How Uber hires drivers where background checks are impossible

Los Angeles, 2019, is now your computer, 2015.
Los Angeles, 2019, is now your computer, 2015.
Image: Warner Bros
We may earn a commission from links on this page.

A month after an Uber driver in Delhi was charged with the rape of a female passenger, Uber beefed up its background checks for prospective drivers in India.

This is unlikely to help. Outside of the developed West, few countries have centralized criminal records databases. Nor are certificates of good character worth much. And clean background checks are easy to obtain for the right sum of money passed under the right table.

As an alternative, companies that rely on local labor are turning to a new method—one that Uber itself uses in places like Mexico, and to great effect. In a country where, as Britain’s Home Office warns travelers, passengers “have been robbed and assaulted by unlicensed taxi drivers,” Uber has built a reputation for safety.

The technique it uses is called psychometrics, and it’s gaining ground in hiring, lending, and even marketing.

Psychometrics 101

Psychometric tests are somewhat like aptitude tests, in that they consist of a series of questions and quizzes—not about the matter at hand (in this case, a job) but about more general things.They are designed to figure out what sort of person you are and how likely you are to exhibit certain behaviors, which might tie into your propensity to succeed at a job, pay back a loan, or purchase a specific item.

According to the British Psychological Society, they are called psychometric tests “because psychological theories of human behaviour and its measurement have been used in their construction.” Think of it as sort of a real world Voight-Kampff test.

How Uber does it

Uber uses two types of standardized psychometrics tests in Mexico to gauge how prone somebody is to violence, criminal association, robbery, and substance abuse, says Juan Jose Fernandez Gallardo, who runs safety and operations for the company in Latin America.

The test results help create a shortlist of applicants who are interviewed by psychologists. Finally, all applicants still in the running go through a drug test.

Acceptance rates vary by city—and by the specific circumstances of the city. “In Guadalajara everyone knows somebody who knows somebody in a drug cartel,” says Gallardo. In cities like that, where organized crime has made deep in-roads, acceptance rates tend to be lower, perhaps around 60% of applicants. In places like Mexico City, acceptance rates tend to be higher, around 75%. Uber now uses psychometric testing in six cities in Mexico as well as in cities in Peru and Chile, with Brazil to join soon.

The pspread of psychometrics

Psychometric testing is increasingly seen as the answer to a lack of good data in poor countries. In Peru, three banks—BanBif, Banco Financiero, and Caja Trujillo—have all started using psychometric testing to help figure out whom they should lend to.

According to Daniel Schydlowsky, Peru’s chief financial regulator, psychometric testing will be the next big thing for lending, especially in countries where credit scoring is difficult, often for the same reasons that background checks are. Already one Peruvian company is combing through lists of bad debtors, applying psychometrics, and going to banks to let them know which of those customers are actually good bets, he says.

In London, a company called VisualDNA figures it can use psychometrics for marketing. It came up with this formulation:

Big data + psychology + digital identity + internet of things = better understanding ( = a huge opportunity)

“A black swan just swam into view,” the company’s head of planning wrote in a marketing document.

That may not just be a marketer’s excitement talking. According to a paper by academics from Cambridge and Stanford, published in Proceedings in the National Academy of Science, “Computer-based personality judgments are more accurate than those made by humans.” Indeed, the authors don’t beat around the bush; that’s the title of the paper. It basically builds a case for the type of testing Uber and others are using in Mexico City. Says Michal Kosinski, one of the authors of the paper, “We can now take psychology out of labs and put [it] on the streets.”