Officer Diversity May Reduce Black Americans’ Fear of the Police (Re-Blog)
(Re-posted from Dr. Nix’s website). NOTE: This paper was originally posted as a pre-print on SocArXiv on August 25, 2022. On July 27, 2023, it was accepted for publication by Criminology. It is scheduled to appear in the February 2024 issue. In the meantime, you can download the pre-print using the button above.
Calls to diversify the police date back to at least the 1960s, when President Johnson’s Commission on Law Enforcement and Administration of Justice argued that a lack of racial diversity in police departments signals to Black communities that their neighborhoods are “being policed, not for the purpose of maintaining law and order, but for the purpose of maintaining … [the] status quo” (p. 101). That same year, the Kerner Commission noted that “for police in a [Black] community, to be predominantly white can serve as a dangerous irritant” (p. 165). Drawing on the theory of representative bureaucracy, both commissions recommended intensifying recruitment of Black applicants, fully and visibly integrating Black officers (e.g., not strictly assigning them to patrol primarily Black neighborhoods), and reviewing promotion policies to ensure Black officers are not unfairly prevented from obtaining supervisory or leadership roles. According to the Kerner Commission:
Some cities have adopted a policy of assigning one white and one [Black] officer to patrol cars, especially in ghetto areas. These assignments result in better understanding, tempered judgment and increased ability to separate the truly suspect from the unfamiliar.
Neither Commission was concerned at the time about the near complete absence of women in policing, but a series of lawsuits alleging gender-based discrimination in the 1960s and 70s, alongside Executive Order 11478 and the Crime Control Act of 1973, opened the doors for women to begin integrating into policing (for more, see Archbold & Schulz, 2012).
Though there has been a lot of research done on the effects of police racial and gender diversity, findings have been mixed. Nearly all studies have been non-experimental and cross-sectional, and nearly all of them focus on policing outcomes (e.g., arrests, shootings). Outcomes are often measured at the aggregate level, making it difficult to isolate causal mechanisms and introducing concerns about ecological fallacies. And finally, almost all prior studies are unable to control for patrol assignments or selection bias in administrative data, prohibiting “apples to apples” comparisons (one notable exception is this article by Bocar Ba and colleagues). That said, the weight of the evidence suggests that male and female officers treat people differently. By contrast, the existing evidence is less clear about whether White and non-White officers treat people differently, as well as about the effects of aggregate-level police racial or gender diversity (see pp. 10-13 of our pre-print).
Couple this mixed evidence with continued highly visible examples of police misconduct and violence, such as the beating of Tyre Nichols by five Black police officers in Memphis, and it is understandable that some are skeptical of diversification as a meaningful police reform. But we believe that it can improve outcomes via reducing fear. We tested the following hypotheses:
- Black Americans will be less afraid when officers are non-White (Black or Hispanic/Latino) instead of White.
- Black Americans will be less afraid when officers are female instead of male.
And because Black Americans are the group that most fears and distrusts the police, it is among them that we expected the effects of officer diversity to be largest.
We embedded two experiments in a YouGov survey fielded between April 21 and May 2, 2022, oversampling Black Americans to ensure similarly sized analytic samples of Black and non-Black Americans (N=511 and 589, respectively).1
Experiment 1 was a paired-profile conjoint analysis, where each respondent evaluated five conjoint tables, each containing two officer profiles (varied by officer race and sex, presence of a body-worn camera, officer age, body type, education, and past complaints). Respondents were asked to indicate which officer in each pairing would make them the most afraid of being mistreated.
Experiment 2 used a different design. Each respondent was shown a picture of a two-officer team and asked to indicate their fear of experiencing four forms of mistreatment (e.g., excessive force, wrongful arrest) upon being stopped by those officers (see Figure 1). We randomized the race and sex of each officer as well as the location of the stop (empty vs. busy street).
First, our baseline measures replicate the findings from Pickett and colleagues' 2021 study, demonstrating that the American racial divide in police-related fear was not a temporary artifact of the events of 2020-21, but a persistent social fact in the United States. Black respondents were significantly more fearful than White respondents of police, but were no different from other groups in terms of their fear of crime.
Turning to our experimental results, we found that net of other randomized factors, Black respondents expressed less fear of officers who were Hispanic/Latino or Black (compared to White), and less fearful of female officers (compared to male officers). In Experiment 1, we also found that Black and non-Black respondents were less fearful of officers who were wearing body-worn cameras, and more fearful of officers who were known to have a prior complaint for being disrespectful, using excessive force, or both.
Note: You can right-click the image above and open it in a new window to make it larger.
Every poll and every survey question to our knowledge has shown the same thing - namely, that police-related fear is widespread in Black America. As my colleague points out, if we believe White Americans who say they aren’t afraid of the police, we need to believe Black Americans who say they are. So, what to make of our key findings?
The key policy implication of our finding is that we should continue increasing the representation of Black Americans and women in policing, as doing so may reduce public fear and by extension, improve police-citizen interactions. This might be achieved legislatively, via earmarked funding, or organizationally, via diversity initiatives. The National 30x30 Initiative, for example, aims to increase the representation of women in policing from 12% to 30% by 2030. To date, over 300 agencies have signed the pledge, which requires them “to report on their efforts to identify and address the obstacles that women officers face in recruitment and throughout their careers.” In less than two years, the first police chief to sign the pledge nearly quadrupled the number of women employed by his agency - from 4 to 15 - after revising or eliminating antiquated policies that he believed discouraged women and people of color from applying. In other cases, it has taken lawsuits to motivate police departments to diversify. Hopefully, however, the supporting empirical evidence herein and in other recent work (Riccucci et al., 2018; Peyton et al., 2022) can further stimulate self-directed diversification efforts within police departments.
From the standpoint of symbolic representation, and for signaling normative alignment with civilians, it would also be beneficial for police agencies to communicate clearly to the public (e.g., via official social media posts) that they are committed to racial and gender diversification and are taking specific steps to increase the diversity of their officers.
And finally, our results suggest that requiring officers to wear (and notify citizens of) body-worn cameras might reduce fear in the community. There seems to be a lot of disagreement about whether body-worn cameras have “lived up to the hype.” I think that stems in part from a lack of clarity about what we expected from them 10+ years ago. But in any event, there is definitely evidence that in the right circumstances, they can improve citizen perceptions of police (as well as reduce complaints, reduce use of force incidents, and perhaps even save agencies money in the long run).
For more about the sample see pp. 13-14 of the pre-print as well as the supplemental appendix. ↩︎