Performance of online reputation mechanisms under the influence of different types of biases

Rating mechanisms are becoming an essential tool for online shoppers. There have been many reputation mechanisms proposed in the literature that try to mitigate the sometimes adverse effects of strategic behavior of the traders to alter their true reputation scores. At the same time, several studies have shown the presence of bias (often unintentional) in product ratings and their harmful effect on buyers and product sales. To date there has not been any work done that studies the performance of reputation mechanisms in the presence of various types of biases. This study is a first attempt to fill that gap. We present three reputation mechanism models that have been discussed in the literature, discuss three types of biases that have been studied in the past, and study the performance of the three reputation mechanisms in the presence of these biases by using a multi-agent system that simulates the interactions among buyers and sellers in an electronic marketplace. Using replicator dynamics we also show the effect of these models on the behavior of buyers and sellers as the buyers choose the type of reputation models they prefer to use and the sellers change their strategic behavior in response.