![Sample our Social Sciences journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/110822/TOC-Subject-Sample-2021-Social-Sciences.jpg"})
Abstract
We use citizen complaint data from the Chicago Police Department and Civilian Office of Police Accountability filed between 2013-2020 to determine the extent to which Body-worn camera (BWC) footage enhances the efficacy of evidence used to formulate a conclusion of responsibility, and whether racial disparities in investigation outcomes would subsequently be reduced. Accordingly, we exploit the staggered deployment of BWCs across 22 Chicago police districts over time to estimate the effect of BWCs on these outcomes. Findings indicate BWCs led to a significant decrease in the dismissal of investigations due to insufficient evidence (“not sustained”) and a significant increase in disciplinary actions against police officers (“sustained” “outcomes”) with sufficient evidence to sanction their misconduct. We further find that disparities in complaints across racial groups for the “unsustained” category faded away with the implementation of BWCs. The article concludes with theoretical and research implications based on the findings.
Acknowledgements
We wish to acknowledge researchers at the Chicago Crime Lab, Zoe Russek-Sobol and Emma Nechamkin for guiding our understanding of the data employed in this paper.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 African Americans in particular, are skeptical of routine policing practices and feel racially profiled by them (Warren, 2011; Horrace & Rohlin, Citation2016). Several studies also found that African Americans are disproportionately targeted in stop-and-frisk practices (Meares, Citation2009; Lotan, Citation2020; NYCLU, 2012, page 5). For instance, African Americans are three to five times more likely than Whites to rate the occurrence of police misconduct as “very often” than Whites (Weitzer & Tuch, Citation2004).
2 The deadly shooting of Michael Brown in 2014 heightened public expectations of police accountability and both the public and the federal government urged the police in the United States to widely adopt BWCs (Bowling & Iyer, 2019; Culhane et al., Citation2016). Subsequent police shootings of unarmed Black men have further intensified such expectations.
3 Many would argue in fact that the historical precursor for the expectation that visual footage could serve as a medium for exposing police misconduct were the “Rodney King” tapes, depicting the beating of King by members of the LAPD shot by videographer George Holliday from his balcony in 1991 (see Granot et al., Citation2018; Stuart, Citation2011). The tapes received extensive airplay in media outlets and engendered expectations of convictions for the officers involved. When they were acquitted, Los Angeles and other cities in the US experienced rioting and protests.
4 This is in line with explanations for claims of a “Ferguson Effect,” where crime rates were assumed to have increased as a result of police tamping down their enforcement engagement (“de-policing”) out of concerns that they would be held responsible for exercising discretion during encounters with the public, and an overall sense that their efforts as societal guardians were being disparaged by constituencies that did not understand the day-to-day rigors and stresses of police work (see, e.g. Nix & Wolfe, Citation2018) combined with residents refraining from reporting crimes and cooperating with police due to increased mistrust (“de-legitimizing”).
5 For instance, in their news article covering the New York City Police Department’s (NYPD) unwillingness to share BWC footage with the Civilian Complaint Review Board (CCRB), Umansky and Simon (Citation2020) discussed how the NYPD withheld video evidence and paper records from the CCRB in at least a couple of incidents despite the law requiring the NYPD’s cooperation with the agency. They also noted that the percentage of sustained cases doubled when the investigators had access to BWC footage.
6 Notably, Barak and colleagues (2019) raised concerns about treatment contamination in this research due to the involvement of more than one officer in the use of force incidents or other incidents that led to complaints. They suggested that the RCT’s ability to capture the true effect was diminished because of the spillover effect.
8 Kalven v. The City of Chicago, 2014 Ill. App. (1st) 121846.
9 https://invisible.institute/police-data.
10 The data includes information about the final findings of the investigation and information on the pending cases and cases that ended with “no findings.”
12 They also find that police officers' race matters in investigation outcomes; complaints against White officers are less likely to be sustained than Black police officers, suggesting a systemic issue.
14 The BWC implementation dates for each district is given in Table A1. We obtained the CPD’s BWC implementation timeline for the 22 police districts from OIG File #18-0103 (City of Chicago OIG, 2019).
15 CPD members must inform the person(s) that they are activating their BWC and recording their interactions. They cannot deactivate their BWC until the assignment is cleared, the police officer leaves the scene, or the arrestee is secured and transferred to processing. Also, a CPD officer may deactivate BWC recording if requested by a victim, witness of crime, or a community member who wishes to report a crime. A police officer can also deactivate the camera when they interact with a confidential informant. In such circumstances, the officer must justify their deactivation with a verbal statement on the BWC before turning the camera off (pages. 28-29).
17 There are also 730 complaints filed against police officers who are allegedly involved in domestic violence. These cases are excluded from our analysis as our focus is the effect of BWCs on investigation outcomes for citizen complaints against police and it is unlikely that police officers use BWC during their interactions with and violence against their intimate partners. In a similar vein, another 691 complaints occurred while the complainant was in custody are also excluded from the analysis since detention and jails are continuously monitored with mounted in-door cameras regardless of a BWC.
18 However, we present the results for the “Exonerated” and “Unfounded” outcomes in Table A2. Note that given the small sample size for the “Exonerated” outcome, the non-parametric model for this outcome is not reliable. As shown in Table A2, there appears to be no relationship between BWC and either of these outcomes as the estimates are neither statistically significant nor sizeable.
19 Our results are not sensitive to controlling for this variable.
20 Goodman-Bacon (Citation2021) develops an alternative method of detrending, in which he suggests estimating separate pre-treatment trends based on the timing of treatment. Then these pre-treatment trends are projected onto the post-treatment period and used to detrend the outcome variable. Our results are robust to using this alternative method of accounting for pre-trends. These results are available upon request.
21 For example, the estimate on BWC policy is -0.007 (p = 0.845) in the model for the likelihood of the complainant being White, − 0.018 (p = 0.641) for the likelihood of the complainant being Black, and -0.022 (p = 0.851) for the likelihood of complainant being Hispanic.
22 Our conversations with the COPA officials indicated that the average duration of time from a misconduct incident and filing of a complaint is 11 days. To assess the sensitivity of our results to this misalignment in some of our observations, we re-estimated our models excluding the observations within 10, 20, 30, and 45 days of the introduction of BWCs. These results are very similar to those presented in Table 2.
23 It is not straightforward to interpret the interaction terms in the non-parametric model specified by equation (2). To keep it tractable, we only consider an augmented version of equation (1) in which we interact BWC with race and ethnicity of the complainants.