Online survey data integrity: lessons learnt from investigating a bot attack on a small-scale educational study

A notable and growing risk in online research is fraudulent respondent activity using artificial intelligence. Given that researchers often rely on anonymous recruitment methods to reach hidden and diverse populations, it is essential to develop contextualised anti-bot protocols sensitive to research scale, topic and population, allowing sufficient data to be collected while detecting bot activity and safeguarding data integrity. This paper reports the investigation of a bot attack on a small-scale survey capturing educators’ and parents’ views regarding the inclusion of same-sex parented families in Australian early childhood settings. The survey was administered using LimeSurvey, configured to maintain anonymity and distributed via multiple free and paid channels, with an incentive to participate. This paper presents a post-hoc data cleaning protocol used to filter bot-automated responses. Lessons learned highlight the potential of balancing anonymity, response rates and data integrity through a nuanced, multifaceted approach.