Online Review Generation in the Digital Age: A Meta-Analytic Study of Online Review Behavior
Abstract
A large body of literature on online review generation has investigated various antecedents to online review characteristics. However, the mixed and even contradictory findings in this domain have hampered a clear understanding of consumers’ online review behavior for both researchers and practitioners, thereby complicating actionable strategies. To address this issue, we conduct a comprehensive meta-analysis of 154 independent samples from 152 empirical papers, synthesizing the relationships between 28 antecedent factors and 12 review features, which we categorize into rating-based (evaluative) and non-rating (expressive) dimensions. Our results indicate that rating attributes are dominantly driven by stable reviewer tendencies and market-level reputation cues, whereas non-rating features—such as text length and image inclusion—are primarily influenced by reviewer social embeddedness and platform engagement. Crucially, we identify that inconsistencies in prior literature can be explained by critical boundary conditions, including platform type (business-hosted vs. third-party), goods type (service vs. tangible), and cultural background (Asia vs. North America). By distinguishing between evaluative and expressive review components, our study reconciles mixed findings and uncovers underlying mechanisms such as expectation-based evaluation and self-presentation. The findings provide a robust theoretical foundation for empirical model specification and offer actionable insights for retailers and platforms to optimize reputation management and system design. Ultimately, our work offers a holistic framework for understanding online review behavior and proposes a rich agenda for future research in the digital marketplace.

