Body-Worn Cameras and Transparency: Experimental Evidence of Inconsistency in Police Executive Decision-Making

Abstract

Body-worn cameras (BWC) have diffused rapidly throughout policing as a means of promoting transparency and accountability. Yet, whether to release BWC footage to the public remains largely up to the discretion of police executives, and we know little about how they interpret and respond to BWC footage – particularly footage involving critical incidents. We asked a nationally representative sample of police executives (N = 476) how supportive they were of legislation that would mandate releasing BWC footage upon request as public information, and presented them with an experimental vignette about BWC capturing one of their officers fatally shooting an [armed/unarmed] [Black/White] suspect. Results indicated inconsistency in executives’ attitudes and decision-making: (1) less than one-third of executives supported such legislation, (2) suspect race and armed/unarmed status shaped how executives felt media would cover the incident and whether they would state publicly that the shooting was justified, and (3) agency size conditioned the effects of armed/unarmed status on executives’ perceptions.

Keywords:

• Body-worn cameras
• transparency
• experiment
• media
• officer-involved shootings

Disclosure statement

No potential conflict of interest was reported by the authors.

Acknowledgment

We appreciate Brad Campbell’s help in conceiving the original idea for this project, and in developing an earlier version of the vignette, which was administered to a different sample in 2016.

Notes

1 See http://www.ncsl.org/research/civil-and-criminal-justice/body-worn-cameras-interactive-graphic.aspx#/.

2 See also “Weekly Crime Reports” at https://www.atlantapd.org/i-want-to/crime-data-downloads. In the first 12 weeks of 2020, Atlanta police officers averaged more than 3,300 traffic stops per week. Even while residents were ordered to stay at home and businesses were closed due to COVID-19, officers never made fewer than 1,000 stops in a week. Yet, from 6/14/2020 to 6/20/2020 – the week immediately following the shooting of Rayshard Brooks – they made just 50. Arrests similarly plunged following the Brooks shooting.

3 We described a home invasion in the vignette after consulting with police officers during pilot testing. The officers noted that a home invasion in progress would likely involve dispatch relaying information from the 911 caller, and that officers would approach it with a sense of urgency due to the rather uncommon nature of the call.

4 Generally speaking, small agencies serve less populous jurisdictions than large agencies. The “100 or more officers” cutoff is common in police surveys (O’Shea & Nicholls, 2003). We use it for the largest stratum because of concerns about statistical power. Just 45 agencies report having 1,000 or more officers, and only another 154 agencies have more than 250 officers (Hyland & Davis, 2019). This means we would have needed to survey every agency with 250+ officers and achieved a 60% or better response rate in order to have sufficient power to detect medium-sized effects.

5 We removed two ineligible respondents (who did not work for a municipal agency) and eight respondents who failed to provide enough information to determine the stratum from which they were sampled. For the purposes of our analyses, we restricted the sample to Chiefs (including Interim and Acting Chiefs), which reduced the sample size to N = 476.

6 Surprisingly, national data on officer-involved shootings, disaggregated by the agencies involved and/or their respective populations, are hard to come by. Nix et al. (2017, p. 323) reported that 68% of fatal shootings in 2015 involved departments with fewer than 1,000 full-time officers. Analyzing seven months of shootings reported by The Washington Post, Sherman (2018) found that the majority of fatal police shootings – 51% – occurred in communities with fewer than 50,000 residents, and 70% occurred outside of major cities. Similarly, according to Mapping Police Violence, 74% of police-involved fatalities from 2013 to 2019 occurred outside of the 100 largest U.S. cities (see https://mappingpoliceviolence.org/cities). Finally, Edwards et al. (2018) showed that two-thirds of all police killings happen in suburbs, smaller cities, and rural counties.

7 Response rates did not vary as a function of these manipulations. The response rates for each treatment condition were as follows: White suspect 27.8%, Black suspect 26.12%; armed suspect 26.84%, unarmed suspect 27.08%. Our experiment achieved good balance—respondent characteristics (gender, race/ethnicity, and education) and agency use of body-worn cameras were very similar across experimental groups. Additionally, the sample mirrors the population of local police chiefs in terms of race/ethnicity (91% White) and gender (96% male; see Hyland & Davis, 2019). See Tables A1-A3 in the supplemental appendices.

8 We also performed each factor analysis separately for each stratum. In each stratum, the items used to create the anticipated media coverage and additional communication scales loaded onto a single factor and exhibited adequate internal consistency. We also re-estimated our regression models using weighted factor scores instead of mean scales for these two outcomes. Results were substantively similar (see Tables B1 and C1 in the supplemental appendices).

9 We estimated supplemental models for the ordinal outcome using ordered logistic regression. The results were substantively similar (see Table D1 in the supplemental appendices).

10 When testing interactions in the full sample, where the four strata are combined, we do not use this Bonferroni corrected alpha.

11 Model fit statistics (adjusted R², AIC) support the decision in the interaction models to treat agency size as ordinal for anticipated media and as categorical for state justified.

12 See https://www.bwcscorecard.org/static/policies/2016-08-17%20Las%20Vegas%20-%20BWC%20Video%20Request%20Policy.pdf.

Additional information

Notes on contributors

Brandon Tregle

Brandon Tregle, J.D., is a doctoral student in the School of Criminology and Criminal Justice at the University of Nebraska Omaha. His research interests include body-worn cameras, legislation, and criminal law. His recent work appears in Police Quarterly and Journal of Crime and Justice.

Justin Nix

Justin Nix, Ph.D., is an associate professor in the School of Criminology and Criminal Justice at the University of Nebraska Omaha. His research interests include police legitimacy and police use of deadly force. His recent work appears in Journal of Research in Crime and Delinquency, Journal of Experimental Criminology, and Criminology & Public Policy.

Justin T. Pickett

Justin Pickett, Ph.D., is an associate professor in the School of Criminal Justice at SUNY Albany. His research interests include police-community relations, public opinion, survey research methods, and theories of punishment. His recent work appears in Criminology, Journal of Quantitative Criminology, and Annual Review of Criminology.