Unveiling the Cross-Cultural Influence of Sentiment Variability on Online Helpfulness Votes
(A2026-132723)

AUTHORS

David DeFranza, Michael Smurfit Graduate Business School, University College Dublin; Pronobesh Banerjee, Indian Institute of Management Kozhikode; Promothesh Chatterjee, University of Utah

ABSTRACT

Consumer helpfulness votes assigned to reviews are a critical component of the shopping ecosystem as they influence profitability through product visibility, consumer trust, and purchasing behavior. However, what makes consumer reviews helpful both within and across cultures has received little attention. We find within-review variability is a predictor of review helpfulness, both within and across cultures. Using natural language processing across two large datasets and a lab experiment, we find that for reviews written in English there is a negative association between variability and helpful votes, consistent with Aristotelian non-contradiction. In contrast, for reviews written in Japanese, variability is positively associated with helpfulness votes, a pattern consistent with the predictions of naïve dialecticism. This research provides ecological validity to naïve dialecticism and extends the theoretical understanding to the formation and expression of review helpfulness.