Researchers performed a ‘huge litmus test’ for racial discrimination among the U.S. labor market by sending functional applications that had either White or Black names to job ads.

These job application forms were sent in pairs with almost identical qualifications, just their names differed. After performing statistical analysis, scientists found evidence for systematic racial discrimination, with Black applications receiving significantly fewer callbacks from prospective employers than White applications.

The economists at the University of California, Berkeley, and the University of Chicago, conducted the largest discrimination audit so far, sending over 83,000 fake job applications for entry-level positions at 108 companies. The vast majority of these companies were in the top 100 of the Fortunate 500, so the study covers some of the nation’s biggest employers, who by now should have robust anti-discrimination hiring procedures.

While the applicants’ characteristics, such as age, sexual orientation, work experience, and education were randomized, the names were chosen on purpose to ensure the applications came in pairs. One application had a distinctively White name — like Connor or Amy — while the other, very similar application sent to the same job listing had a distinctively Black name — like DeShawn or Aaliyah. Here’s a resume example:

The findings were perhaps not all that surprising. On average, applicants with distinctively Black names were about 10% less likely to be contacted than job seekers with White names.

However, the discrimination was not uniform among the companies that received the fictitious resumes. For many, there was no distinguishable pattern of discrimination beyond random noise, which suggests they’ve done a good job at laying the groundwork for a fair employment process.

That can’t be said about 23 companies that exhibited a significant bias towards selecting applications with distinctively White names despite the fact that they had equally qualified applicants with Black names to choose from. These companies were not explicitly called out in the researchers’ report, but their identities may have been shared with the U.S. Labor Department, which was briefed on the general findings.

The top 20% of firms from the study accounted for roughly 50% of the total discrimination. For every 1,000 applications, White candidates received around 250 callbacks, compared to 230 for Black job seekers. But among one-fifth of companies, this gap grew to 50 callbacks on average. In many instances, it seems, the most impact when writing your resume may be in what name it’s penned under rather than the actual background.

Researchers also found a slight pattern of gender discrimination, but it wasn’t what you probably thought of. Overall, male and female resumes were equally likely to be contacted on average. But when they broke it down, the researchers found that some firms strongly prefer male names, while others strongly prefer female names, revealing a pattern of discrimination in both directions.

There was no sign of geographical discrimination, in that job applications in the South, for instance, fare no worse than anywhere else. Instead, race discrimination seems to cluster in certain industries, particularly auto, retail, and food. Discrimination against Black applications was also more prevalent in companies that are less profitable, are FED contractors, and have a decentralized hiring system.

These latter findings suggest that a more uniform and robust human resources policy may be important in controlling bias and reducing discrimination in the labor market, according to Evan Rose, an economist at the University of Chicago’s Becker Friedman Institute and one of the authors of the new study.

What’s more, concerning the gender discrimination that was also analyzed in the study, men are favored in auto/repair services, building materials, and communications while women are favored in apparel and furnishing stores, health services, and food products.

This isn’t the first study that creatively employed fake resumes to expose racial discrimination in the hiring process. As the New York Times reported, a 2003 study responded with fictitious resumes signed by “very African American sounding name or a very White sounding name” to help-wanted ads in Boston and Chicago newspapers. They found White names received 50% more callbacks for interviews, on average. A more recent similar study from 2014 employed 9,000 resumes from fictitious, recently-graduated job seekers, finding that Black applications receive approximately 14% fewer interview requests than their otherwise identical White counterparts.