Crowdsourcing Intelligent Research Participants: A Student versus MTurk Comparison

Abstract

The use of online workers as research participants has grown in recent years, prompting interest in how online workers compare to traditional accounting research participants. To date, no study has compared the relative intelligence of online workers to student subjects. Such a comparison may be important to behavioral accounting researchers given the homogeneity of accounting students relative to online subject pools and given prior research suggesting accounting students have relatively high analytic ability. While graduate students outperform both undergraduate and MTurk participants on common intelligence tests and across two reasonably complex tasks, we find that MTurk participants perform similarly to undergraduate accounting students. Overall, our results provide incremental assurance that MTurk participants are suitable subjects when accounting expertise is not explicitly required. We also provide evidence that screening MTurk workers on intelligence scores may benefit researchers who require participants with relatively high cognitive ability.