Morris George
Abstract With the exponential growth in the number of reviews for products and services, it is important to identify reviews that are helpful in making purchase decisions. Prior studies have identified several factors influencing helpfulness of a review. However, the impact that negativity and positivity of review comments have on review helpfulness especially of extreme (1-star and 5-star) reviews has not received much attention. This study focuses on assessing the impact of negativity and positivity of extreme review ratings on review helpfulness by building a zero-One Inflated Beta regression model using a data set from Amazon having product reviews from several categories. We find that negativity (positivity) of review text in an extremely positive (negative) review influences helpfulness of the review positively. The study also finds support for the negativity in review text of an extremely positive review to be more helpful compared to positivity in an extremely negative review.