How Review Readers Cast Helpfulness Votes: An Empirical Investigation

Existing empirical studies examined how review and reviewer characteristics impact review helpfulness evaluation, assuming that helpfulness votes are casted based on voters’ perception of the review’s helpfulness and the helpfulness ratio aggregated at the review level can accurately represent review helpfulness. However, this assumption has not been explicitly tested, partially due to the lack of vote-level dataset. In this study, we utilize a unique vote-level dataset and investigate how context-driven biases (i.e., confirmation bias and bandwagon effect) and voter-level heuristics (i.e., cognitive inertia and declining perfectionistic strivings) impact the (un)helpful votes. Our result shows that voters’ helpfulness evaluation is vulnerable to these heuristics and biases. It suggests that past review-level research may over-estimate the impact of review-level and reviewer-level characteristics on review helpfulness evaluation, and the assumption that the review level helpfulness ratio is able to reflect consumers’ perception of a review’s helpfulness may not hold