After reports surfaced that drivers discriminated against certain groups, ridesharing platforms such as Uber and Lyft removed information that could indicate a rider’s gender and race from initial ride requests. But a new study, led by a researcher at the Kelley School of Business at Indiana University, shows that biases against underrepresented groups continued to exist after drivers accepted a ride request — when the rider’s picture would then be displayed.
In other words, the ridesharing platforms’ efforts shifted some of the biased behavior until after the ride was confirmed. This resulted in higher cancellation rates.
The researchers performed a field experiment on a ridesharing platform in fall 2018 in Washington, D.C. They randomly manipulated rider names, using those traditionally perceived to be White or Black, as well as profile pictures to observe drivers’ behavior patterns in accepting and canceling rides. They found that 8 percent of customers from unrepresented racial groups had their rides canceled compared to only 3 percent of White customers.
“Our results confirm that bias at the ride request stage has been removed. However, after ride acceptance, racial biases are persistent, while we found no evidence of gender biases,” said Jorge Mejia, assistant professor of operations and decision technologies at the business school and lead author of the study.
The full study, “When Transparency Fails: Bias and Financial Incentives in Ridesharing Platforms,” was published on the website of the journal Management Science. The research was co-authored by Chris Parker, assistant professor in the information technology and analytics department at American University in Washington, D.C. The study may be accessed here.