Measure and Mitigate the Dimensional Bias in Online Reviews and Ratings

Online word-of-mouth in the form of online reviews and ratings is an increasingly important resource for consumers to acquire product information for their purchase decision. However, dimensional review bias, originated from consumer heterogeneity and their multidimensional product preferences and experiences, have been shown to undermine the information transfer among consumers. Through a novel text mining approach, we identify and quantify two types of dimensional biases from textual reviews: dimensional preference bias and dimensional rating bias. We also introduce a quantitative method to mitigate the dimensional rating bias. We examined the effectiveness and applicability of our bias measures and de-bias method in the context of multi-dimensional and single-dimensional rating systems. Specifically, we focused on the hotel reviews from TripAdvisor.com and Expedia.com. Our preliminary results show promising theoretical and managerial contributions.