When researchers at UC Riverside, the University of Connecticut, Brown University, UCLA, and Fordham University recruited young gay and bisexual men for an online health study, they faced an unexpected challenge: nearly 96 out of every 100 responses were fraudulent, duplicate, or otherwise illegitimate.
The problem reflects a growing tension in modern public health research. Online studies can reach people who face barriers to in-person participation—those living with stigma, privacy concerns, or geographic isolation—yet offering payment for participation creates an opportunity for scammers to game the system with automated bots and duplicate accounts. As research increasingly moves online, especially for sensitive health topics, maintaining data integrity has become as critical as recruiting the right participants.
The multi-university team, led by Brandon Brown, a professor in the Department of Social Medicine, Population and Public Health at UCR School of Medicine, set out to test whether rigorous verification procedures could solve this problem. They published their findings in the journal AIDS and Behavior, analyzing data from 9,321 individuals who completed an initial online eligibility screener for a randomized controlled trial testing a video-based informed consent intervention.
The filtration process was stark: of those 9,321 screener responses, only 2,637 met basic eligibility criteria. After applying automated fraud-detection methods alongside manual reviews and phone or video verification calls, the researchers verified just 251 entries as both legitimate and unique. Of those, 158 participants completed informed consent, and 115 finished the entire study.
What made the difference wasn't relying on any single approach. Automated systems caught most problematic entries, but they weren't enough on their own. Human reviewers and direct contact with participants through phone or video calls provided essential safeguards, revealing patterns that algorithms alone might miss. The researchers concluded that a layered approach—combining automated screening, human review, and participant verification—was necessary to protect research integrity.
"As public health research increasingly relies on online recruitment, the question is no longer whether fraudulent responses will occur when offering payment, but how researchers can identify and address them," Brown said. "Our study shows that a layered approach combining automated screening, human review and participant verification can substantially improve confidence in online research findings."
Yet verification work brought its own complications. Many potential participants simply didn't respond to phone or video calls—an understandable hesitation for people in marginalized communities with legitimate privacy concerns. The researchers had to balance protecting data quality with keeping participation accessible, an especially delicate consideration for populations already underrepresented in health research.
Brown emphasized that these verification procedures shouldn't be treated as optional add-ons but as core components of study design from the start. "Online research allows us to engage populations that are often underrepresented in health studies," he noted, "but maintaining trust in the resulting data requires careful planning and investment."
The authors call on the broader research community to invest in staffing and resources for verification, and to continue investigating which strategies work best across different study designs and platforms. As online health research expands to reach harder-to-reach populations, this rigorous approach to data integrity may be what separates credible findings from noise.