More Progress, More Bias?
How racial progress relates to racial biases.
Posted Jan 29, 2020
Racial biases in the U.S. have improved dramatically since the darkest days of our history, but at this point, many feel that the reduction in racial biases has plateaued, or perhaps even begun creeping back up. What might explain this observation?
Previous research has shown that in states that are more racially diverse, White people tend to show higher levels of racial bias. In other studies, researchers have looked at the impact of exposing people to information, such as Census projections, indicating impending increases in racial and ethnic diversity in the U.S. This kind of research allows us to determine the causal impact of telling White people that racial minority group members will soon make up more than half of the U.S. population. These studies show that this information causes racial biases and support for conservative policies to increase. Similar increases in racial bias are observed when White Americans are reminded of the racial significance of the election of Barack Obama, and the fact that increasing societal power is being afforded to those who are not White.
Taken together, this evidence suggests that when progress is made among minority groups in the U.S., it can heighten racial biases among White Americans. Thus, when I heard that WalletHub had used data from the U.S. Census Bureau, the National Center for Education Statistics, the U.S. Equal Employment Opportunity Commission, the Bureau of Justice Statistics, and the Centers for Disease Control and Prevention to calculate a metric of racial progress across U.S. states, I was interested. Specifically, this metric was designed to quantify changes in racial disparities—such that states with the highest “progress” scores had shown the largest reduction in racial disparities over time. Thus, states with higher racial progress scores are not necessarily the most racially “progressive," but they are the states that have shown the greatest changes in that direction (toward reducing racial disparities). The racial disparities that went into this metric included income, labor-force participation, unemployment rates, home ownership, poverty rates, high school completion, Bachelor’s degrees, standardized test scores, voter turnout, and infant mortality rates, among others.
The question I wanted to answer was whether there was any relation between racial progress and White Americans’ racial biases. Based on the research that I mentioned above, I predicted that in states where racial progress was greatest, racial biases among White people would tend to be higher. Fortunately, Project Implicit collects data on the attitudes and biases of people all over the world, and just so happens to have annual estimates of the racial biases of people of various races across all 50 states and the District of Columbia. To test my question, I combined the racial progress scores calculated by WalletHub with the data on state-level racial biases (for 2018) from Project Implicit. Using this approach, I was able to examine how racial progress in the U.S. relates to the racial attitudes of both White people and Black people.
First, I conducted simple correlation analyses to examine whether there was any association between racial progress and the racial biases of White respondents. As you can see in the figure below, the results provided evidence of the pattern I expected (r = .27). White respondents in states that had seen the biggest reduction in Black-White racial disparities, such as Mississippi and Wyoming, showed the highest levels of implicit pro-White/anti-Black bias. In contrast, White respondents in states that showed the least racial progress (the smallest reduction in racial disparities over time), such as Maine and Washington DC, tended to show the least racial bias.
Next, I conducted another correlation analysis; this time, to see whether there was any relation between racial progress and the racial attitudes of Black respondents. In this case, as can be seen in the figure below, Black respondents in states that had seen more racial progress showed more pro-Black attitudes (r = -.32). In states that had seen very little racial progress, Black respondents showed a tendency for pro-White bias (implicitly favoring White people over Black people).
So, what does this all mean? Taken together, these analyses suggest that racial progress may have positive and negative implications for the racial attitudes of U.S. residents. On the one hand, racial progress may be threatening—resulting in increased anti-Black biases among White people. But, on the other hand, it looks like racial progress may have positive implications for the attitudes and esteem of Black people in the U.S. In societies which grant more power and privilege to people of some races over others, it is not uncommon for members of the disadvantaged group to develop negative attitudes toward their own racial group or favoritism for higher status groups. The U.S. is no exception in this regard; a tendency to favor White people has even been observed among Black children; in fact, this concerning tendency contributed to the historic Brown vs. Board of Education decision. The correlations I have reported here provide some suggestive evidence that racial progress may be relevant to addressing this. To the extent that this tendency to favor White people over Black people still exists among Black Americans, it appears to largely be in the states that have seen the least racial progress.
Many questions remain unanswered, and WalletHub does not publicly provide complete details of the formula they use for calculating racial progress—thus, I cannot vouch for the quality of this metric. My hope is that this simple analysis inspires others to take on this question in a more rigorous way—calculating their own metrics of racial progress and testing how this stacks up against other known predictors of racial bias, such as racial diversity.
I would like to thank Ellen Zhang, Itohan Aigbekaen, and Sydney Phillips of the Georgia Attitude Bias and Behavior Lab for their assistance compiling and checking the data for these analyses, and for helpful feedback on earlier drafts of this post.