What’s in a name? Businesses are reforming hiring practices

September 11, 2017 —

Thanks to David Cameron, name-blind hiring is all the rage in the UK.

“You can give someone all the opportunities in the world… but it’s no good if they’re prevented from getting on because of their gender, race, religion, sexuality or disability,” the prime minister wrote in a 2015 op-ed in The Guardian. “We have managed to get some of the biggest graduate employers to pledge to anonymise their job applications – in other words, make them name-blind.”

The beauty of Cameron’s proposition is that he’s basically asking businesses to do what is in their own self-interest. He provided the political cover for companies to modernize and catch up with three decades of research in behavioral economics. As a result, several major corporations and at least four universities in the UK have adopted name-blind applicant screening.

Name-blind applications aren’t complicated. They include everything you’d expect from an application, except for the applicant’s name.

There is substantial evidence for name-based bias in applications, largely because it’s easy to test. Simply draft identical resumes, slap a different name on each one – say, Jamaal on one, Brian on the other – submit the resumes to public job postings and see who gets a call back.

Unsurprisingly, the phone rings quickly for male and white names, more slowly for women and minorities.

“Job applicants with white names needed to send about 10 resumes to get one callback,” concluded one study by the National Bureau of Economic Research. “Those with African-American names needed to send around 15 resumes to get one callback.”

Similar studies have demonstrated the edge that men enjoy over women.

This research establishes that women and minorities have to clear a higher hurdle to get interviews than do white men. But unlike the ease of identifying the problem, the root cause is more nuanced and complicated.

One possibility is that the people who make screening decisions are sexist and racist. They don’t value diversity and discriminate against underrepresented constituencies.

Another interpretation is that they are well-intentioned but flawed human beings, like the rest of us.

A 2010 study in France showed the complexity of the problem. Name-blind applications failed to help ethnic minorities get interviews, and may even have hurt them.

In this case, name-blind applications may have been counter-productive for companies that already followed effective diversity policies, leading to less-diverse outcomes. A second theory is that, in the absence of a name, employers looked for indicators such as proficiency in Arabic to weed out immigrant applicants.

Disqualifying applicants because of Arabic fluency involves second-order thinking, revealing explicit bias against immigrants. Unfortunately, it’s not realistic to protect against such explicitly held prejudices. If an employer considers race, gender or ethnicity to be qualifying attributes for a job, reforms based on blind applications are not the answer. This enters the realm of policy and legal intervention.

While this remains a real problem, a more immediate challenge for our business community is to help well-intentioned corporations and hiring managers protect themselves from their own flaws.

In his celebrated book, Thinking Fast and Slow, Nobel Laureate Daniel Kahneman discusses an internal conflict in how we make decisions. He summarizes how involuntary instincts, or “fast thinking,” can be at odds with our deliberate, logical, “slow” thoughts.

Instincts are useful when making snap judgments about personal safety. They are less useful when deciding between job applicants. Unfortunately, instincts are subject to biological or cultural biases, such as deference to tall men or willingness to trust physically attractive people.

Even when we consciously disavow or try to suppress their influence, instincts prove powerful and difficult to change.

It’s this internal conflict that leads someone with deeply held feminist convictions to unknowingly favor male applicants. This doesn’t make the screener a bad person. It just means he is flawed, like the rest of us.

Recognizing this vulnerability, many companies are automating the process of applicant screening to remove human bias. However, algorithms have limitations.

In a 2016 HBR article, Gideon Mann and Cathy O’Neil warn that machine learning solutions can perpetuate historical hiring biases, instead of correcting them.

Startups are also wading into the fight. GapJumpers has gained notoriety by administering applicant tests on behalf of hiring companies. Employers receive anonymous test results and use these to screen applicants. This approach is useful to remove bias and assess individual competencies, but is less effective identifying how well a person works with a team, or if she is an effective manager.

Businesses are reforming hiring practices; HBS should, too

A bit closer to home, HBS publishes enrollment breakdowns by gender, ethnicity, and country of origin, while appealing to additional constituencies such as LGBT and former military. In this way, public accountability provides its own discipline to the admissions process. Because HBS values diversity in the case-method classroom, completely anonymizing applications would risk removing a valuable dimension from consideration, rather than eliminating discrimination.

However, HBS also acts on the “supply side” of the talent equation, allowing employers to recruit MBA students for internships and full-time roles. In this position, HBS can still affect hiring outcomes.

A simple reform would be to publish two resume books, one in the traditional style and one name-blind, with a glossary that ties names to page numbers. This would make it low-cost for companies to eliminate bias when they recruit from HBS.

A more involved reform would be to “modularize” our resume information online. Instead of downloading entire resumes, a company could choose only those features it believes are relevant, and prevent resume reviewers from seeing details that may poison the well.

The possibilities are broader than simply omitting names. Companies could decide that undergraduate major and GPA are relevant, for example, but not undergraduate institution to prevent bias against international students or for only “elite” applicants. They might omit “interests” to prevent favoring like-minded applicants, or languages to prevent ethnic identification.

These solutions are not without cost. HBS will have to invest in IT, staff resources, and time spent on communication with both students and employers. However, HBS’ administration, which has demonstrated a commitment to sensitive and equitable treatment of women and minorities, needs to see this as the next logical step in furthering that commitment.

For students, reforming our resume and application system would ensure that the most qualified, desiring students are not denied access to jobs based on their race or gender. For employers, it would eliminate “noise” from the hiring process, improving the quality of new hires. This, in turn, enhances HBS’ brand by improving the perception of its graduates.

It would also enforce that we are an institution that is not content just talking about problems, but solves them.

Robert Carpenter (HBS ’18) worked in strategy consulting in Houston, Texas, before starting at Harvard. He previously attended Texas A&M University, where he was editor in chief of the student publication, The Battalion.