Can We Trust Online Surveys in Community-Recruited Research?

Survey research online offers advantages but is vulnerable to fraudulent responding. Although studies using niche samples, survey panels, or social media recruitment show that fraud is common, few have examined it in online surveys targeting broad populations through community-based recruitment. This study investigated fraudulent responding in an online survey on barriers to job-related training recruited solely through community organizations with minimal eligibility restrictions. A points-based algorithm (30 detection variables) was applied iteratively to assess response legitimacy, and relationships between demographics and validity status were examined using binary logistic regression. Of 1952 surveys, 56.1% (n = 1,096) were deemed fraudulent. Duplicate IP and suspicious e-mail addresses were the most frequent violations. All demographic variables were significantly associated with validity status. Findings show that fraudulent responding is pervasive in community-based online surveys. Strengthening data integrity requires the implementation of flexible, quasi-standardized fraud detection methods and transparent reporting of fraud screening in published online research.