Exploring the relationship between emotionality and product star ratings in online reviews

Abstract

This article investigates the relationship between the review text valence and the product star ratings in user-generated reviews. The analysis performed on the Amazon data set using the generalised logistic regression method confirmed the negativity bias in online consumer reviews. The results indicated that negative sentiments in a review text are more influential in determining the star ratings of the products than the positive sentiments of the same magnitude in the review text. The results recommend that in all managerial decisions, the interventions to reduce negative performance disconfirmation be prioritised over the ones creating positive performance disconfirmation.