Demeanor, Race, and Police Perceptions of Procedural Justice: Evidence from Two Randomized Experiments

Abstract

President Obama’s Task Force on 21st Century Policing recently endorsed procedural justice as a way to restore trust between police and communities. Yet police–citizen interactions vary immensely, and research has yet to give sufficient consideration to the factors that might affect the importance officers place on exercising procedural justice during interactions. Building on research examining “moral worthiness” judgments and racial stereotyping among police officers, we conducted two randomized experiments to test whether suspect race and demeanor affect officers’ perceptions of the threat of violence and importance of exercising procedural justice while interacting with suspicious persons. We find that suspect race fails to exert a statistically significant effect on either outcome. However, demeanor does—such that officers perceive a greater threat of violence and indicate it is less important to exercise procedural justice with disrespectful suspects. These findings have implications for procedural justice training, specifically, and police–community relations more broadly.

Keywords:

• policing
• procedural justice
• race
• demeanor
• moral worthiness

Disclosure Statement

No potential conflict of interest was reported by the authors.

Notes

¹ With respect to ethnicity, however, the evidence is not so clear. For example, Sargeant, Murphy, and Cherney (2014) found that procedural justice was less associated with trust in the police among Vietnamese individuals than the general population. On the other hand, a more recent study suggests procedural justice has a larger effect on legitimacy among ethnic minorities who feel disengaged from police (Madon, Murphy, & Sargeant, in press).

² Interestingly, these studies were completed prior to the August 2014 shooting of Michael Brown in Ferguson, Missouri, which led to unprecedented levels of scrutiny regarding police treatment of minority citizens (Nix & Wolfe, 2016, 2017).

³ Note that Tyler’s framework is intended to complement instrumental-based perspectives like deterrence, not supplant them. In fact, recent research has argued that perceived police effectiveness is simply a component of people’s police legitimacy evaluations (Tankebe, 2013; see also, Beetham, 1991).

⁴ Other research adds to this conclusion. Procedural justice is important when measured both directly (i.e. stemming from recent police contact) and vicariously (Tyler & Huo, 2002; Wolfe, McLean, & Pratt, in press), for the general population and among offenders (Reisig, Wolfe, & Holtfreter, 2011; White, Mulvey, & Dario, 2016), and within various cultural contexts outside of the United States (Jonathan-Zamir & Weisburd, 2013; Murphy & Mazerolle, in press; Murphy, Mazerolle, & Bennett, 2014; Tankebe, 2008).

⁵ Prior to data collection, we performed a power analysis whereby we determined we would need a sample size of approximately N = 200 in order to have 80% power to detect a medium-sized effect at the .05 alpha level (Cohen, 1992).

⁶ The suspect in each vignette was a male, given that according to the 2015 Uniform Crime Report, males accounted for 73% of all arrests and 80% of all arrests for violent crimes.

⁷ Consistent with the best practice in survey research (Dillman, Smyth, & Christian, 2014), prior to finalizing the vignettes, we pilot-tested them with 36 police officers who were attending a class at an advanced police education and training institute in the southeastern United States. We asked the officers to read over the vignettes and report how realistic they seemed on a scale of 1 (not at all realistic) to 100 (completely realistic). On average, the officers scored the vignettes 92 out of 100, indicating that they seemed very realistic. We also asked the officers to provide written feedback about how we could make the vignettes more realistic, and used their feedback to improve the vignettes accordingly prior to administering the survey. It is a common strategy to employ experts, in this case police officers, in methods construction (see e.g. Polit & Beck, 2006).

⁸ We thank an anonymous reviewer for suggesting we calculate standardized mean-difference effect sizes (or Cohen’s d) for each of the experimental manipulations. To do so, we used Wilson’s Practical Meta-Analysis Effect Size Calculator (available at https://www.campbellcollaboration.org/escalc/html/EffectSizeCalculator-SMD-main.php). Cohen (1992) provides the following guidelines for interpreting the magnitude of d: .2 is considered a “small” effect, .5 is considered a “medium” effect, and .8 is considered a “large” effect.

⁹ In a supplemental analysis, we re-ran this model using a weighted factor score instead of a mean index for the dependent variable. The results (available upon request) were substantively identical.

¹⁰ We suspect the difference in response rates between the two surveys is attributable to several factors, including the nature of the two police agencies, type of promotion, and stated survey length. Specifically, the lower response rate was achieved from a much larger and more urban agency, where officers had received a greater number of prior survey requests from other researchers, and where an Assistant Chief (rather than Deputy Chief) helped promote the survey. The survey used for Study 2 also had a longer stated length, as noted on the introductory page.

¹¹ Because indirect effects can exist even in the absence of significant zero-order relationships between variables, we tested for mediation using bias-corrected bootstrap (k = 1,000) confidence intervals (95%; see Hayes, 2013). None of the experimental manipulations had statistically significant indirect effects (Hypotheses #4 and #6 not supported; results available upon request).

¹² We re-ran Models 1 and 2 using a weighted factor score instead of a mean index for the dependent variable. The results for Model 1 were substantively identical to those reported in-text. In Model 2, the only difference was that the p-value for symbolic disrespect increased slightly from .046 to .072. These results are available upon request.

Additional information

Notes on contributors

Justin Nix

Justin Nix is an assistant professor in the School of Criminology and Criminal Justice at the University of Nebraska at Omaha. He received his PhD in Criminology and Criminal Justice from the University of South Carolina in 2015. His research interests include procedural justice, police legitimacy, and police use of force.

Justin T. Pickett

Justin T. Pickett is an associate professor in the School of Criminal Justice at the State University of New York at Albany. His research interests include survey research methods and public opinion about criminal justice.

Scott E. Wolfe

Scott E. Wolfe is an associate professor in the School of Criminal Justice at Michigan State University. His research focuses on organizational justice, procedural fairness, and police–citizen interactions.

Bradley A. Campbell

Bradley A. Campbell is an assistant professor in the Department of Criminal Justice at the University of Louisville. He received his PhD in Criminal Justice from Sam Houston State University in 2015. His research focuses on police investigations, decision-making, and responses to victims.