Racial Bias in Hiring

Are Emily and Brendan More Employable than Lakisha and Jamal?

Originally published in Capital Ideas, University of Chicago Newsletter—excerpted from Bertrand, Marianne, and Sendhil Mullainathan. “Are Emily and Greg More Employable Than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination.” The American Economic Review, 2004, 94 (4), pp. 991.

Though racial inequality in the U.S. labor market is understood as a persistent problem even today, it has been difficult to measure how such discrimination works. Do employers actively discriminate against African-American job applicants? Can such discrimination be proven? What is the effect of improved credentials for African-Americans? A recent study offers the answers.

For most job applicants, getting called for an interview is the first major step toward getting a job. But what if the call never comes? Can the name listed on a resume and the perceptions of race implied by this name hinder an applicant’s chances before even getting his or her foot in the door?

In the study “Are Emily and Brendan More Employable than Lakisha and Jamal?” University of Chicago Graduate School of Business professor Marianne Bertrand and Sendhil Mullainathan of Massachusetts Institute of Technology use a field experiment to measure the extent of race-based job discrimination in the current labor market.

From July 2001 to May 2002, Bertrand and Mullainathan sent fictitious resumes in response to 1,300 help-wanted ads listed in the Boston Globe and the Chicago Tribune. They used the callback rate for interviews to measure the success of each resume. Approximately 5,000 resumes were sent for positions in sales, administrative support, clerical services, and customer service. Jobs ranged from a cashier at a store to the manager of sales at a large firm.

The catch was that the authors manipulated the perception of race via the name of each applicant, with comparable credentials for each racial group. Each resume was randomly assigned either a very white-sounding name (Emily Walsh, Brendan Baker) or a very African-American-sounding name (Lakisha Washington, Jamal Jones).

The authors find that applicants with white-sounding names are 50 percent more likely to get called for an initial interview than applicants with African-American-sounding names. Applicants with white names need to send about 10 resumes to get one callback, whereas applicants with African-American names need to send about 15 resumes to achieve the same result.

In addition, race greatly affects how much applicants benefit from having more experience and credentials. White job applicants with higher-quality resumes received 30 percent more callbacks than whites with lower-quality resumes. Having a higher-quality resume has a much smaller impact on African-American applicants, who experienced only 9 percent more callbacks for the same improvement in their credentials. This disparity suggests that in the current state of the labor market, African-Americans may not have strong individual incentives to build better resumes.

“For us, the most surprising and disheartening result is seeing that applicants with African-American names were not rewarded for having better resumes,” says Bertrand.

Statistically, the authors found that discrimination levels were consistent across all the occupations and industries covered in the experiment. Even federal contractors (for whom affirmative action is better enforced) and companies that explicitly state that they are an “Equal Opportunity Employer” did not discriminate less.

Creating the Job Applicants

In order to determine the effect of racially distinctive names on callback potential, Bertrand and Mullainathan needed to generate realistic and representative resumes that also would not thwart actual job seekers. They began with resumes posted at least six months beforehand on two major job search Web sites. For the cities in the experiment, Boston and Chicago, the authors used Boston resumes as templates for the Chicago resumes, replaced the employer names and school names of one city with those of the other, and purged the resumes of real names and contact information.

Within the four occupational categories (sales, administrative support, clerical services, and customer services), the authors classified each occupational category into two groups–high-quality and low-quality–using criteria such as job experience, gaps in employment history, and level of skill. For each high-quality resume, the authors also added extra credentials such as summer employment experience, volunteer experience, more computer skills, certification for administrative positions, special honors, or military experience. E-mail addresses were used almost exclusively for the high-quality resumes. The high- and low-quality resumes differed substantially in terms of skill but did not include a large gap in education level, to avoid making applicants under- or overqualified for a given position.

The result was a pool of distinct but realistic looking resumes, similar in their education, experience, and personal profiles to the potential population of job-seekers.

The choice of both first names and last names was crucial to the experiment. To choose names that were distinctively white or distinctively African-American, the authors referenced all Massachusetts birth certificates from 1974 to 1979, and tabulated the names that appeared relatively most frequently for each racial group, male and female. Frequently used white-sounding names in the study include Anne, Emily, and Allison for women, and Neil, Todd, and Matthew for men. Frequently used African-American-sounding names include Tamika, Latoya, and Latonya for women, and Tyrone, Tremayne, and Rasheed for men.

Applicants in each race/sex/city/resume quality group were assigned the same phone number so that the authors could track employer callbacks in each group, even if they were not able to match callbacks to specific resumes. The resulting bank of names, phone numbers, addresses, and e-mail addresses were then randomly assigned to the template resumes when responding to employment ads.

For each ad, the authors used the bank of resumes to sample four that fit the job description and requirements as closely as possible: two resumes with white names and two with African-American names, and one high-quality and one low-quality resume for each group. They used the voicemail and e-mail messages sent by employers to match the calls to specific resumes and ads.

The Callback Gap

By isolating elements of the resumes in this fashion, the authors can attribute the 50 percent lower callback rate for African-American applicants to name manipulation. While the cost of sending additional resumes might not be large, this gap can be substantial in relation to the rate of new job openings.

The results suggest a significant amount of discrimination in this first stage of the job recruiting process. Furthermore, the study measured how employers responded to improvements in the African-American applicants’ credentials.

The average resume lists eight years of experience. The addition of e-mail addresses, honors, and special skills had a significant effect on the likelihood of white applicants being called, but a statistically insignificant effect for African-American applicants. Employers simply seem to pay less attention or discount the additional characteristics listed on the resumes with African-American sounding names.

“The question may become, ‘Do I really want to invest the time to take an evening class to build my resume?'” notes Bertrand. “The payback that an African-American applicant gets from building these skills is much lower than the payback a white applicant would get.”

Bertrand cautions that employers may infer not just the race of the applicant, but also social class, assuming that certain African-American sounding names are associated with having more underprivileged backgrounds. The results do not delve into the larger issue of hiring rates or earnings gaps, or how African-American applicants might fare using other channels for their job searches.

Though it has been suggested that choosing more race-neutral names is the answer for African-Americans, Bertrand regards such suggestions as ridiculous.

“Names are about identity,” says Bertrand. “We do not advocate changing names to fit the system, and that is certainly not the point of our study.”

Is Awareness Enough?

While the chances for getting the job can change dramatically at the interview stage, the study shows that getting to that stage requires overcoming significant hurdles for African-Americans.

The evidence suggests that discrimination is an important factor in why African-Americans do poorly in the labor market as compared to whites, and indicates one possible reason for the persistence in racial inequality over time.

Though training alone may not be enough to alleviate the barriers raised by discrimination, the study still may be useful for training human resources managers on issues of diversity.

“We’re not claiming that employers engage in discriminatory behavior consciously, or that this is necessarily an issue of racism,” says Bertrand. “It is important to teach people in charge of hiring about the subconscious biases they may have, and figure out a way to change these patterns.”

The nature of the problem also proves to be a dilemma beyond any quick fix.

“I think most African-Americans already realize they need to work much harder than whites to get a job,” says Bertrand. “They will have to send more resumes and fight to get that first job interview.”


This topic is also discussed in an article on the Slate e-zine site in a review of Freakonomics, by Steven D. Leavitt and Stephen J. Dubner: http://slate.com/id/2116449.

Marianne Bertrand is Professor of Economics and Neubauer Family Faculty Fellow at the University of Chicago Graduate School of Business.