Race-based discrimination is common in the hiring process.

For example, racial minorities are less likely than whites to receive a callback when they apply for a job. There are also wide earning gaps, with African-Americans and Latinos earning a fraction of what whites and Asians do.

Yet despite laws that aim to reduce employment discrimination and improve attitudes toward diversity, these patterns have not changed for decades.

When analyzing these problems, researchers and others tend to focus on how the experiences of racial minorities compare with those of whites. Often missing is whether there are differences among individuals of the same racial group in terms of how they experience bias.

That is where my new study, which focuses on perceptions of others’ racial identities, comes in.

Perceived identities

People have more than one identity, such as being a mom, a Muslim, an athlete, a scientist and so on.

Just as we commonly think about the importance of each of our identities to who we are – such as being a dad or very religious – we make the same assessments of other people. That is, we evaluate other people’s identities to understand which ones are most fundamental to who they are.

And it turns out, the conclusions we come to about each other’s “perceived identities” can have a big effect on how we interact with them.

As a researcher who has spent the last 19 years examining diversity and inclusion, I was interested in how perceptions of identity affected a racial minority’s prospects as a job applicant. More specifically, I wanted to know if the perception that an applicant has a strong racial identity affected her ability to get a job and how much she’d get paid.

Presumed identity

Past research has shown that our inferences about others’ personal identities can influence how we interact with them.

In some cases, people might talk about how their identity is important to them, or how it reflects a critical part of who they are as a person. In other cases, we make assessments based on cues. For example, we might think someone strongly identifies as Latino when they are members of a Latino student organization. Or, we might infer a weak identity among people who engage in actions that are seemingly contrary to the interests of their group.

For example, psychologists Cheryl Kaiser and Jennifer Pratt-Hyatt found found that whites interact more positively with racial minorities they believe weakly identify with their race – and more negatively with those with stronger racial identifies. Specifically, whites expressed more desire to be their friends and offer favorable ratings of their personality.

Presumed identity and work

Drawing on their work, Astin Vick, a former student of mine, and I examined whether African-American women’s and Latinas’ presumed racial identity affect their job ratings.

Using an online data collection platform, we asked 238 white people who indicated that they currently or previously worked in the fitness industry to review the application of someone applying to be a club manager. They were told to review a job description, a hiring directive from the club owner, a summary of each applicant’s relevant background and a picture.

All applicants had the same experience, work history and education. The pictures were used to indicate an applicant’s race. Most importantly, we varied each applicant’s relevant affiliations and community service to suggest whether she had a strong identification to her racial group or a weak one.

For example, membership in the Latino Fitness Instructors Association or volunteering for former President Barack Obama’s campaign would signal a strong identification to an applicant’s Latina or black racial group. Belonging to the neutral-sounding Intercollegiate Athletics Coaches Association or volunteering for Obama’s opponent in the 2012 presidential campaign, Mitt Romney, would signal a weak one.

The participants then filled in a questionnaire to measure their perceptions of the applicant they reviewed, including work attributes such as “untested” or “expert,” hiring recommendation and suggested salary.

Our results showed that most people did in fact use cues from the application file to form views of the applicant’s racial identity, which in turn informed their hiring and salary recommendations. Essentially, as we expected, applicants perceived as identifying strongly with their racial group were less likely to be recommended for a job. And, when they were, received lower suggested salaries – on average US$2,000 less – than those signaling weak associations.

The story does not end there, though, since we also knew each participant’s gender. And we found that men showed a slightly different pattern than the one described above.

Men recommended roughly the same salaries for African-American women and Latinas who identified weakly with their racial groups. But for those with strong perceived identifies, they penalized Latinas far more than African-Americans. That is, they recommended the club pay Latinas with a strong racial identify about $5,000 less than African-Americans.

These small changes can add up over time. Over a 15-year tenure with a company, that difference results in $96,489 difference in inflation-adjusted earnings.

The impact

Our study illustrates several key points.

First, though racial minorities, as a collective, face bias in employment, there is considerable within group variability. An applicant’s specific race matters, as does her or his presumed racial identity.

Second, raters use cues on a resume to infer a job applicant’s racial identity. They then use this information in their decision-making. Aware of this pattern, some job seekers remove race-related activities on their resumes, what Sonia Kang, an associate professor of organizational behavior, refers to as racial whitening.

Finally, research has shown that diversity in the workplace leads to greater organizational performance and employee well-being. As such, employers would be wise to be on the lookout for biases like the one we found that are likely to lead to less diverse workforces and take steps to overcome them when hiring new workers.

George B. Cunningham does not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and has disclosed no relevant affiliations beyond their academic appointment.