Reading 11.2 A Quasi-Experimental Evaluation of the Effects of Police Body-Worn Cameras (BWCs) on Response-to-Resistance in a Large Metropolitan Police Department

Police departments have adopted body-worn cameras (BWCs) to both foster officer accountability and provide officers protection against false or exaggerated claims of brutality. Jennings, Fridell, Lynch, Jetelina, and Reingle Gonzalez (2017) studied whether officers from the Tampa, Florida, police department who used BWCs would differ from other officers in the same city in their frequency of engaging in physical response-to-resistance. A classical experiment was not possible since officers were given the opportunity to volunteer to wear BWCs. Flad the researchers compared the BWC volunteers to a random sample of officers who did not volunteer, there would be a question of whether any differences found between the two groups was a product of the BWCs or due to fundamental differences in the temperament and job performances of the officers. To compensate for the lack of randomization, the authors used propensity scores to pair the BWC officers with a matched group of police not wearing cameras during the same time period. As you will see, the results of this study would have been quite different had the researchers not accounted for the differences between the two groups with a matching procedure.

A Quasi-Experimental Evaluation of the Effects of Police Body- Worn Cameras (BWCs) on Response-to-Resistance in a Large Metropolitan Police Department

Wesley G. Jennings, Lorie A. Fridell, Mathew Lynch, Katelyn K. Jetelina, and Jennifer M. Reingle Gonzalez

Introduction

Given the recent and growing number of tragic incidents such as Michael Brown in Ferguson, Missouri; Eric Garner in New York; Alton Sterling in Baton Rouge, Louisiana; and Philando Castile in St. Paul, Minnesota, where minority individuals are being victims of violence perpetrated by the police, as well as the recent shooting and killing of five police officers in Dallas, Texas, it is clearly a time when people are searching for answers and explanations and justice system reform and prevention strategies. Coinciding with these events, and often in response to these events, public and police interest in police body-worn cameras (BWCs) is emerging. In fact, Lum and colleagues (2015) recently identified 12 existing BWC empirical studies (Ariel, Farrar, and Sutherland 2015; Ellis, Jenkins, and Smith 2015; Goodall 2007; Grossmith et al. 2015; Jennings, Fridell, and Lynch 2014; Jennings, Lynch, and Fridell 2015; Katzet al. 2015; ODS Consulting 2011; Owens, Mann, and Mckenna 2014; Ready and Young 2015; Roy 2014; Young and Ready 2015) and another 30 ongoing empirical studies (for review of the ongoing BWC studies, see Lum et al. 2015).

Relatedly, a number of recent studies have shown that officers are generally supportive of BWCs, and that while their initial perceptions may be less optimistic about the benefits of their use, their favorability ratings toward BWCs generally improves once they become more acquainted with and accustomed to the technology (Jennings et al. 2014, 2015; Katz et al. 2015). For example, Young and Ready (2015) reported that Mesa, Arizona, officers who wore BWCs were more likely to rate BWCs as being helpful in their day-to-day activities, and Gros- smith et al.’s (2015) analysis of officers in London revealed that the officers reported BWCs to be useful in handling complaints more effectively, particularly in cases where evidence would have been lacking in the absence of BWCs. Katz et al. (2015) also indicated that the majority of surveyed officers perceived BWCs as a tool to provide a more accurate depiction of events, that the BWC video increases the quality of the evidence that can be obtained, and that nearly 60% of the officers felt that BWCs were comfortable to wear, but were also hoping for some technological improvements for added comfort as well. In addition, emerging research has provided some preliminary evidence that public attitudes generally favor the use of BWCs in police departments and perceive that BWCs will increase the quality of police-citizen interactions. However, some citizens have also noted their concerns about victim and witness cooperation (Sousa, Miethe, and Sakiyama 2015).

Turning toward BWC outcome evaluations specifically, Young and Ready (2015) relied on quasi-experimental field data from Mesa, Arizona, where 50 officers were assigned BWCs and 50 officers represented a matched comparison group who did not wear BWCs. Their results indicated that the BWC officers performed significantly fewer stop and frisks, issued significantly more citations, initiated significantly more contacts with citizens, and rated the BWCs as being particularly helpful in their encounters with citizens. Ariel, Farrar, and Sutherland (2015) conducted a randomized controlled trial of BWCs in Rialto, California, where 54 Rialto police officers (the entire police department) were randomly assigned to shifts where they either wore or did not wear BWCs. Their results suggested that response-to- resistance was twice as likely during shifts where the BWC was not worn relative to the shifts where the BWC was worn. In contrast, Jennings et al. (2015) performed a randomized controlled trial where officers (not shifts) were randomized in Orlando, Florida. The results from this study revealed that the 46 officers who were randomly assigned to wear BWCs had significantly fewer incidents of response-to-resistance in the 12 months post-BWC implementation compared with the 46 officers who were randomly assigned to not wear BWCs, and that wearing BWCs led to a 53.4% reduction in the response- to-resistance for officers who were wearing the BWCs when comparing their response-to- resistance in the 12 months pre-BWC implementation to their response-to-resistance in the 12 months post-BWC implementation.

The Current Study

In recognition of the growing but small amount of existing evaluation research surrounding BWCs, the current study offers an examination of the impact of BWCs on police response-to- resistance. Specifically, relying on data from 60 BWC officers from the Tampa Police Department in Tampa, Florida, and using a propensity score matching (PSM) technique to identify' a statistically comparable matched sample of 60 non- BWC officers, this study evaluates the effect of BWCs on the frequency of response-to- resistance in the 12 months post-BWC implementation and compares this frequency with the frequency of response-to-resistance in the 12 months prior to BWC implementation.

Methods

Site and Sample

Tampa, Florida, is a mid-to-large-size metropolitan city with greater than 350,000 residents, and the city is fairly diverse, with roughly 40% of the population being minorities (African American or Hispanic). The sample for the current study includes 761 Tampa Police

Department (TPD) officers that were working for the TPD during the 24-month study period.

Dependent Variable

The dependent variable is a frequency (count) of the number of incidents of physical response - to-resistance that the officers were involved in during the 12 months pre-BWC implementation (March 2014-February 2015) and in the 12 months post-BWC implementation (March 2015-February 2016). Specifically, physical response-to-resistance incidents included behaviors such as empty hand “soft” control techniques (i.e., grabs, holds, and joint locks), empty hand “hard” control techniques (i.e., punches, kicks, countermeasures), less-than- lethal weapons (i.e., batons, projectiles, chemical sprays, conducted energy devices), and firearms.

Independent Variable

Sixty BWCs were purchased by TPD according to a normal bidding process with a host of vendors, and volunteers were recruited from the three police districts in Tampa (e.g., approximately 20 officers per district) to wear the BWCs. As such, the officers were separated into two groups of interest: BWC officers and non- BWC officers.

Covariates

There are a series of officer demographics included as covariates in the analysis, including officer gender (male = 1; female = 0), officer race/ethnicity (White Officer, African American Officer, Hispanic Officer), officer age (measured continuously in years), and officer years of law enforcement experience at TPD (measured continuously in years based on the number of years between TPD hire date and study end date). In addition, the frequency (count) of the number of incidents of physical response-to- resistance that the officers were involved in during the 12 months pre-BWC implementation is included as a covariate in certain analyses as well.

Analytic Procedure

This analysis proceeds in a series of stages. In the first stage, we present descriptive summary statistics for the officers. Next, we employ PSM techniques to approximate a quasi-experimen- tal research design in order to be better positioned to derive more precise estimates between statistically matched samples (see Jennings et al. 2013; Richards et al. 2014; Rosenbaum and Rubin 1983). Specifically, the utilization of PSM permits the ability to remove any observable and systematic differences that may exist between the officers who were wearing a BWC relative to the officers that were not wearing a BWC. In this vein, relying on an R program (Ho et al. 2007, 2011) and statistical convention in the propensity score matching literature including the use of a strict .10 caliper and nearest neighbor matching methods (Austin 2009; Rosenbaum 2002), we derive our propensity score estimates from a logistic regression model where wearing a BWC is considered the “experimental condition,” and not wearing a BWC is considered the “control condition” (for further discussion on PSM, see Ho et al. 2011; Pearl 2009; Rosenbaum 2002). This specified propensity score algorithm is also based on the inclusion of a series of officer demographic factors (gender, race/ethnicity, age, and years of experience) and the frequency of response-to- resistance incidents that the officers were involved in during the 12 months pre-BWC implementation. Following the estimation of PSM, we make statistical comparisons via t-tests to compare the demographic factors and the frequency of response-to-resistance pre-BWC implementation between the BWC officers and the non-BWC officers prior to and post-PSM. In the final stage of analysis, we perform statistical comparisons between the BWC officers and the matched sample of non-BWC officers via a t-test for the frequency of response-to-resistance in the 12 months post-BWC implementation and through an examination of percentage changes in the frequency of response-to- resistance pre- to post-BWC implementation for the BWC officers and the matched sample of non-BWC officers.

Results

Summary statistics are presented in Table 1. As illustrated, the majority of officers were male (83.8%), and 71.8% were White, 12.8% were African American, and 15.4% were Hispanic. On average, the officers were 40.11 years of age (SD = 7.83; range = 23-63 years of age), and had been working as police officers at TPD for 11.81 years, on average (SD = 6.62; range = 2-31 years). Overall, the officers were involved in an average of 3.05 incidents where physical force was used during the 12 months pre-BWC implementation, and in an average of 2.73 incidents where physical force was used during the 12 months post-BWC implementation.

Table 2 provides the bivariate comparisons for the covariates (officer gender, officer race/ ethnicity, officer age, years of experience, and the frequency of incidents where physical force was used during the 12 months pre-BWC implementation) before and after applying PSM techniques. As can be seen in columns 1 and 2, prior to matching, most of the covariates significantly differed when comparing the 60 BWC officers with the 701 non-BWC officers. Specifically, a significantly lower percentage of BWC officers were White and a significantly greater percentage of the BWC officers were Hispanic compared with their non-BWC counterparts. Also, the non-BWC officers, on average, were significantly older and had a greater number of years of law enforcement experience at TPD relative to the BWC officers. Finally, the BWC officers had been involved in a significantly greater mean number of incidents where physical response-to-resistance was used in the 12 months pre-BWC implementation compared to the non-BWC officers. Turning toward columns 3 and 4, the results demonstrated that after applying PSM and generating a statistically comparable sample of 60 non-BWC officers in terms of officer demographics and the frequency of physical response-to-resistance pre-BWC implementation that none of the pre-existing significant differences between the two groups remained.

Table 1 Summary Statistics (n = 761)

M/%

SD

Minimum

Maximum

Officer gender Male

83.8

Female Office race/ ethnicity

16.2

White

71.8

-

-

-

African American

12.8

Hispanic

15.4

Officer age

40.11

7.83

23.00

63.00

Number of years as a TPD officer

11.81

6.62

2.00

31.00

Response-

to-resistance

frequency:

Pre-BWC

implementation

3.05

3.69

0.00

29.00

Response-

to-resistance

frequency:

Post-BWC

implementation

2.73

3.78

0.00

30.00

Table 2 Bivariate Comparisons of Officer Demographics and Response-to-Resistance Frequency Before and After Matching

Before

Matching

After

Matching

n = 60

n— 701

n =60

n -60

Variables

BWC

officer

Non-

BWC

officer

BWC

officer

Non-

BWC

officer

Male officer

.80

.84

.80

.74

White officer

.57

.73

.57

.53

African

American officer

.18

.12

.18

.15

Hispanic officer

.25

.15

.25

.31

Officer age

38.37

40.26

39.23

38.37

Number of years as a TPD officer

10.13

11.95

10.13

11.70

Response-

to-resistance

frequency:

Pre-BWC

implementation

3.72

2.99

3.72

4.31

Note: Significant mean differences (p < .10) in bold italics.

Pre- and Post-BWC Implementation Comparisons in the Frequency of Response-to-Resistance for Matched Experimental and Control Groups (N = 120)

Figure 1 Pre- and Post-BWC Implementation Comparisons in the Frequency of Response-to-Resistance for Matched Experimental and Control Groups (N = 120)

Percentage Change in the Frequency of Response-to-Resistance from Pre- to Post-BWC Implementation for Matched Experimental and Control Groups (N = 120)

Figure 2 Percentage Change in the Frequency of Response-to-Resistance from Pre- to Post-BWC Implementation for Matched Experimental and Control Groups (N = 120)

Or in other words, prior to matching, the comparisons between BWC officers and non-BWC officers were in effect apples to oranges comparisons, and, after matching, the comparisons were now apples to apples comparisons.

Figure 1 graphically displays the average number of physical response-to-resistance incidents that the 60 BWC officers and the 60 non-BWC matched officers were involved in during the 12 months pre-BWC implementation and in the 12 months post-BWC implementation. As illustrated, the BWC officers were involved in an average of 3.72 physical response-to-resistance incidents in the 12 months prior to BWC implementation, and this involvement declined to an average of 3.41 physical response-to-resistance incidents in the 12 months post-BWC implementation. Comparatively, the non-BWC matched officers were involved in 4.31 physical response-to- resistance incidents pre-BWC implementation and 4.46 physical response-to-resistance incidents post-BWC implementation. Statistical results from a t-test comparison revealed that the post-BWC physical response-to-resistance incidents significantly differed and were significantly less for the BWC officers relative to the non-BWC matched officers. Finally, the percentage changes in the mean frequency of physical response-to-resistance incidents from pre-BWC implementation to post-BWC implementation are displayed in Figure 2. This graph demonstrates that there was an 8.4% decrease in the average number of physical response-to- resistance incidents pre- to post-BWC implementation for the 60 BWC officers compared with a 3.4% increase in the average number of physical response-to-resistance incidents pre- to post-BWC implementation for the matched sample of 60 non-BWC officers.

Discussion

The current study set out to provide an evaluation of the impact of BWCs on police response- to-resistance. Relying on 24 months of response-to-resistance data (12 months pre- BWC implementation and 12 months post- BWC implementation) and PSM techniques, a number of key findings emerged that are highlighted below.

For example, the initial comparisons revealed that the 60 BWC officers significantly differed from the 701 non-BWC officers not only in the extremely unequal sample sizes but also, more importantly, on a host of officer demographics and in their average number of physical response - to-resistance incidents pre-BWC implementation. As a result, propensity score matching techniques were applied in an effort to generate a matched sample of 60 non-BWC officers that were statistically similar to the 60 BWC officers in terms of their gender, race/ethnicity, age, years of law enforcement experience at TPD, and their average number of physical response-to-resis- tance incidents pre-BWC implementation. Once this PSM algorithm and analysis were performed, the results demonstrated that the 60 BWC officers were involved in a significantly fewer number of physical response-to-resistance incidents post-BWC implementation, on average, compared to the matched sample of 60 non-BWC officers. And the BWC officers exhibited an 8.4% decrease in their mean number of physical response-to-resistance incidents when comparing their behavior in the 12 months pre- BWC implementation to their behavior in the 12 months post-BWC implementation. In contrast, the mean number of physical response-to- resistance incidents increased by 3.4% for the matched sample of non-BWC officers when comparing their pre-BWC implementation frequency to their post-BWC implementation frequency. Although only speculative, if this 8.4% reduction in the mean number of physical response-to-resistance incidents (a reduction of approximately 20 physical response-to-resistance incidents) observed among these 60 BWC officers in the 12 months post-BWC implementation was scaled up to the entire sample of 761 TPD officers, then this would translate to slightly more than a reduction of 250 physical response- to-resistance incidents per year for the department.

These findings notwithstanding, it is important to note several limitations of the current study. First, this study is based on results from one large metropolitan police department in one city, and a city that is a mid- to large-size city with a diverse demographic population. The degree to which these findings would be observed in smaller, less diverse, and more rural police departments and cities may be limited. Second, although a rigorous statistical procedure (i.e., propensity score matching) was used to derive a statistically comparable matched sample of non- BWC officers, these results were not generated from a randomized controlled trial such as those found elsewhere in the literature (see Jennings et al. 2014, 2015). As such, caution should be taken when making comparisons to these studies and in the interpretation of this study’s results as the influence of unobserved and unmeasured differences between the BWC officers and the matched sample of non-BWC officers is unknown. Third, this study (similar to other studies in the literature) did not have information from the incidents where response-to- resistance was used regarding suspect demeanor, suspect race, suspect gender, and so on. As these variables are also potentially relevant factors for influencing an officer’s decision to use force, it is recommended that future research incorporate these data when available.

Ultimately, the results from this study contribute to the growing body of evidence in support of the utility of police body-worn cameras to reduce police response-to-resistance (Ariel et al. 2015; Grossmith et al. 2015; Jennings et al. 2015). However, it is important to acknowledge that BWCs are not a panacea. In light of the growing number of tragic incidents that are occurring in the United States where minority individuals and police officers are being targeted for violence, it is necessary that police departments consider a range of prevention strategies for reducing officer response-to-resistance and increasing officer safety such as the adoption of BWCs and an increase in police officer training in general and implicit bias training specifically, along with other recognizable and successful practices to improve police—community relations such as community policing and other related strategies (Chappell and Piquero 2010; Zavala and Kurtz 2016). Without this constellation of efforts, the effectiveness of any one strategy may prove relatively futile. In this same vein, society and the police cannot simply expect that putting a BWC on every officer, in and of itself, will provide an immediate cure for the problem of racially biased policing practices, police excessive response-to-resistance, and police officers being targeted for being police officers, but it is perhaps a step in the right direction.

Discussion Questions

  • 1. What were the variables that the authors used to match the BWC officers with police who were not using cameras? Can you think of any additional variables that might have been helpful?
  • 2. What kind of impact did the matching procedure have on the results?
  • 3. For this study, the researchers obtained agency records for their data. Can you think of any other types of data that they could have used to do this study?

References

Ariel, Barak, William A. Farrar, and Alex Sutherland. 2015. “The Effect of Police Body- Worn Cameras on Use of Force and Citizens’ Complaints Against the Police: A Randomized Controlled Trial.” Journal of Quantitative Criminology 31:509-535.

Austin, Peter C. 2009. “Some Methods of Propensity- Score Matching had Superior Performance to Others: Results of an Empirical Investigation and Monte Carlo Simulations.” Biomedical Journal

51:171-184.

Chappell, Allison T. and Alex R. Piquero. 2010. “Applying Social Learning Theory to Police Misconduct.” Deviant Belutvior 25:89-108.

Ellis, Tom, Craig Jenkins, and Paul Smith. 2015. Evaluation of the Introduction of Personal Issue Body Worn Video Cameras (Operation Hyperion) on the Isle of Wight: Final Report to Hampshire Constabulary. Portsmouth, UK: University of Portsmouth: Institute of Criminal Justice Studies.

Goodall, Martin. 2007. Guidance for the Police Use of Body-Wom Video Devices. London: Home Office. Retrieved July 1, 2016 (http://library.college. police, uk/docs/homeoffice/guidance -body-worn- deviccs.pdf).

Grossmith, Lynne, Catherine Owens, Will Finn, David Mann, Tom Davies, and Laura Baika. 2015. Police, Camera, Evidence: London's Cluster Rattdomised Controlled Trial of Body Worn Video. London, UK: College of Policing and the Mayor’s Office for Policing and Crime (MOPAC).

Ho, Daniel E., Kosuke Imai, Gary King, and Elizabeth A. Stuart. 2007. “Matching as Nonparametric Preprocessing for Reducing Model Dependence in Parametric Causal Inference.” Political Analysis 15:199-236.

Ho, Daniel E., Kosuke Imai, Gary King, and Elizabeth A. Stuart. 2011. “Matchit: Nonparametric Preprocessing for Parametric Causal Inference.” Journal of Statistical Software 8:1-28.

Jennings, Wesley G., Lorie A. Fridell, and Mathew D. Lynch. 2014. “Cops and Cameras: Officer Perceptions of the Use of Body-Worn Cameras in Law Enforcement.” Journal of Criminal Justice 42:549-556.

Jennings, Wesley G., Mathew D. Lynch, and Lorie A. Fridell. 2015. “Evaluating the Impact of Police Officer Body-Worn Cameras (BWCs) on Response-to-Resistance and Serious External Complaints: Evidence from the Orlando Police Department (OPD) Experience Utilizing a Randomized Controlled Experiment.” Journal of Criminal Justice 43:480-486.

Jennings, Wesley G., Tara Richards, Elizabeth Tomsich, Angela Gover, and Rachael Powers. 2013. “A Critical Examination of the Causal Link between Child Abuse and Adult Dating Violence Perpetration and Victimization from a Propensity- Score Matching Approach.” Wmien and Criminal Justice 23:167-184.

Katz, Charles M., Mike Kurtenbach, David E. Choate, and Michael D. White. 2015. Phoenix, Arizona, Smart Policing Initiative: Evaluating the Impact of

Police Officer Body-Worn Cameras. Washington, DC: U.S. Department of Justice, Bureau of Justice Assistance.

Lum, Cynthia, Christopher Koper, Linda Merola, Amber Scherer, and Amanda Reioux. 2015. Existing and Ongoing Body Worn Camera Research: Knowledge Gaps and Opportunities. Fairfax, VA: George Mason University: Center for Evidence- Based Crime Policy.

ODS Consulting. 2011. Body Worn Video Projects in Paisley and Aberdeen, Self Evaluation. Glasgow: ODS Consulting.

Owens, Catherine, David Mann, and Roy Mckenna. 2014. The Essex BWV Trial: The Impact of BWV on Criminal Justice Outcomes of Domestic Abuse Incidents. London, UK: College of Policing.

Pearl, Judea. 2009. Causality: Models, Reasoning and Inference. 2nd ed. New York: Cambridge University Press.

Ready, Justin T. and Jacob T. Young. 2015. “The Impact of On-Officer Video Cameras on Police- Citizen Contacts: Findings from a Controlled Experiment in Mesa, AZ " Journal of Experimental Criminology 11:445-458.

Richards, Tara, M. Dwayne Smith, Wesley G. Jennings, Beth Bjerregaard, and Sondra Fogel. 2014. “An Examination of the Sex Disparity in Capital

Sentencing: A Propensity Score Matching Approach.” American Journal of Criminal Justice 39:681—697.

Rosenbaum, Paul R. 2002. Observatkmal Studies. 2nd ed. New York: Springer-Verlag.

Rosenbaum, Paul R. and Donald B. Rubin. 1983. “The Central Role of the Propensity Score in Observational Studies for Causal Effects.” Biometrics 70:41-55.

Roy, Allyson. 2014. On Officer Video Cameras: Examining the Effects of Police Department Policy and Assignment on Camera Use and Activaticm. (Unpublished master’s thesis). Arizona State University, Phoenix, AZ.

Sousa, William H., Terance D. Miethe, and Mari Sakiyama. 2015. Body Worn Cameras on Police: Results from a National Survey of Public Attitudes. Las Vegas: Center for Crime and Justice Policy, University of Nevada, Las Vegas.

Young, Jacob T. and Justin T. Ready. 2015. “Diffusion of Ideas and Technology: The Role of Networks in Influencing the Endorsement and Use of On-Officer Video Cameras.” Journal of Contemporary Criminal Justice 31:243-261.

Zavala, Egbert and Don L. Kurtz. 2016. “Applying Differential Coercion and Social Support Theory to Police Officers' Misconduct.” Deviant Behavior 37:877-892.

Reading 11.3 Psychometric Properties of the UNCOPE

Risk assessment tools are valuable assets for treatment providers as they work to identify appropriate correctional programs. These instruments are of varying effectiveness, however, since they need to be validated with the population for which they are going to be used. Proctor, Kopak, and Hoffmann (2017) tested the validity and reliability of the Unintended use, Neglect of responsibilities, desire/attempts to Cut down, Objection by others, Preoccupation, and Emotional distress (UNCOPE) screening instrument when used with justice-involved juveniles. UNCOPE is a very short, six-item tool intended to identify substance use disorders in respondents. To test this, the authors collected data on UNCOPE scores for a sample of juveniles who also participated in the Practical Adolescent Dual Diagnostic Interview (PADDI), which provides treatment staff much more information than the UNCOPE but also takes much longer to administer. The researchers obtained their data with the cooperation of a state juvenile justice agency. This article serves as a good example of a test of validity and reliability, but the authors also raise some other very important topics for us to consider when we think about administering screening tools.

 
Source
< Prev   CONTENTS   Source   Next >