What Makes Consumer Reviews Trustworthy?

Keywords: Consumer Reviews, Variable Selection, Trust in Online Communities

Abstract

Consumer reviews are ubiquitous in online shopping. They help consumers access product qualities and make purchase decisions. However, consumers are also complaining that consumer reviews are likely to be untrustworthy. While they read reviews, they can vote for a trustworthy review by clicking the 'Helpful' button. Using the number of votes as the proxy for review trustworthiness, this paper develops an ensemble variable selection method to evaluate what characteristics of the reviews form the trustworthiness. Based on the analysis using sample reviews on Amazon, the paper highlights that consumers' trust formation for reviews is likely conditional on the review ratings. The method and result could potentially improve Amazon and other e-commerce websites' algorithm for recommending most trustworthy reviews for consumers.