Annotated Bibliography for research proposal

Dissecting the Complexities of the Relationship Between Police Officer–Civilian Race/Ethnicity Dyads and Less-Than-Lethal Use of Force

Katelyn K. Jetelina, PhD, MPH, Wesley G. Jennings, PhD, Stephen A. Bishopp, PhD, Alex R. Piquero, PhD, and Jennifer M. Reingle Gonzalez, PhD

Objectives.To examine how sublethal use-of-force patterns vary across officer–civilian

race/ethnicity while accounting for officer-, civilian-, and situational-level factors.

Methods. We extracted cross-sectional data from 5630 use-of-force reports from the

Dallas Police Department in 2014 and 2015. We categorized each officer–civilian in-

teraction into race/ethnicity dyads. We used multilevel, mixed logistic regression models

to evaluate the relationship between race/ethnicity dyads and the types of use of force.

Results. Forty-eight percent of use-of-force interactions occurred between a White

officer and a non-White civilian (White–non-White). In bivariate models, the odds of

hard-empty hand control and intermediate weapon use were significantly higher among

White–Black dyads compared with White–White dyads. The bivariate odds of in-

termediate weapon use were also significantly higher among Black–Black, Hispanic–

White, Black–Hispanic, and Hispanic–Black dyads compared with White–White dyads.

However, after we controlled for individual and situational factors, the relationship

between race/ethnicity dyad and hard-empty hand control was no longer significant.

Conclusions. Although we observed significant bivariate relationships between race/

ethnicity dyads and use of force, these relationships largely dissipated after we con-

trolled for other factors. (Am J Public Health. 2017;107:1164–1170. doi:10.2105/

AJPH.2017.303807)

Because of recent high-profile events, 1

public attention has increasingly focused on use-of-force interactions between police officers and civilians. Public opinion has attributed the degree of force used in any given interaction to the race/ethnicity of the civilian and police officer. This is expected given the fact that the race/ethnicity of a civilian is the strongest demographic factor that predicts public opinions, whether that be positive or negative, about the police.2–4

However, because of methodological in- consistencies and gaps in the extant research, the relationship among officer race/ethnicity, civilian race/ethnicity, and use of force is unclear.5

Police officer characteristics, and specifi- cally race/ethnicity, have received consider- able scientific attention and many studies have found no relationship between officer race/ethnicity and use of force.6–10 For

example, in a multisite study, Paoline and Terrill observed no racial or ethnic differences in coercive practices after they controlled for officer-, suspect-, and encounter-level characteristics.7 Also, Lott and Moody found no difference in fatal shootings of Black suspects among Black and White officers and concluded that their “findings [were] inconsistent with taste-based racial discrimi- nation against blacks by white police offi- cers.”11(p20) On the other hand, Garner et al. found that Hispanic officers had an increased

odds of using force compared with White officers, although no significant effects were observed for force severity by officer race/ethnicity.12 In addition, McElvain and Kposowa found that White officers were more likely to use deadly force than Hispanic officers and not more likely than Black officers.13

Literature evaluating the relationship between civilian race/ethnicity and use of force is also mixed.5 Many studies report no relationship between civilian race/ethnicity and force used by police officers.6,8,9,14–16

However, results from a recent meta-analysis suggest that minority civilians are more likely to be arrested, a form of force, than White civilians.17 Furthermore, a recent, highly publicized report found that Black civilians are more likely to be touched, handcuffed, and pepper-sprayed, but not more likely to encounter lethal forms of force.18

A valid assessment of the relationship between officer and civilian characteristics related to use of force is important to inform preventive efforts and to mend the tense community relationships between the police and the public.18,19 Although numerous studies have examined the relationship between officer or civilian race/ethnicity and use of force, none of these studies to our knowledge have explored how use-of- force patterns vary as a function of the officer– civilian race/ethnicity dyad, specifically beyond Black and White dyads. In other

ABOUT THE AUTHORS Katelyn K.Jetelina and Jennifer M. Reingle Gonzalez are withthe Universityof Texas Health Science Center, School of Public Health–Dallas. Wesley G. Jennings is with the University of South Florida, Department of Criminology, Tampa. Stephen A. Bishopp is with the Caruth Police Institute/Dallas Police Department, Dallas. Alex R. Piquero is with The University of Texas at Dallas, Program in Criminology, Richardson.

Correspondence should be sent to Katelyn Jetelina, 6011 Harry Hines Blvd, Dallas, TX 75390 (e-mail: katelyn.kassarjian@ utsouthwestern.edu). Reprints can be ordered at http://www.ajph.org by clicking the “Reprints” link.

This article was accepted March 16, 2017. doi: 10.2105/AJPH.2017.303807

1164 Research Peer Reviewed Jetelina et al. AJPH July 2017, Vol 107, No. 7

AJPH RESEARCH

words, officer and civilian race and ethnicity have been considered as independent risk factors for use of force; however, use of force is inherently nested in an officer–civilian dyad (e.g., force patterns may be different when a White officer is interacting with a Black civilian compared with a Black officer interacting with a Black civilian). To address this important problem, we examined how sublethal use-of-force patterns vary across officer–civilian race/ethnicity dyads, while accounting for officer-, civilian-, and situational-level factors.

METHODS Dallas, Texas, is the ninth largest city in the

United States (with a 2010 population of approximately 1.2 million) and is policed primarily by the 3551 sworn members of the Dallas Police Department (DPD). In February 2016, 82% (n = 2900) of DPD were male and 51% (n = 1818) identified as non-Hispanic White.

Since January 1, 2013, officers at DPD have been mandated to complete a use-of- force report, called the “Blue Team report,” every time they engage in use of force beyond compliant arrest. These reports document individual (e.g., complainant or civilian and officer characteristics), situational, and envi- ronmental factors related to each use-of-force altercation.20 Examples of force include, but are not limited to, deployment of physical defensive tactics (e.g., “take downs”), use of pressure points, oleoresin capsicum spray deployment (i.e., pepper spray), and drawing and discharge of weapons—batons, Tasers, and firearms. Each Blue Team report is linked to the completing officer’s badge number to allow for an assessment of how use-of-force patterns vary within and between police officers.

Blue Team reports are completed im- mediately after the incident (with the ex- ception of firearm discharges) and forwarded to field supervisors for approval. These re- ports are periodically uploaded to a publicly available database. Instances involving firearm discharges or in-custody-death in- cidents were excluded from these records. Officer-involved shooting incidents (i.e., shots fired by officers or at officers) and deaths occurring after an individual has

been detained or arrested are routinely investigated by the Special Investigative Unit. Detectives within those investigative areas are responsible for use-of-force entries into the Blue Team database after the initial investigation of each event. All other in- cidents involving a use of force greater than compliant handcuffing are recorded by either the officers involved or the responding supervisor immediately after the preliminary investigation. We used these abstracted records in the current study, which included data from 5630 Blue Team reports submitted by 1693 DPD officers. This represents all reports filed from January 2014 to December 2015.

Measures Outcome. Force used was recorded by

officers immediately after each incident. In the Blue Team database, officers were pre- sented with a drop-down menu of 40 types of force used to select from, including baton display, verbal command, handcuffing take down, leg restraint system, pressure points, and oleoresin capsicum spray (among others). In accordance with DPD policies and pro- cedures, 40 force options were categorized to represent the force continuum21: (1) verbal direction (e.g., verbal command and combat stance), (2) soft-empty hand control (e.g., held suspect down, pressure points, threatened to use Taser, and displayed Taser), (3) hard-empty hand control (e.g., joint locks or weapon display at person), and (4) in- termediate weapon use (e.g., pepperball saturation and used Taser). Deadly force was not included in this data set because of the rare occurrence of these cases (n = 26 in DPD from 2014 to 2015) and because of the fact that they are reported through a specialized unit within DPD. Because the force contin- uum variables are not mutually exclusive, we dichotomized the types of force used in each of the 4 outcomes for analysis (e.g., 0=no verbal direction used;1=verbal direction used).

Individual factors. We gathered officer gender, officer race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, other [including Native American and Asian]), and tenure (< 5 years, 5 to <10 years, ‡ 10 years), calculated from hire date, from Blue Team reports.

Civilian gender and civilian race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, other [including Native American

TABLE 1—Sample Description: Dallas Police Department, Dallas, TX, 2014–2015

Characteristics No. (%)

All use-of-force reports (n = 5630)

Race/ethnicity dyads (officer-citizen)

White–White 765 (14)

White–Non-White (Black/Hispanic/

other)

2685 (48)

White–Black 1806 (32)

White–Hispanic 764 (14)

White–Other 115 (2)

Black–Black 556 (10)

Hispanic–Hispanic 292 (5)

Black–White 160 (3)

Hispanic–White 210 (4)

Black–Hispanic 177 (3)

Hispanic–Black 563 (10)

Other (and missing) 222 (4)

Force continuum

Verbal 1879 (33)

Soft emptya 3070 (55)

Hard emptyb 2210 (39)

Intermediate weapon usec 391 (7)

Civilian under influence 2411 (43)

Civilian mental health symptomology 841 (15)

Unique officers involved in use-of-force

altercations (n = 1693)

Male gender 1461 (87)

Race/ethnicity

White non-Hispanic 935 (55)

Black non-Hispanic 360 (21)

Hispanic 337 (20)

Other 62 (4)

Tenure, y

< 5 425 (25) 5 to < 10 587 (35) ‡ 10 682 (40)

No. of use-of-force incidents

1 570 (34)

2 340 (20)

3 229 (14)

‡ 4 554 (32)

aFor example, held suspect down, pressure points, threatened to use Taser, and displayed Taser. bFor example, joint locks or weapon display at person. cForexample,pepperballsaturationandusedTaser.

AJPH RESEARCH

July 2017, Vol 107, No. 7 AJPH Jetelina et al. Peer Reviewed Research 1165

and Asian]) were self-reported by the officer as a part of the Blue Team report.

Race/ethnicity dyads. We matched the main exposure, the dyad of officer race/ ethnicity and civilian race/ethnicity per Blue

Team report, and categorized them into 10 groupings with officer race/ethnicity first and civilian race/ethnicity second (White–White [referent], White–Black, White–Hispanic, Black–White, Black–Black, Black–Hispanic,

Hispanic–White, Hispanic–Black, Hispanic– Hispanic, other).

Situational factors. Calls for service were categorized by the responding police officer into 8 mutually exclusive groups:

1. arrest, 2. call for assistance, 3. crime in progress, 4. service call, 5. suspicious activity, 6. traffic stop, 7. warrant execution, and 8. other.

These data provide requisite context about the motivation for the officer–civilian interaction.

Other risk factors. Officers reported whether they perceived that the complainant was under the influence of alcohol or drugs (dichotomized) or appeared to be suffering from severe mental health symptomology (dichotomized) during the incident. We in- cluded these variables in the model as con- founders because civilian impairment by substances and mental health symptoms has been associated with elevated use of force by police officers.14,22–24

Statistical Analysis To describe the prevalence of racial/ethnic

dyads in use-of-force incidents, we matched officer race/ethnicity and civilian race/ ethnicity and categorized them into 10 groupings. We used tabulations to describe the distribution of race/ethnicity dyads, use-of-force typologies within the force continuum, civilian and police officer de- mographics, and officer tenure. Because multiple officers could complete Blue Team Reports for the same incident, multivariate analyses accounted for the 2-stage clustering design (e.g., officers’ Blue Team reports were nested within each call).

We used SAS command glimmix (version 9.4; SAS Institute Inc, Cary, NC) to model 2-level, logistic mixed effects (random intercept) regression to evaluate the re- lationship between race/ethnicity dyads (White–non-White vs White–White; in- dependent variable) and types of sublethal force used (dependent variable). Model building commenced by fitting bivariate

TABLE 2—Prevalence and Adjusted Odds Ratios of Hard-Empty Hand Control and Intermediate Weapon Use Among White–Non-White Compared With White–White Officer–Citizen Dyads: Dallas Police Department, Dallas, TX, 2014–2015

Hard Empty (n = 2210)

Intermediate Weapon Use (n = 391)

Variable No.a AOR (95% CI) No.b AOR (95% CI)

Race/ethnicity dyads (officer–citizen)

White–White 284/481 1 (Ref) 35/730 1 (Ref)

White–Non-White 1146/1539 1.15 (0.94, 1.42) 183/2502 1.73 (0.98, 3.04)

Other 780/1399 0.89 (0.73, 1.10) 173/2006 1.94 (1.11, 3.42)

Officer level

Gender

Female 210/469 1 (Ref) 56/623 1 (Ref)

Male 2000/2950 1.53 (1.25, 1.89) 335/4615 0.89 (0.55, 1.44)

Tenure, y

< 5 609/1148 1 (Ref) 92/1665 1 (Ref) 5 to < 10 1103/1381 1.51 (1.29, 1.76) 179/2305 1.89 (1.26, 2.83) ‡ 10 498/891 1.07 (0.90, 1.28) 120/1269 1.97 (1.27, 3.06)

Civilian level

Gender

Female 360/730 1 (Ref) 46/1044 1 (Ref)

Male 1844/2680 1.33 (1.12, 1.59) 344/4180 2.26 (1.34, 1.82)

Situational level

Civilian under influence

No 1317/1902 1 (Ref) 205/3014 1 (Ref)

Yes 893/1518 0.74 (0.64, 0.85) 186/2225 1.35 (0.93, 1.96)

Civilian mental health symptomology

No 1960/2829 1 (Ref) 335/4454 1 (Ref)

Yes 250/591 0.53 (0.43, 0.66) 56/785 1.08 (0.63, 0.84)

Service type

Arrest 968/1464 1 (Ref) 168/2264 1 (Ref)

Call for cover 95/177 0.76 (0.55, 1.04) 25/247 1.76 (0.86, 3.61)

Crime in progress 151/120 1.72 (1.25, 2.38) c/262 0.45 (0.16, 1.26)

Service call 624/1015 0.88 (0.75, 1.04) 112/1527 0.93 (0.62, 1.41)

Suspicious activity 46/85 0.65 (0.41, 1.03) c/126 0.24 (0.06, 1.01)

Traffic stop 115/121 1.16 (0.82, 1.62) 10/226 0.49 (0.17, 1.40)

Warrant execution 17/21 1.13 (0.51, 2.48) c/36 1.33 (0.18, 9.80)

Other 194/416 0.62 (0.49, 0.78) 60/550 1.84 (1.09, 3.12)

Note. AOR = adjusted odds ratio; CI = confidence interval. To calculate crude odds ratios, divide the no. fraction for the predictor category by the no. fraction of the referent category. The crude odds ratios may not match Tables A and B (available as a supplement to the online version of this article at http://www.ajph.org) because they do not take into account the clustered study design. aNo. among reports with hard-empty hand control/no. among reports without hard-empty hand control. bNo. among reports with intermediate weapon use/no. among reports without intermediate weapon use. cn < 10.

AJPH RESEARCH

1166 Research Peer Reviewed Jetelina et al. AJPH July 2017, Vol 107, No. 7

regression models for each independent variable. Because significant relationships emerged in these bivariate models, we ex- panded models incrementally to include individual and situational covariates (officer demographics, tenure, civilian gender, civilian impairment by substances or mental health symptomology, and service type).

RESULTS A description of the sample is provided in

Table 1. From January 1, 2014, through December 31, 2015, 5630 unique Blue Team use-of-force reports were filed by 1693 unique DPD officers. White officers’ in- teractions with non-White civilians repre- sented 48% of the total interactions; of these, 67% were White–Black (White officer–Black civilian; n = 1806). Black–Black and His- panic–Black dyads each represented 10% of the sample. Officers reported use of soft-empty hand control in 55% of reports, hard-empty hand control in 39% of reports, verbal force in 33% of reports, and in- termediate weapon deployment in 7% of reports. Of the officers, 43% reported that the civilian was under the influence of alcohol or drugs and 15% reported the civilian to be exhibiting symptoms of a mental health condition.

Of the 1693 unique police officers in our sample, 87% were male and the ma- jority (55%) were non-Hispanic White.

Twenty-five percent of officers were employed by DPD for less than 5 years and 32% of officers had more than 4 use-of- force incidents recorded over the 2-year study period.

There was no significant bivariate re- lationship between White–non-White dyad and White–White dyad encounters in re- lation to verbal or soft-empty hand control (data not shown). Because significant bivariate models emerged betweenrace/ethnicity dyad and hard-empty hand control and interme- diate weapon use (Tables A and B, available as a supplement to the online version of this article at http://www.ajph.org), models were expanded to include officer-, civilian-, and situational-level confounders (Table 2). After we controlled for covariates, differ- ences in the odds of hard-empty hand control between White–non-White dyad and White–White dyad encounters were not significant (odds ratio [OR] = 1.15; 95% confidence interval [CI] = 0.94, 1.42). Similarly, variations in the odds of intermediate weapon use between White– non-White dyad and White–White dyad encounters were not significant after we controlled for officer-, civilian-, and situational-level factors (OR = 1.73; 95% CI = 0.98, 3.04).

Table 3 expands the race/ethnicity dyads to examine how these use-of-force patterns vary according to specific combinations of officer and civilian race/ethnicity. Results from bivariate models suggest that use of verbal commands and soft-empty hand

control techniques were significantly less frequently reported in Black–Black dyad encounters and Hispanic–Hispanic dyad en- counters compared with White–White dyad encounters. The odds of hard-empty hand control were significantly higher among White–Black dyad encounters compared with White–White dyad encounters. Furthermore, Black–Black, White–Black, Hispanic–White, Black–Hispanic, and Hispanic–Black dyad encounters had higher odds of intermediate weapon use than White–White dyad encounters.

Table 4 displays results from the multilevel, multivariate regression analyses examining the persistence in variation between race/ ethnicity dyads and use of force. After we controlled for officer-, civilian-, and situational-level covariates, the odds of hard-empty hand control were significantly lower among Black–Black dyad encounters compared with White–White dyad en- counters (OR = 0.74; 95% CI = 0.56, 0.98). The odds of intermediate weapon use were significantly higher among White–Black (OR = 2.34; 95% CI = 1.22, 4.53), Hispanic– White (OR = 4.14; 95% CI = 1.62, 10.60), Black–Hispanic (OR = 3.09; 95% CI = 1.08, 8.85), and Hispanic–Black (OR = 3.18; 95% CI = 1.47, 6.89) dyad encounters compared with White–White dyad encounters.

DISCUSSION The use of force by police, when

inappropriate, is a significant concern— especially when such instances are dispro- portionately experienced by certain races/ ethnicities. The detrimental effect of such circumstances has the power to strain confi- dence and trust in the police, which is essential to building effective crime-control partner- ships between civilians and the police.25–27 In accordance, empirical analysis of this issue is critical for theory and especially for police policy and practice.

In this study, we used the population of use-of-force incidents from the DPD in 2014 and 2015 to examine how the race/ethnicity of both the officer and the civilian were in- terrelated across a range of use-of-force in- cidents. Not only does our work present an important contribution to a salient national

TABLE 3—Crude Odds Ratios of Types of Sublethal Use of Force Across Officer–Civilian Race/Ethnicity Dyad: Dallas Police Department, Dallas, TX, 2014–2015

Variable Verbal, OR (95% CI)

Soft Empty, OR (95% CI)

Hard Empty, OR (95% CI)

Intermediate Weapon Use, OR (95% CI)

Race/ethnicity dyads

(officer–citizen)

White–White (Ref) 1 1 1 1

Black–Black 0.74 (0.56, 0.99) 0.74 (0.56, 0.97) 0.76 (0.58, 1.01) 2.22 (1.04, 4.77)

Hispanic–Hispanic 0.43 (0.29, 0.62) 0.64 (0.45, 0.90) 1.07 (0.76, 1.49) 0.58 (0.19, 1.75)

White–Black 0.93 (0.75, 1.16) 0.86 (0.69, 1.07) 1.31 (1.06, 1.62) 2.25 (1.19, 4.27)

Black–White 0.57 (0.36, 0.90) 0.87 (0.57, 1.34) 0.65 (0.42, 1.01) 2.06 (0.63, 6.69)

White–Hispanic 0.93 (0.71, 1.20) 0.82 (0.63, 1.06) 1.18 (0.92, 1.52) 1.57 (0.75, 3.29)

Hispanic–White 0.89 (0.60, 1.32) 0.84 (0.58, 1.23) 0.88 (0.60, 1.28) 3.94 (1.58, 9.83)

Black–Hispanic 0.76 (0.50, 1.17) 0.92 (0.61, 1.40) 0.84 (0.56, 1.27) 3.07 (1.09, 8.69)

Hispanic–Black 1.00 (0.76, 1.32) 0.79 (0.60, 1.04) 1.02 (0.78, 1.34) 2.60 (1.22, 5.51)

Note. CI = confidence interval; OR = odds ratio.

AJPH RESEARCH

July 2017, Vol 107, No. 7 AJPH Jetelina et al. Peer Reviewed Research 1167

issue, but its use of a broad conception of race and ethnicity in general and in examining various race dyads also specifically represents a key extension over much of the previous work in this area because of its concentrated focus on Whites and Blacks. Research more generally on criminological and criminal

justice topics has, unfortunately, not con- sidered Hispanics in large part because of data sources not containing the relevant information.28,29

The overarching finding from our analysis of these data showed that, although there were significant bivariate relationships

between race/ethnicity dyads and use of force, control for other factors related to both officers and civilians rendered most early relationships insignificant. Among hard-empty hand control, relationships interestingly dissipated once situational fac- tors, such as the citizen being under the in- fluence of alcohol or drugs, were introduced into the model. These results notwithstand- ing, the odds of intermediate weapon use among Hispanic–White, Black–Hispanic, White–Black, and Hispanic–Black dyads remained significant in complex multivariate models, and this is certainly not an in- consequential finding.

Limitations Findings from this study should be con-

sidered in light of 3 limitations. First, all data were obtained from police officer self- reports for which no information about validity or reliability of reporting practices is available. A number of safeguards are in place to prevent widespread falsification of doc- uments. Use-of-force reports are routinely reviewed and approved by police supervi- sors, a process likely to identify and flag details not consistent with the associated narrative. Falsified records are subject to internal affairs investigation and disciplinary action for all involved officers, a ramification that may deter officers from providing false information. That said, it is very difficult to obtain objective data on the entire range of use-of-force incidents in a single, large city over an extensive time period. The widespread use of body cameras may play a role in validating the reporting practices of police officers in future research.

Second, use of data from a single police department limits the external validity of these findings. In the absence of national (or state) surveillance systems tracking police use of force, this limitation will continue to limit our knowledge about officer–civilian interactions. Finally, the use-of-force in- cident reports did not contain all possible factors that play a role in police–citizen interactions. For example, community-level factors, such as crime rates; individual-level factors, such as civilian or officer body height or weight, physical fitness, and visible tattoos; or situational-level factors, such as day or night shift, may all

TABLE 4—Adjusted Odds Ratios of Hard-Empty Hand Control and Intermediate Weapon Use Across Officer–Civilian Race/Ethnicity Dyad: Dallas Police Department, Dallas, TX, 2014– 2015

Variable Hard Empty, AOR (95% CI) Intermediate Weapon Use, AOR (95% CI)

Race dyads (officer–citizen)

White–White (Ref) 1 1

Black–Black 0.74 (0.56, 0.98) 2.00 (0.90, 4.42)

Hispanic–Hispanic 1.01 (0.71, 1.42) 0.69 (0.23, 2.09)

White–Black 1.18 (0.95, 1.47) 2.34 (1.22, 4.53)

Black–White 0.70 (0.45, 1.09) 2.11 (0.64, 7.00)

White–Hispanic 1.05 (0.80, 1.36) 1.58 (0.75, 3.36)

Hispanic–White 0.93 (0.63, 1.36) 4.14 (1.62, 10.6)

Black–Hispanic 0.85 (0.56, 1.28) 3.09 (1.08, 8.85)

Hispanic–Black 0.95 (0.72, 1.25) 3.18 (1.47, 6.89)

Officer level

Gender

Female (Ref) 1 1

Male 1.52 (1.23, 1.89) 1.00 (0.59, 1.70)

Tenure, y

< 5 (Ref) 1 1 5 to < 10 1.51 (1.29, 1.77) 2.00 (1.28, 3.13) ‡ 10 1.08 (0.90, 1.30) 2.45 (1.49, 4.01)

Civilian level

Gender

Female (Ref) 1 1

Male 1.34 (1.12, 1.61) 2.28 (1.26, 4.12)

Situational level

Civilian under influence

No (Ref) 1 1

Yes 0.76 (0.65, 0.88) 1.31 (0.86, 1.97)

Civilian mental health symptomology

No (Ref) 1 1

Yes 0.53 (0.43, 0.66) 1.01 (0.55, 1.84)

Service type

Arrest (Ref) 1 1

Call for cover 1.79 (0.57, 1.09) 1.93 (0.87, 4.23)

Crime in progress 1.74 (1.24, 2.44) 0.31 (0.09, 1.03)

Service call 0.88 (0.74, 1.03) 1.03 (0.65, 1.65)

Suspicious activity 0.63 (0.39, 1.01) 0.21 (0.05, 0.97)

Traffic stop 1.27 (0.89, 1.82) 0.78 (0.25, 2.39)

Warrant execution 1.02 (0.45, 2.29) 1.43 (0.18, 11.6)

Other 0.62 (0.49, 0.79) 1.96 (1.09, 3.53)

Note. AOR = adjusted odds ratio; CI = confidence interval.

AJPH RESEARCH

1168 Research Peer Reviewed Jetelina et al. AJPH July 2017, Vol 107, No. 7

influence how use-of-force situations play out on the streets during police–citizen encounters. Because officer behavior could vary according to these circumstances, future research should strive to collect those data and append them to force incident reports.30

Public Health Implications Ultimately, these findings have important

implications for policy within the DPD. For example, although the majority of use-of- force variability dissipated across race/ ethnicity dyad once a host of covariates was introduced into the models, it is important for the DPD to continue to improve trust and legitimacy within the police–citizen in- teraction, especially with respect to use of force. This has been, and should continue to be, accomplished through public in- formational transparency via open data from most operational units within DPD. Yet, to the general civilian, deciphering police ter- minology is often difficult. The provision of codebooks that operationalize variables consistent with requirements for most grant-funded social research would increase DPD’s data transparency and utility. Un- fortunately, DPD and most other policing agencies do not have sufficient resources to gather, manipulate, analyze, and present data in a publicly accessible, user-friendly manner.

The results of this study also have policy implications for DPD training. Looking to improving police–citizen interaction dy- namics with an eye toward reducing excessive use of force, new implicit bias training has begun to incorporate officers’ thought pro- cesses and decision-making through high- quality, scenario-based judgment training.31

Training programs for DPD should be tied with findings from the neuropsychological literature. For example, meta-cognitive ap- proaches, such as coaching DPD officers with the constructive and deliberate practice of recognizing the way they think and why, may allow officers to act and reflect more objec- tively in high-intensity situations. Officers will gain increased self-awareness, which, in turn, may significantly improve officer– citizen interactions, specifically surrounding race/ethnicity and intermediate weapon use variability, as seen in this current study. This method has improved performance around

other high-stress occupations, such as trauma care teams.32

In sum, although in this study we used data from 1 large, urban police department, the issues raised are relevant to all police agencies, especially those serving diverse communities. President Obama’s Task Force on 21st Century Policing has recently made a num- ber of recommendations surrounding “shoot/don’t shoot” scenarios, de-escalation tactics, and the use of less-than-lethal alternatives to deadly force to reduce the occurrence of police use of force.27 In addition, the President’s Task Force recom- mendations have led to 130 police agencies making much of their data openly available to the public. Encounters involving police use of force and their relationship to citizen race/ethnicity, real or perceived, coupled with publicly available data, fervently calls for research like the current study to ex- amine the nuances of those often unstable encounters.

In this same vein, there is a growing and encouraging amount of evaluation research highlighting the effectiveness of police officer body-worn cameras for reducing the fre- quency and prevalence of the use of force and civilian-generated external complaints.33

Thus, it is likely that the implementation of strategies and initiatives along the lines out- lined and recommended by President Oba- ma’s Task Force in addition to the adoption of police officer body-worn cameras will be a step in the right direction to prevent officer misuse of force, reduce the likelihood of force rising to the level at which deadly force is necessary, and improve police–community relations, particularly in racial/ethnic mi- nority communities.

CONTRIBUTORS K. K. Jetelina managed the data set, ran the analyses, and drafted the article. S. A. Bishopp provided feed- back on the data set and specific Dallas Police De- partment policies and implications. W. G. Jennings and A. R. Piquero drafted the Discussion and assisted in editing throughout. J. M. Reingle Gonzalez provided critical analysis, feedback, and edits throughout the development of the article.

ACKNOWLEDGMENTS Funding for this study was provided by the University of Texas School of Public Health’s Front of the Envelope Innovation Award.

HUMAN PARTICIPANT PROTECTION Institutional review board approval was received from the University of Texas Health Science Center.

REFERENCES 1. The DPD Beat. Investigative update regarding the deadly attack on police officers. Dallas, TX: Dallas Police Department; 2016.

2. Webb VJ, Marshall CE. The relative importance of race and ethnicity on citizen attitudes toward the police. Am J Police. 1995;14(2):45–66.

3. Schuck AM, Rosenbaum DP, Hawkins DF. The in- fluence of race/ethnicity, social class, and neighborhood context on residents’ attitudes toward the police. Police Q. 2008;11(4):496–519.

4. Lai Y-L, Zhao JS. The impact of race/ethnicity, neighborhood context, and police/citizen interaction on residents’ attitudes toward the police. J Crim Justice. 2010; 38(4):685–692.

5. Klahm CF, Tillyer R. Understanding police use of force: a review of the evidence. Southwest J Crim Justice. 2010;7(2):214–239.

6. Lawton BA. Levels of nonlethal force: an examination of individual, situational, and contextual factors. J Res Crime Delinq. 2007;44(2):163–184.

7. Paoline EA, Terrill W. Police education, experience, and the use of force. Crim Justice Behav. 2007;34(2): 179–196.

8. McCluskey JD, Terrill W. Departmental and citizen complaints as predictors of police coercion. Policing. 2005; 28(3):513–529.

9. McCluskey JD, Terrill W, Paoline EA. Peer group aggressiveness and the use of coercion in police–suspect encounters. Police Pract Res. 2005;6(1):19–37.

10. McElvain JP, Kposowa AJ. Police officer character- istics and internal affairs investigations for use of force allegations. J Crim Justice. 2004;32(3):265–279.

11. Lott JR, Moody CE. Do White police officers unfairly target Black suspects? 2016. Available at: https://ssrn. com/abstract=2870189. Accessed January 3, 2017.

12. Garner JH, Maxwell CD, Herauz CG. Characteristics associated with the prevalence and severity of force used by the police. Justice Q. 2002;19(4):705–745.

13. McElvain JP, Kposowa AJ. Police officer character- istics and the likelihood of using deadly force. Crim Justice Behav. 2008;35(4):505–521.

14. Morabito MS, Socia KM. Is dangerousness a myth? Injuries and police encounters with people with mental illnesses. Criminol Public Policy. 2015;14(2): 253–276.

15. Phillips T, Smith P. Police violence occasioning citizen complaint: an empirical analysis of time–space dynamics. Br J Criminol. 2000;40:480–496.

16. Sun I, Payne B. Racial differences in resolving con- flicts: a comparison between Black and White police officers. Crime Delinq. 2004;40:516–541.

17. Kochel TR, Wilson DB, Mastrofski SD. Effect of suspect race on officers’ arrest decisions. Criminology. 2011;49(2):473–512.

18. Fryer RG. An empirical analysis of racial differences in police use of force. Cambridge, MA: The National Bureau of Economic Research; 2016.

19. Wolfe SE, Nix J. The alleged “Ferguson Effect” and police willingness to engage in community partnership. Law Hum Behav. 2016;40(1):1–10.

20. Ariel B, Farrar WA, Sutherland A. The effect of police body-worn cameras on use of force and citizens’ com- plaints against the police: a randomized controlled trial. J Quant Criminol. 2015;31(3):509–535.

AJPH RESEARCH

July 2017, Vol 107, No. 7 AJPH Jetelina et al. Peer Reviewed Research 1169

21. General order: 900.00 response continuum, 901.04. Dallas, TX: Dallas Police Department; 2015.

22. Kaminski RJ, Digiovanni C, Downs R. The use of force between the police and persons with impaired judgment. Police Q. 2004;7(3):311–338.

23. Smith MR, Kaminski RJ, Alpert GP, Fridell LA, MacDonald J, Kubu B. A multi-method evaluation of police use of force outcomes. Washington, DC: US Justice Department; 2009.

24. Johnson RR. Suspect mental disorder and police use of force. Crim Justice Behav. 2011;38(2):127–145.

25. Tyler T. Why People Obey the Law. Princeton, NJ: Princeton University Press; 1990.

26. National Research Council. Identifying the Culprit: Assessing Eyewitness Identification. Washington, DC: The National Academies Press; 2014.

27. President’s Task Force on 21st Century Policing. Final Report of the President’s Task Force on 21st Century Policing. Washington, DC: Office of Community Oriented Po- licing Services; 2015.

28. Piquero AR. Disproportionate minority confine- ment. Future Child. 2008;18(2):59–79.

29. Piquero AR. Understanding race/ethnicity differ- ences in offending across the life course: gaps and op- portunities. J Dev Life Course Criminol. 2015;1:21–32.

30. Smith DA. The neighborhood context of police behavior. In: Morris N, Tonry M, eds. Crime and Justice: An Annual Review of Research. Chicago, IL: University of Chicago Press; 1986: 313–341.

31. Fridell L, Lim H. Assessing the racial aspects of police force using the implicit- and counter-bias perspectives. J Crim Justice. 2016;44:36–48.

32. Fairchild RM. Practical ethical theory for nurses responding to complexity in care. Nurs Ethics. 2010;17(3): 353–362.

33. Jennings WG, Fridell LA, Lynch M, Jetelina KK, Reingle Gonzalez J. A quasi-experimental evaluation of the effects of police body-worn cameras (BWCs) on response-to-resistance in a large metropolitan police department. Deviant Behav. 2016; Epub ahead of print, November 7, 2016.

AJPH RESEARCH

1170 Research Peer Reviewed Jetelina et al. AJPH July 2017, Vol 107, No. 7

Reproduced with permission of copyright owner. Further reproduction prohibited without permission.