Detecting and responding to fake responses to an online infant care survey in the UK

<ns3:p>Background This paper outlines the steps taken to identify fraudulent responses to an online survey of infant care practices and account for this in the subsequent analysis. A survey of infant care practices in England was conducted in 2022, offering a prize draw as an incentive, and promoted via social media with a public link. During the data cleaning process, it became clear that the survey contained fake responses and a process to identify and remove suspicious submissions was developed. Methods This method involved a 5-step process: 1) verifying genuine responses, 2) removing duplicate responses, 3) assessment of red and amber flags, 4) adding back in validated responses, 5) analysis of included and excluded responses. Results Overall, 209/3409 (6.1%) of responses were identified as suspicious and removed. Conclusion We present our reasoning at each stage of the process and suggest some principles which may be helpful for other research teams faced with a similar predicament.</ns3:p>