Improving data credibility in online recruitment: Signs and strategies for detecting fraudulent participants when...

Online recruitment platforms are valuable tools that allow researchers to efficiently reach large pools of potential participants for clinical trials and observational studies. However, challenges associated with fraudulent respondents and bots threaten data credibility. This paper discusses signs we identified for recognizing potentially fraudulent respondents when using one online recruitment tool, ResearchMatch, to recruit study participants.

Participants were recruited via ResearchMatch for a study focused on stress and coping among U.S. young, low-income parents. Participants completed an online survey via Qualtrics. To screen for fraudulent respondents, we compared reports generated from Qualtrics with those from ResearchMatch.

We identified six signs of fraudulent respondents, including 1) suspicious metadata, 2) fake addresses listed in ResearchMatch profile, 3) a common Western name for both first and last names, 4) inconsistent and duplicate data, 5) extensive list of medications and/or medical conditions, and 6) short survey completion time. We contacted 63,284 accounts through ResearchMatch and received 928 survey responses. About 46 % (n = 425) of responses were deemed fraudulent.

Fraudulent respondents and bots undermine the integrity of data, which may result in adverse implications for research, policy, and patient health. Given that nearly half of the survey respondents in our study were deemed fraudulent, it is evident that this is a significant issue for recruiting research participants via online recruitment platforms. To ensure reliability and accuracy of study findings, it is critical for researchers to thoroughly examine their data for signs of fraudulence.