Online Review Censorship

Ample anecdotal evidence in the media notes that many businesses seek to ‘silence’ negative reviews, e.g., via legal threat. Despite attention toward this issue, we are aware of no systematic analyses addressing it. We address that gap here, leveraging review data from TripAdvisor.com. First, we estimate that ~1% of truthful reviews are deleted within six months of posting and that negative reviews are significantly more likely to be deleted, consistent with a mechanism of censorship. The effect is substantial; we estimate that a 1-star decrease in rating valence is associated with an approximate 25% (0.25pp) increase in the probability of deletion. Second, we examine how freedom of expression (FoE) in a country associates with characteristics of (uncensored) online reviews. We find that FoE associates with larger review volumes, lower review valence, and faster review posting. We discuss implications for online ratings platforms, consumers, and research opportunities