Iterative Effect-Size Bias in Ridehailing: Measuring Social Bias in Dynamic Pricing of 100 Million Rides

Cornell University logo
June 08, 2020

In 2015, the dynamic algorithmic pricing used by the Princeton Review was found to be charging Asian-Americans higher prices for their test-preparation services twice as often as non-Asian customers, often regardless of a customer’s income. The use of geographic location in their dynamic pricing model, which adjusted prices based on where customers resided, was determined to be at the root of the change in prices due to ethnicity. While neither illegal nor intentional, the bias exhibited by the Princeton Review’s use of a dynamic pricing model had considerable financial consequences for their customers. As the use of algorithmic dynamic pricing increases, unintended consequences such as racially based price disparities, as in the case of the Princeton Review, must be found and accounted for to ensure the fairness of service pricing.

Read More