How online review environment affects review generation behavior: A competing for attention perspective

Abstract

From the perspective of competing for attention, this study attempts to examine how a potential reviewer’s review content in terms of topic diversity and topic popularity is affected by the review environment, which is characterized by review volume, review variance, and time distance. The empirical analysis is based on 70,383 restaurant reviews collected from Yelp. The Latent Dirichlet Allocation (LDA) model is adopted to conduct review text mining. Our empirical findings indicate that reviewers are more likely to evaluate the product on a wider range of topics when exposed to a larger volume or lower variance of existing reviews. Our findings also show that reviewers prefer to talk about popular topics as the volume of prior reviews increases or when prior reviews exhibit higher variance, but they tend to discuss unpopular topics when time distance from the first review increases.