Selection biases in crowdsourced big data applied to tourism research: An interpretive framework

Abstract

Crowdsourced big data has faced growing criticism due to its quality issues, particularly selection biases. We propose an interpretive framework for understanding selection biases in crowdsourced big data applied to tourism research. Inspired by medical terminology, the framework was structured according to external manifestations, internal causes, and potential influencing factors. Using illustrative case data from six websites, the framework demonstrates the emergence and impact of selection biases. Specifically, crowdsourcing-based tourism analysis can be notably affected by online platforms and destination contexts. Crowdsourced samples may not provide a perfect representation of actual travelers due to skewness in gender, age, origin, etc. Tourism researchers and stakeholders are urged to acknowledge selection biases and respond judiciously in their academic and practical efforts. Our research addresses a timely data science issue and offers insights for advancing knowledge innovation and technological improvements in tourism.