Let's Talk aBOT Scam Online Survey Completions in Health Behavior Research: Tutorial With Case Studies, Practical Guidance, and a Checklist for Researchers.

BACKGROUND: Online data collection can reach large populations efficiently and cost-effectively. However, the increase in bots and scammers (ie, person- or software-based fraudulent completions) completing online surveys raises data integrity issues and wastes scarce research resources. OBJECTIVE: This paper aims to describe case studies and experiences in which bot or scam completions of online surveys occurred within the health behavior field (specifically physical activity and nutrition). Lessons learned and a checklist of strategies to assist researchers before, during, and after data collection to reduce the incidence of and identify bot or scam completions are provided. METHODS: Four diverse case studies are presented from studies that used online recruitment and data collection methods for cross-sectional surveys by parents about children's screen time, cross-sectional surveys by adults about transport-related physical activity, qualitative interviews for a proposed trauma-informed physical activity program for female victim-survivors of intimate partner violence, and the Australian component of a large multicountry prospective study targeting university students. The strategies used to identify and prevent bot or scam online survey completions are explored. RESULTS: High levels (7%-80%) of suspected bot or scam completions were identified in a number of these studies. Participant characteristics and outcome variables were significantly different between included and excluded participants (eg, excluded responses had a higher percentage of male parents and children, higher social media use, and lower physical activity guideline adherence). The learnings from these case studies and the wider literature are combined to create a checklist of strategies that researchers can use to prevent and identify bot or scam completions. These include strategies before data collection (when creating study collateral), during survey design and development (including the use of inbuilt platform functions and the design of the survey questions and structure), following data collection (indicators of potential bot or scam completions), and recommendations for reporting of bots or scams. CONCLUSIONS: The checklist, based on the included case studies and wider literature, can be used to help researchers who use online recruitment and data collection methods at each stage, from planning and conducting through to analyzing and reporting their findings. Researchers should include several steps to prevent and identify fraudulent survey responses when creating surveys and completing data cleaning. This checklist should also be considered in grant applications and ethics applications. This will provide greater confidence in the research findings and reduce unnecessary waste of research time and resources.