Weighting Ratings: Are People Adjusting for Bias in Extreme Reviews?

37 Pages Posted: 19 Oct 2022

Date Written: October 12, 2022

Abstract

Consumer ratings distributions have increased in importance due to the ever-increasing proportion of transactions that are made online. This raises the question of whether consumers are adjusting for potentially fake or biased reviews when evaluating products from ratings. Two online experiments tested treatments that manipulated how the extremes of rating distributions (i.e. 1-star and 5-star) were presented. Neither removing extreme reviews while preserving the mean rating, flagging suspicious extreme reviews in the distribution, nor priming individuals about review manipulation significantly affects perceived product quality on average (when the mean rating is accounted for). Furthermore, people appear to apply different weights to ratings of different scores to evaluate product quality. The estimated weighting profile is hump-shaped and asymmetric, implying that consumers are particularly susceptible to being misled by disingenuous 1-star reviews. These results may indicate a need to change how summary rating statistics are conveyed in order to prevent undesirable or wasteful purchases.

Keywords: rating distributions, fake reviews, consumer behaviour, weighted mean

Suggested Citation

Ocean, Neel, Weighting Ratings: Are People Adjusting for Bias in Extreme Reviews? (October 12, 2022). Available at SSRN: https://ssrn.com/abstract=4245795 or http://dx.doi.org/10.2139/ssrn.4245795

Neel Ocean (Contact Author)

University of Warwick

Gibbet Hill Rd.
Coventry, West Midlands CV4 8UW
United Kingdom

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
17
Abstract Views
77
PlumX Metrics