Caution when Crowdsourcing: Prolific as a Superior Platform Compared with MTurk
Abstract
Many researchers host surveys on online crowdsourcing platforms, such as Amazon’s Mechanical Turk (MTurk) and Prolific. Online platforms promise a convenient way to meet sample size needs while drawing on diverse pools that might not otherwise participate in science. Yet, the quality of data obtained from these platforms is often questionable, so the collection must be closely monitored and reviewed. This study aimed to independently determine which crowdsourcing pool best serves researchers who plan to recruit for online surveys. To achieve this aim, we analyzed data from a recently completed study that drew participants from both MTurk and Prolific. We screened the collected data for both cost and quality, focusing on measures of attention, duration, and internal consistency. We found that only 9.89% of MTurk participants (N = 354) and 43.34% of Prolific participants (N = 345) produced high-quality data; Prolific also proved to be the more affordable option. Researchers considering these platforms for recruitment may weigh the evidence to make decisions when developing their own recruitment strategies. Finally, we highlight best practices for social scientists conducting online research, including additional survey and screening techniques.
Study specs
Data from participants recruited via MTurk and Prolific were analyzed for cost, attention measures, participation duration, and internal consistency.
- Authors
- D OConnell,A Bautista
- Institution
- University of Houston,Webster University
- Sample Size
- N=699
- Study Type
- Evaluation Study
- Year
- 2025
- Human Data Platform
- Prolific
- Source
- View Source Google Scholar
Measured Outcomes
Comparison of data quality and cost-effectiveness between MTurk and Prolific for online survey recruitment.
Peer Review & Critical Discussion
Potential Selection Bias in 2023 Cohort
The participant pool shows a concerning overrepresentation of users from high-income demographics. Looking at Table 3, we can see that 78% of respondents had annual incomes above $75k, which significantly limits the generalizability of these findings to broader populations.
Non-naive Participants Issue
I've noticed a methodological concern regarding participant naivety. Given that Prolific users often complete multiple studies, there's a real risk that participants had prior exposure to similar experimental paradigms, which could confound the results.
RLHF Applicability to This Study Design
The implications for RLHF training pipelines are understated. If we accept the authors' conclusions about preference stability, this has direct consequences for how we should structure reward model training. The temporal decay effect described in Section 4.2 is particularly relevant.
Verify your expertise to join discussion
Create an account and verify your credentials to participate in peer discussions.