The sociologist's research found significant problems with a prior study that claimed to show the disappearance of hiring discrimination.
Raj Ghoshal, assistant professor of sociology, has published new research. His article “Flawed Measurement of Hiring Discrimination against African Americans,” that appears in the Fall 2019 issue of the peer-reviewed journal Sociation suggests that discrimination against African Americans persists.
The article addresses a debate over whether employers still illegally discriminate when making hiring decisions. Many experiments in the past 15 years have tested for discrimination by creating two fake identities with equal-quality resumes and applying to the same set of job listings with both identities.
In these experiments, known as audit studies, “Steven Smith” and “DeShawn Jackson” might apply to the same 1,000 jobs. Researchers then measure how much interest each resume generates. This line of research has generally found that black Americans need to send out significantly more applications than white Americans to get the same number of callbacks.
However, a 2016 study by economists at the University of Missouri-Columbia argued that these studies used overly stereotypical names to signal race in ways that exaggerated their results. The economists conducted their own study using what they considered more realistic names, and found no difference in employers’ response rates by race. But Ghoshal’s work finds significant flaws in the methodology the economists used.
The 2016 study had used names like “Chloe Jackson” and “Ryan Washington” for their African American job candidates because the last names Jackson and Washington typically belong to black individuals, while “Chloe” and “Ryan” were purportedly race-neutral. Though the economists are correct to see Washington and Jackson as typically black last names, Ghoshal hypothesized that very few Americans would know this and interpret the names as intended. He therefore conducted a 1,050-person national survey which asked respondents to guess the race of people with the exact names the 2016 study had used.
Survey findings show that about 60 percent of people do not interpret the economists’ study names as intended, and frequently see the names as belonging to white individuals. Further, those individuals most likely to make hiring decisions make just as many errors as others. The level of error is sufficient that the 2016 study is not merely invalidated. Rather, its results, properly interpreted, suggest that black job applicants need to send out about 50 percent more applications to have an equal chance of response.
Overall, the findings suggest that racial discrimination remains an important concern that individuals, employers, and government should address.
The research was supported by Elon’s Faculty Research & Development.