non-Hispanic)

Figure 2: Gender, Political Party, Race, and Ethnicity

Stacked bar chart comparing age distributions of MTurk participants between 2019 and 2020, showing very similar patterns with the largest percentage of participants in the 25-34 age range, followed by 35-44, with minimal changes during the pandemic

Figure 3: Participant Age

Stacked bar chart comparing household income distributions of MTurk participants between 2019 and 2020, showing consistent patterns across income brackets with no significant shifts despite the economic impacts of the pandemic

Figure 4: Household Income


Participant Join-Rates Showed Seasonality

The rate of new people joining MTurk throughout 2020 largely mirrored the month-by-month progression of 2019. In the second half of the year, and especially in the fourth quarter of the year, new workers joined MTurk at a slower pace than they did in 2019. Overall, however, the rate of new people joining shows a seasonal effect that was largely similar in both years.  

Line graph comparing the number of new MTurk workers joining per month during 2019 (blue line) and 2020 (red line), showing similar seasonal patterns but a slower join rate in the second half of 2020 compared to 2019

Researchers Ran More Online Studies

Finally, at the risk of sounding obvious, researchers ran more online studies in 2020 than 2019. A month-by-month comparison shows a sharp rise in April 2020—the month after all in person data collection stopped—before a leveling off in the summer and fall months.

Although COVID-19 dramatically changed how researchers gather data, it also led to some remarkable opportunities like this study that used CloudResearch to study how lockdowns affected people’s level of paranoid thinking.

Line graph comparing the number of studies run on MTurk during 2019 (blue line) and 2020 (red line), showing a sharp increase in April 2020 after in-person data collection stopped due to COVID-19 lockdowns, with consistently higher study volumes throughout 2020 compared to 2019

Learn how CloudResearch can make it easy to run studies on Mechanical Turk.

Share this blog

Related Articles

Best Practices That Can Affect Data Quality on MTurk
Best Practices That Can Affect Data Quality on MTurk

Amazon’s MTurk has given researchers a powerful tool for connecting to survey respondents online. See how six simple best practices can help improve your data quality. ...

Read More >
How Has COVID-19 Affected MTurk Worker Demographics?
How Has COVID-19 Affected MTurk Worker Demographics?

The COVID-19 pandemic has affected everything about how we work. New research from the CloudResearch team dives into its effects on MTurk worker demographics ...

Read More >

SUBSCRIBE TO RECEIVE UPDATES

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.