Though rideshare apps were intended to be a more egalitarian transportation option than traditional cab services, they have a demonstrated history of bias against minority passengers before pickup. Numerous studies have established that black passengers, for instance, faced longer wait times and were subject to more frequent cancellations than white riders.
In recent years, platforms responded to reports of the bias by removing information about riders’ gender and race from the ride requests presented to drivers. Yet, a recent study shows that these adjustments have not diminished bias across the board and that black riders, riders perceived to be lesbian, gay, bisexual, transgender or queer, and riders who are perceived to be LGBTQ supporters are canceled on more frequently before pickup than white and perceptibly straight passengers.
“By removing the ability to see information before the drivers accepted a ride request, the hope was that all of the bias we were observing would cease to exist,” Chris Parker, an assistant professor of information technology and analytics at American University and the co-author of the study, “When Transparency Fails: Bias and Financial Incentives in Ridesharing Platforms,” told NBC News. “But after the change was instituted, we suspected that there’s still the problem of some drivers not wanting to pick up certain passengers.”
In order to test their hypothesis, Parker and co-author Jorge Meija, an assistant professor of operations and decision technologies at Indiana University, created an account with multiple rideshare profiles. Though each passenger had a 4.8 rating (out of a 5.0 maximum), the profiles differed in passenger names and pictures. The researchers used four pictures to create the profiles: one for a black woman named Keisha or Latoya, one for a black male named Rasheed or Jamal, one for a white female named Emily or Allison and one for a white male named Brad or Greg. At least 92 percent of people perceive Emily, Allison, Brad and Greg as Caucasian names and Keisha, Latoya, Rasheed and Jamal to be black names, according to a 2004 study, which is why the researchers used these names.
Parker and Meija also added a rainbow filter on some of the profiles — using the same names and pictures — to denote that a potential passenger was LGBTQ or supported the LGBTQ community.
From early October to mid-November of last year, the researchers ordered rides from a major and unnamed ride-sharing platform in Washington, D.C., at a fixed central Metro stop. Throughout the experiment, they altered the gender, race, rainbow filer and timing of the ride requests. After calling each ride, the researchers waited three minutes to allow a driver to cancel if they no longer wanted to accept a ride — enough time for them to determine the riders’ race, gender and perceived sexuality or LGBTQ support. If the driver has not canceled at this point, Parker and Meija would cancel the rides themselves so that the driver could receive compensation in the form of a cancellation fee.
After 3,200 observations during both peak and nonpeak hours of the day, the researchers determined that while there was no significant evidence of bias against women riders compared to male riders of both races, both black men and black women were nearly three times more likely to be canceled on than their white peers. Their tests also showed that signaling support for the LGBTQ community with a rainbow filter resulted in the passenger nearly doubling their chance of being canceled on across races.
“We know that LGBTQ riders face discrimination with these rideshare apps, but we thought that it was an interesting little twist, that even just signaling your support for the LGBTQ community could result in a canceled ride,” Parker said.
The research did not show that riders who were both black and signaled support for the LGBTQ community faced greater cancellations.
One especially instructive aspect of the research, according to Parker, was that black riders were less likely to get canceled on during peak hours — likely because drivers view the heightened cost of such rush-hour trips as increased incentive. Yet signaling support for the LGBTQ community with the rainbow filter resulted in similar cancellation rates during peak and nonpeak hours.
Based on this evidence, Parker said it appears that providing information about customer demographics may enable drivers’ biases. He recommends platforms and policymakers reflect on the type and timing of information they give to rideshare drivers and suggests that platforms might consider penalizing drivers by moving them down the priority list when they exhibit biased cancellation behavior or rewarding them when they exhibit especially low cancellation rates for minority riders.
“There’s a lot of next-step actions platforms might consider to ensure a good outcome and that everybody has a safe, comfortable, noncombative ride,” Parker said.
Though this study looks at bias that occurs before an LGBTQ passenger or ally enters a car, instances of discrimination in the backseat of rideshares have been well-documented.
In December 2018, a gay couple in Houston said they were kicked out of an Uber ride after sharing a quick kiss. Then in May of this year, a gay couple in Indianapolis claimed they were booted from a Lyft ride for sharing a “short kiss on the lips.” In June, a gay man said a New York City taxi driver told him, “I don’t drive gays,” and that same month, a lesbian couple claimed they were thrown out of an Uber ride for sharing a “peck.” In perhaps the most frightening incident, a gay man last November claimed a Lyft driver in Miami pulled a gun on him after exclaiming, “I want to kill everyone that’s gay.”
And the discrimination is not only limited to riders; LGBTQ drivers face bias as well. Earlier this month, a transgender woman quit driving for Lyft after an intoxicated passenger allegedly hit her head and told her she was “nothing but a man.”