Our choices as employees may lead to workplace segregation

Business people working on laptops during a meeting

There is widespread ethno-racial segregation between workplaces in the US, even within the same job sector. Research suggests that as many as 50 percent of all African American and Hispanic workers would have to change jobs to achieve integrated workplaces.

Employers contribute to this segregation through hiring decisions, but what is the role of the choices that we, as employees, make when looking for a workplace? People tend to live, work and socialize within their own ethno-racial group. This has both positives and negatives.

On the positive side, a segregated city may offer a safe zone for individuals belonging to discriminated minorities. In a community with similar others, spaces can be created where the larger society’s racial hierarchy does not apply. On the negative side, the problem is that in societies where there is a lot of ethno-racial inequality, segregation tends to reinforce these differences.

In segregated societies, minorities tend to live in impoverished areas with high crime rates and poor public and commercial services. We know from previous research that ethno-racial inequality is also connected to occupational segregation: children of poor families tend to aim for blue-collar jobs, while children of wealthy families tend to aim for high-level white-collar jobs.

What has been less known, is what causes segregation between workplaces in the same sector. The aim of our study was to find another piece of the puzzle.

Workplace segregation does not concern people’s choice of profession. The puzzle of workplace segregation is that it somehow channels people within the same profession to different firms.

And just like with other types of segregation, this leads to an additional disadvantage for minorities. Minorities somehow end up in the firms with the lowest pay and the worst working conditions. In short, minorities tend to draw the short straw when it comes to effects of segregation. So how do workplaces become segregated?

First, an ethno-racially unequal and segregated society provides people with different pathways into the labor market. Second, when some employers discriminate against minorities and some do not, the discriminated groups will end up at some workplaces and the non-discriminated groups at others. Third, employees might jointly maintain workplace segregation by their choice of employer.

Even though employees cannot always choose where to work, they have at least some freedom to do so. In particular, they can influence workplace segregation by their tenure at a workplace: by staying longer in segregated workplaces and being employed for a shorter period in integrated workplaces.

Previous research has shown that the first and second explanations are indeed descriptions of what happens on the labor market. In our study, we explore the basis for the third explanation: the role of employees. Employee choice is typically underpinned by some type of attitude or preference for something. So, we have asked: to what extent could workplace segregation be the result of employee attitudes and preferences?

We addressed this question by measuring general feelings for having colleagues of different ethno-racial identities, and if they could choose, who would they want as a colleague? Once we knew people’s ethno-racial preferences, we put those numbers into a simulation model to estimate the potential consequences of these preferences for workplace segregation. Can the choices of employees influence workplace segregation along race-ethnic lines, and are they then contributing to segregation or holding it back?

Drawing on results from an online survey of 1,435 American respondents, we found, perhaps not surprisingly, that people do prefer their own ethno-racial group as colleagues. This was especially true of European Americans. Being given a secondary choice, the respondents preferred the outgroup with the highest labor market status: European Americans and Asians.

Thus, people prefer colleagues who are like themselves, and thereafter people who have a high societal status. Overall, female respondents were more positive about diverse workplaces than men. Another noteworthy finding was that the respondents generally preferred minority women as colleagues over minority men.

Many people expressed fairly positive attitudes toward ethno-racial diversity. Even though many European Americans did prefer homogeneous workplaces, it was much more common among them to prefer mixed workplaces, where there were almost as many employees from minority groups as there were European Americans.

There is, however, a clear limit at 50%. Almost no one of the European Americans wanted to be in the minority at their workplace. And, if they could not get a position at a company with an even mix, they would almost always opt for a company with less diversity, rather than one with more.

This brings us to the famous residential segregation model by Thomas Schelling. His model illustrates that even if people were to prefer to live in a diverse neighborhood, they will still end up in segregated neighborhoods if they cannot, at the same time, tolerate being in too small a minority themselves.

We designed and implemented a model based on a similar idea, but for workplace segregation. In this model, people have a ranking of preferences. Their first company of choice is the one with a degree of employee diversity that best matches their ranking. If the company is full, they opt for the second best choice, and so on. We ran the model using data from our study on actual preference rankings.

Recall that even though our respondents expressed fairly positive attitudes toward diversity and wanted less segregation than is the case between workplaces today, at the same time they avoided highly diverse companies. In our simulation model, this meant that primary choices for less segregation were not enough to desegregate workplaces, since secondary segregating choices came into effect.

When based on the most common preferences among our respondents (that is, removing those with more individually unique preferences), the simulations even suggested that these preferences can cause and increase segregation.

It is possible that we have not captured entirely representative preferences of the American population. If there is a bias, however, we believe that our results overestimate positive attitudes and preferences toward diversity, rather than underestimate them because of social desirability bias in these types of studies.

That is, people tend to display more positive attitudes and preferences concerning diversity than they actually have, in part because people tend to adjust their attitudes and preferences to what they perceive as the purpose of the study. Even so, these attitudes and preferences we found were not positive enough to contribute to the desegregation of workplaces.

From earlier studies, we know that employers tend to favor majority group applicants in hiring decisions. A take-home message from this and our results is that desegregation will not occur through a laissez-faire approach where employers and employees freely choose their employees and their colleagues.