Public Attitudes About Body-Worn Cameras in Police Work: A National Study of the Sources of Their Contextual Variability

Abstract

Based on a large national sample of U.S. adults, the current study examines the nature and correlates of public support for body-worn cameras (BWCs) in various policing activities. Multivariate analyses were performed to assess the direct and moderating effects of individuals’ socioeconomic characteristics, general police attitudes and experiences, and specific beliefs about benefits of BWCs on the level of public support for this technology. Strong public support for BWC usage is found across different areas of police work. However, substantial contextual variability in this support is also evident when the analysis focused on the conjunctive influences of individuals’ level of confidence in social institutions, personal involvement in these institutions, and beliefs about police legitimacy and their effectiveness. These results are discussed in terms of their implications for future research on the sources of public receptivity, resistance, and change in these attitudes about BWCs over time.

Keywords

body cameras public attitudes contextual variability

Body-worn cameras (BWCs) have become a central feature of urban policing in contemporary American society. Their acceptance and usage by police departments derives from the multiple presumed benefits associated with this technology. For example, BWCs are widely regarded for their potential in increasing transparency in police–citizen encounters, moderating citizen and police behavior (e.g., reduces use of force, citizen complaints, lawsuits), building trust and police legitimacy, gathering visual evidence at crime scenes, and recording interviews with witnesses and victims (Ariel, Farrar, & Sutherland, 2015; Crow, Snyder, Crichlow, & Smykla, 2017; Sousa, Miethe, & Sakiyama, 2015, 2017; White, 2014). However, concerns about BWCs’ usage also exist. These concerns include privacy issues, the selective framing of these video accounts, the economic costs associated with the technology, and the greater public reluctance to provide information to the police (Miller, Tolliver, & Police Executive Research Forum, 2014; White, 2014).

Despite these negative concerns, public support for BWCs in police work is extremely strong. More than 80% of adult respondents are consistently found in previous national surveys to support BWCs’ usage by police (e.g., Moore, 2015; Pew, 2017; Sousa et al., 2017). County and citywide surveys indicate a similar pattern of high support for BWCs and strong agreement about their benefits to police–citizen relations (Crow et al., 2017; Lawrence, Peterson, & Thompson, 2018). However, most of these surveys on public support for body cameras were conducted several years ago (e.g., 2015, 2016), a time period in which BWCs were still an emergent technology. Given greater visibility of BWCs in media accounts and more empirical research on their actual effects in police practices, it is important to investigate whether public receptivity and resistance to this technology has changed in more recent times (i.e., 2017–2018) and examine some of the contextual factors that may impact public sentiments.

The objective of the current study is to explore the level and correlates of public support for BWC usage across different areas of police work. Based on a large national online sample, it also evaluates whether these public attitudes about BWCs are moderated by individuals’ confidence in social institutions, personal involvement in these institutions, and beliefs about police legitimacy and effectiveness. The results of these analyses are discussed in terms of their limitations and implications for future research.

Public Support for BWCs and Its Correlates

Over the last decade, a number of local and national surveys have been conducted to explore the nature of public support for BWCs and the correlates of these attitudes. The results of this previous research are summarized below.

Public Attitudes About BWCs in Police Work

Public attitudes about BWCs have been examined in several national and local survey projects. These surveys have elicited general opinions about “requiring police to wear body cameras” (Moore, 2015; Sousa et al., 2015, 2017), “the use of body cameras by police” (Pew, 2017), and whether “video cameras should be worn by all officers” (White, Todak, & Gaub, 2017). These surveys also typically include questions about specific benefits of BWCs, including the level of public agreement that this technology “increase(s) transparency,” “reduce(s) use of excessive force,” “improve(s) police–citizen relations,” “increase(s) citizen’s trust,” “decrease(s) racial tension,” “make(s) officers act more professionally,” and “increase(s) citizen’s cooperation with police” (Sousa et al., 2017; White et al., 2017). Survey items that tap negative perceptions about this technology include questions about their impact on “invading privacy” and “hurting police–community relations” (Crow et al., 2017; White et al., 2017).

Regardless of its national or local scope, survey data indicate wide public support for BWCs in police work. The level of support for BWCs in national surveys ranges from a low of 85% in May 2015 (Sousa et al., 2017) to a high of 93% in September 2016 (Pew, 2017). Among citywide surveys, White, Todak, and Gaub (2017) found that 86% of a sample of Spokane residents supported BWCs usage by their local police.

Public attitudes about the benefits of BWCs also reflect high support, but these opinions vary across different domains of these benefits. For example, Sousa, Miethe, and Sakiyama (2017) found substantial public agreement that BWCs “increase transparency” (91%) and “reduce use of excessive force” (80%), but far lower support for the ideas that BWCs “improve police–citizen relations” (66%), “increase citizen’s trust” (61%), and “decrease racial tension” (36%). Similarly, Crow, Snyder, Crichlow, and Smykla (2017) found extremely high agreement among sample respondents that BWCs will “assist in the collection of quality evidence” (88%) and “improve police officer behavior” (87%), but less support for statements that BWCs will “improve residents’ behavior” (79%) and “improve views about police legitimacy” (78%). The results of multiwave samples of Milwaukee resident (Lawrence et al., 2018) indicate that about 87% of sample respondents in June 2018 believed that “BWCs hold Milwaukee police accountable for their behavior” and 84% of them also believed that this technology improved police relationships with community members. In contrast, previous studies also reveal that less than one fourth of adult survey respondents believe that BWCs violate the personal privacy of officer, crime victims, or witnesses (Crow et al., 2017; Sousa et al., 2017).

Correlates of BWCs’ Attitudes

Coupled with the dominant finding of high public support, previous research has also identified sources of variability in attitudes about BWCs. These correlates of support for BWCs examined in previous studies include (1) individuals’ sociodemographic characteristics (e.g., gender, age, race/ethnicity, education, income, residency, crime concerns) and (2) police-related experiences and attitudes (e.g., police interactions, beliefs about police performance, procedural justice).

Among individuals’ sociodemographic attributes, public beliefs about BWCs’ usage and their benefits are most strongly influenced by the respondent’s race and age, but the nature of these effects vary across studies. For example, Sousa et al. (2017) found that Black residents had the greatest support for BWCs, but they also had less positive views than other racial/ethnic groups about the presumed benefits of this technology (e.g., increasing transparency, improving police–citizen relations, increasing citizen’s trust in police). Adults under 30 years old in this study also had more positive views about BWCs’ benefits than older persons. Other sociodemographic characteristics (e.g., gender, education, income, urban residency) have little impact on attitudes about BWC usage and their perceived benefits.

Previous research has shown that public attitudes about BWCs and other emergent technology in policing (e.g., aerial drones) are influenced by various police-related attitudes and experiences. These police-related factors may also impact the nature of the sociodemographic correlates of public support. For example, Crow et al. (2017) found that beliefs about police fairness and greater police interactions were associated with more positive views about the benefits of BWCs. The effects of individuals’ age and race on beliefs about BWC benefits were also found in this study to be primarily indirect, transmitted through the impact of these demographic variables on individuals’ beliefs about procedural fairness, crime concerns, and police performance.

Extensions of Previous Research

Previous national and local surveys provide some empirical evidence of public attitudes about BWCs usage, their perceived benefits, and correlates. However, to better understand the nature of public receptivity and resistance to this technology, several measurement, theoretical, and analytic issues underlying previous research warrant further investigation. These areas of extension of previous research are the focus of the current study and are summarized below.

Measurement issues

Previous surveys of BWC attitudes have focused primarily on general measures of public support for BWCs (e.g., “requiring police to wear body cameras”), specific measures of their benefits (e.g., transparency, reducing use of force and misconduct, improving police–citizen relations), and potential costs (e.g., violating privacy). Previous survey research, however, has not explored whether support for BWCs varies across different domains of police work (e.g., routine traffic stops, neighborhood patrol, crime scene investigation, crowd monitoring).

There are several substantive reasons why public support for BWCs may exhibit domain-specific differences. First, previous survey research (Sousa et al., 2015) indicates that support varies widely across different types of public safety officials (e.g., 95% support of BWCs for local police, 81% support for TSA officers, 56% for EMT officials). This variation across different officials may reflect differences in the specific aspects of their respective work. Second, previous research (Heen, Lieberman, & Miethe, 2017) indicates that public support for aerial drones is far greater in reactive policing activities (e.g., search/rescue activity, crime scene photography) than proactive activities (e.g., crowd monitoring, detecting traffic violations). Drone surveillance also elicits less public support when it involves monitoring activity within neighborhoods rather than business areas (Sakiyama, 2017). Given that drone usage is similar to body cameras as an emergent visual technology, these differences by type of police activity and its location may also characterize public support for BWCs across different domains of police work. For these reasons, it is important to explore the robustness of the substantive findings about the level of public support for BWCs across different areas of policing.

Theoretical and conceptual issues

Previous research on public perceptions of BWCs has used a path analytic framework and structural equation models to conceptualize and estimate the relationships between individuals’ sociodemographic attributes, police-related factors, and BWCs support (see Crow et al., 2017; Sousa et al., 2017). Within this analytic framework, Crow et al. (2017) found that measures of police performance exhibit a significant direct effect on individuals’ perceptions of the benefits of BWCs and mediate the impact of ratings of procedural fairness and crime concerns. However, it is also possible that wider social structural conditions associated with public confidence in social institutions, police legitimacy, and individual’s personal involvement in activities to support social institutions moderate the nature and magnitude of the observed relationships between these police-related factors and support for BWCs. The theoretical basis for this expectation is presented below.

National polls in the United States over the last three decades have documented the erosion of public confidence in social institutions (Norman, 2016). The level of public trust and confidence in these social institutions is important because it is a major component of various theories of institutional legitimacy. Habermas (1975), for example, argues that a “legitimacy crisis” occurs when the state does not maintain the requisite level of mass public loyalty necessary to perform its functions. LaFree (1998) makes a similar argument about “losing legitimacy” by highlighting the empirical linkage between changes in U.S. crime rates and the decline in social institutions. Building on Merton’s (1938) classic work, Messner and Rosenfeld (2013) suggest another type of institutional crisis (i.e., institutional anomie) that derives from (1) an overemphasis on the economy that is not counterbalanced by attention to noneconomic institutions (e.g., family, education, religion) and (2) the intrusion of economic success ideology into the functions of these other institutions.

Based on these theoretical arguments, we contend that individual differences in their levels of institutional anomie (i.e., low confidence in social institutions) and personal detachment from these social institutions (i.e., lack of personal involvement in activities to support these activities) may (1) directly influence public views about BWCs and (2) moderate the causal significance of police-related factors (e.g., views about police legitimacy, police effectiveness) on the public support for this technology. In particular, among individuals with low confidence and involvement in social institutions, attitudes about BWCs may be largely unaffected by police-related factors because their impact is nullified by these individuals’ general feelings of institutional and personal detachment. However, for those more firmly attached to social institutions, police-related attitudes may have a greater impact on support for BWCs because positive views about police legitimacy and effectiveness provide further affirmation of the trust and involvement in these social institutions. These types of moderating effects have not been investigated in previous multivariate analyses of public attitudes about BWCs.

Analytical issues

Within traditional methods of multivariate analysis, the nature and magnitude of moderating effects are often assessed through the estimation of group-specific models, interactive effects, and formal tests of the equality of sets of coefficients across these models (e.g., Chow, 1960; Kim, Kaye, & Wright, 2001; Paternoster, Brame, Mazerolle, & Piquero, 1998). However, for studying the context-specific sources of public support and opposition to BWCs, these traditional analytic methods are limited in several respects. First, these methods assume causal symmetry (i.e., A → B implies Not A → Not B). For example, if high ratings of police legitimacy are linked to high support for BWCs, then low ratings of legitimacy are also linked to low support. Second, the principles of causal equifinality (i.e., the presence of multiple pathways leading to same outcome) are not easily evaluated within these traditional context-specific analyses.

As an alternative to these traditional statistical models, the method of conjunctive analysis is specifically designed to evaluate the nature and magnitude of causal complexity among discrete qualitative variables (see Hart & Miethe, 2014; Miethe, Hart, & Regoeczi, 2008; Rennison & DeKeseredy, 2017; Venger, 2017). When applied to the study of possible moderating effects in public support for BWCs, conjunctive analysis provides empirical answers to several basic research questions. First, what is the nature and magnitude of contextual variability in support for BWCs (i.e., what are the particular combinations of sociodemographic attributes and police-related attitudes that have the highest and lowest levels of support)? Second, are particular subsets of cases (e.g., persons with high ratings of police legitimacy and effectiveness) always associated with high levels of support for BWCs? Does causal symmetry also typify these particular subsets (i.e., low legitimacy and low police effectiveness are always associated with low levels of support for BWCs and the converse is also true)? Third, does a pattern of invariance or context-specific effects best represent the impact of particular sociodemographic attributes and police-related attitudes on public support and opposition to BWCs? By evaluating these types of questions, conjunctive analysis offers a more nuanced method for exploring the sources and magnitude of variability in public attitudes about this technology.

The Present Study

Previous research reveals high levels of public support for BWCs and general agreement about the perceived benefits of this technology. However, the extant research has not examined whether this public support is consistent across different types of police work. The use of body cameras in police work was also an emergent technology when most survey data on them were gathered (i.e., 2013–2016), so it is largely unknown whether public receptivity and resistance to this evolving technology has changed in more recent times (i.e., 2017–2018). No research to date has examined whether the impact of police-related factors (e.g., beliefs about police legitimacy, effectiveness) on support for BWCs is moderated by individuals’ confidence and involvement in social institutions. The conjunctive impact of various sets of variables on support for BWC has also been an unexplored area in previous research.

Using a national online survey design, the current study examines two research questions about public attitudes about the use of BWCs in police work. First, what sociodemographic characteristics, police-related attitudes and experiences, and specific beliefs about BWCs are significant predictors of public support for this technology in different areas of police work? Second, what is the nature and magnitude of the moderating effects of individuals’ attitudes about BWCs’ benefits, their level of confidence in social institutions, personal involvement in activities supporting these institutions, and beliefs about police legitimacy and effectiveness on public support for this technology? The results of this study are discussed in terms of their implications for future research on the sources of receptivity, resistance, and change in public support for BWCs in contemporary policing.

Data and Method

Sample

A multiwave national online sampling design was used to address the research questions underlying the current study. Each of the three separate waves of the sample were generated through a panel of survey participants, developed by Qualtrics, a national company that specializes in online research platforms and sample construction. Study participants also used Qualtrics as the electronic platform to gain access and complete the survey instrument.

The three waves of the survey were launched in the following time periods: Wave 1 (June 1–June 15, 2017), Wave 2 (October 25–December 20, 2017), and Wave 3 (April 16–June 28, 2018). Within each wave of the survey, the sampling frame was stratified by gender (male, female), race/ethnicity (i.e., non-Hispanic White, Black, Hispanic) and income (i.e., <US$30,000 household income, >US$30,000 household income). At least 80 respondents were obtained from each of the 12 strata within each wave. Over 1,000 online surveys were completed for each wave, with total sample sizes of 1,286 respondents for Wave 1, 1,008 for Wave 2, and 1,012 for Wave 3. When respondents with any missing data were excluded, the sample size across the three survey waves was reduced by 5–10% (N’s [no missing] = 3,051 [overall]; 1,215 [Wave 1]; 923 [Wave 2]; and 913 [Wave 3]).¹

To adjust for the stratified sampling design, poststratification weighting by gender, race/ethnicity, and household income was conducted using the 2017 population estimates from the U.S. Census. The sample data were also weighted by the respondent’s age to further increase the convergence between the sample and U.S. population distributions. The final sample weights were readjusted to retain the total sample size of 3,051 cases. All analyses in the current study are based on these weighted sample data.

A comparison of the level of public support for BWCs in different domains revealed some significant differences across the three survey waves (e.g., support for BWCs for crowd management and routine traffic stops decreased over time). However, these “significant” differences were substantively small (i.e., in the range of 3–4% differences) and limited to these two particular domains. As a result, the combined data across waves are used here for examining the research questions underlying this study.

Measures

Dependent variables

Survey participants were asked to indicate their level of support or opposition to BWCs in six different domains of police work on a 5-point scale (1 = strongly oppose to 5 = strongly support). The areas of BWCs’ usage in police work include (1) routine traffic stops, (2) neighborhood police patrols involving “stop and questioning,” (3) crime scene investigations, (4) crowd management in public areas, (5) interviewing crime victims, and (6) all areas of police work. The particular measure about using BWCs in “all areas of police work” is most similar to the items used in previous research and public polls (e.g., Pew, 2017; Sousa et al., 2017). As shown in Table 1, each of these dependent variables was recoded into binary categories (0 = oppose/unsure, 1 = support).

Table 1.

Coding and Univariate Statistics for All Variables.

Variable	Coding	n	Mean or %
BWC support—All areas of policing	1 = support	2,536	83%
	0 = unsure	310	10%
	0 = oppose	205	7%
BWC support—Traffic stop	1 = support	2,651	87%
	0 = unsure	141	5%
	0 = oppose	259	8%
BWC support—Neighborhood patrol	1 = support	2,554	84%
	0 = unsure	178	6%
	0 = oppose	319	10%
BWC support—Crime scene work	1 = support	2,760	90%
	0 = unsure	140	5%
	0 = oppose	151	5%
BWC support—Crowd monitoring	1 = support	2,721	89%
	0 = unsure	145	5%
	0 = oppose	185	6%
BWC support—Interview victim	1 = support	2,679	88%
	0 = unsure	181	6%
	0 = oppose	191	6%
Gender	1 = male	1,554	51%
Gender	0 = female	1,497	49%
Age	1 = <30 years old	464	15%
Age	0 = ≥30 years old	2,587	85%
Race/ethnicity	1 = White	2,113	69%
	2 = Black	374	12%
	3 = Hispanic	564	19%
Annual household income	1 = <US$30,000	898	29%
Annual household income	0 = ≥US$30,000	2,153	71%
Educational attainment	1 = college graduate	1,318	43%
Educational attainment	0 = <college graduate	1,733	57%
Political party identification	1 = democrat	1,085	36%
Political party identification	0 = other	1,966	64%
Years living in neighborhood	1 = >5 years	1,694	56%
Years living in neighborhood	0 = ≤5 years	1,357	44%
Size of city population	1 = >1 million	1,680	55%
Size of city population	0 = <1 million	1,371	45%
Region of country of resident	1 = South	1,251	41%
Region of country of resident	0 = Non-South	1,800	59%
Confidence in social institutions (high confidence)	Mean of 7 items (0–2 point scale)	3,051	1.04
	1 = rating > 1 (“some”)	1,931	63%
	0 = rating ≤ 1	1,120	37%
Personal involvement in institutions (high personal involvement)	Mean of 8 items (0–2 point scale)	3,051	1.06
	1 = rating > 1 (“moderate”)	1,920	63%
	0 = rating ≤ 1	1,131	37%
Crime concerns	Mean of 5 items (0–2 point scale)	3,051	0.79
Level of neighborhood crime	1 = “high” crime area	568	19%
Level of neighborhood crime	0 = <“high” crime area	2,483	81%
Rating of prior police experience	1 = “poor”	333	11%
Rating of prior police experience	0 = “fair to excellent”/none	2,718	89%
Police effectiveness (high perceived effectiveness)	Mean of 4 item (1–5 point scale)	3,051	3.67
	1 = rating > 3 (“good/excellent”)	2,164	71%
	0 = rating ≤ 3 (“poor/fair”)	887	29%
Procedural justice (high procedural justice)	Mean of 6 items (1–5 point scale)	3,051	3.74
	1 = rating > 3 (“agree”)	2,268	74%
	0 = rating ≤ 3 (“unsure/disagree”)	783	26%
Police legitimacy (high police legitimacy)	Mean of 5 items (1–5 point scale)	3,051	3.64
	1 = rating > 3 (“agree”)	2,230	73%
	0 = rating ≤ 3 (“unsure/disagree”)	821	27%
BWC reduce use of force	1 = yes	2,206	72%
BWC reduce use of force	0 = no/unsure	845	28%
BWC violate suspect’s privacy	1 = yes	785	26%
BWC violate suspect’s privacy	0 = no/unsure	2,266	74%
BWC accurate account of events	1 = yes	2,334	76%
BWC accurate account of events	0 = no/unsure	717	24%

Note. BWC = body-worn camera.

The specific binary coding of the dependent variables in the current study was guided by several considerations. First, most previous national surveys emphasize a binary representation of “support” for BWC either directly by its initial level of measurement or indirectly in the language used to provide narrative summaries of the results (see Cato, 2016; Moore, 2015; Sousa et al., 2017). The binary coding of support versus nonsupport (i.e., unsure/oppose) in the current study provides a direct comparative basis to past research. Second, this distinction between support and nonsupport reflects a clearer line of demarcation for discussions about public attitudes toward BWCs’ usage than is true of the categories in the original Likert-type scale. Third, as shown in Table 1, each of the particular categories within the reference group (i.e., unsure/oppose) represents only a small proportion (i.e., 5–10%) of sample respondents. By combining these unsure/oppose responses into a singular reference category, a sufficient sample size is retained for analytic purposes. It is for these reasons that the particular binary coding of the dependent variables was used in this study.

Independent variables

Measures of respondent’s sociodemographic characteristics, confidence and personal involvement in social institutions, concerns about crime, police-related experiences and attitudes, and beliefs about BWCs’ benefits and costs are the major predictors of public support for BWCs in this study. The coding of the respondent’s sociodemographic attributes and other single-item measures is self-explanatory (see Table 1), but the items and procedures used to develop several composite measures require further comments.

Confidence in social institutions

Individuals’ ratings of their confidence in social institutions were measured on a 3-point scale (1 = no confidence to 3 = a great deal of confidence). The specific social institutions include (1) the federal government, (2) local governments, (3) economic institutions, (4) public schools, (5) religious institutions, (6) the family, and (7) television news and other mass media. A composite measure of institutional confidence was developed by averaging these ratings across all item. The average interitem correlation (r = .38) and other measures of the internal consistency (Cronbach’s α = .81) provide empirical justification for this composite measure.

Personal involvement in social institutions

Individuals’ level of personal involvement in social institutions was measured on a 3-point scale (1 = not involved to 3 = very involved). The eight different activities that represented personal involvement in these institutions include (1) participation in local community activities (school/housing/neighborhood programs), (2) doing work that makes you happy, (3) maintaining strong family relations, (4) helping less fortunate people in your community, (5) strengthening your religious/spiritual beliefs, (6) doing whatever it takes to make more and become financially successful, (7) trying to make positive changes in your community, and (8) trying to make yourself a better person. A composite measure of personal involvement based on the average ratings across these 8 items exhibited a high internal consistency (α = .84) and average interitem correlation (r = .40).

Crime concerns

Individuals’ concerns about victimization by particular crimes were measured on a 3-point scale (1 = not concerned to 3 = highly concerned). The specific items in this composite measure include the following crimes: (1) a street robbery/mugging, (2) residential burglary, (3) vandalism, (4) car theft, and (5) physical assault by a stranger. The 5-item composite measure had high scale reliability (α = .91) and average interitem correlation (r = .68).

Police effectiveness

Using a 5-point scale (1 = poor to 5 = excellent), a composite measure of individuals’ perceptions of police effectiveness was developed from the average ratings of their local police in the following areas: (1) working together with residents to solve local problems, (2) preventing crime in your neighborhood, (3) reducing your fear of crime, (4) arresting/catching criminals, and (5) overall rating of your local police. This 5-item composite measure exhibited high scale reliability (α = .87) and average interitem correlation (r = .57).

Procedural justice

Similar to its measurement in previous research (Hinds & Murphy, 2007; Reisig, Bratton, & Gertz, 2007), individuals’ perceptions of procedural justice were indicated by their level of agreement with the following statements about their local police: (1) treat citizens with dignity and respect, (2) treat people fairly, (3) take time to listen to people, (4) provide the same quality of service to all people, (5) make decisions only on the facts of the case, and (6) explain decisions to the people they deal with (e.g., citizens, suspects). The average ratings of these 6 items on a 5-point Likert-type scale (1 = strongly disagree to 5 = strongly agree) were used to construct a composite measure of procedural justice (α = .88; average interitem correlation = .57).

Police legitimacy

Consistent with its application in previous research (see Heen et al., 2017; Tyler, 2003), individuals’ perceptions of police legitimacy were measured by their level of agreement with the following statements: (1) people’s basic rights are well protected by police, (2) the police can be trusted to make decisions that are right for your community, (3) you should accept police decisions even when you think they are wrong, (4) the police have the same sense of right and wrong that I do, and (5) I have great respect for the police. The average ratings of these 5 items on a 5-point Likert-type scale (1 = strongly disagree to 5 = strongly agree) were used to construct a composite measure of police legitimacy. This composite measure exhibited a high internal consistency (α = .79) and average interitem correlation (r = .44).

Analytic Approach

Three analytic approaches addressed the research questions underlying this study. First, the magnitude of public support for BWCs was examined across the different domains of police work. Second, logistic regression models assessed the net effects of various sociodemographic characteristics, police-related attitudes, and beliefs about BWCs’ consequences on public support for this technology across different areas of police work. Third, conjunctive analysis was used to explore (1) the moderating effect of attitudes about social institutions and personal involvement in them on how police-related factors influence support for BWCs and (2) the nature and magnitude of interrelationships among these variables that affect the different causal pathways that lead to high and low levels of support for this technology.

Results

BWC Support Across Areas of Police Work

Consistent with estimates in earlier years using similar online sampling designs (e.g., Sousa et al., 2015), national public support for BWCs in police work remains high over the 2017–2018 time period. As shown in Table 1, about 83% of the survey respondents expressed support for the use of BWCs “in all areas of police work,” and this level of support was consistent across each of the three waves of data collection. Support for BWCs was stronger than the overall rate within specific areas of police work (e.g., 90% for crime scene investigations, 89% for crowd management). Thus, despite recent increases in media scrutiny of BWC footage and its consequences (i.e., benefits and costs), these national findings indicate that public support for this technology remains incredibly strong.

Predictors of Support for BWCs

Table 2 displays the odds ratios (ORs) and confidence intervals from a series of logistic regression analyses of the predictors of BWCs’ support in general and specific areas of police work. Although some differences are found in specific domains, the direction and magnitude of the net effects for many of these variables are largely consistent across the estimated models.

Table 2.

Logistic Regression Analysis of Predictors of Body-Worn Cameras (BWC) Support in Different Areas of Police Work.

Variables	All Areas of Police Work	Routine Traffic Stops	Neighborhood	Crime Scene Activity	Crowd Monitoring	Interview
	All Areas of Police Work	Routine Traffic Stops	Local Patrol	Crime Scene Activity	Crowd Monitoring	Crime Victims
	Odds Ratio (95% CI)	Odds Ratio (95% CI)	Odds Ratio (95% CI)	Odds Ratio (95% CI)	Odds Ratio (95% CI)	Odds Ratio (95% CI)
Male	1.11	0.63*	0.81	0.77	0.88	1.08
Male	[0.91, 1.38]	[0.50, 0.81]	[0.66, 1.01]	[0.59, 1.02]	[0.68, 1.15]	[0.85, 1.36]
<30 Years old	0.75*	0.69*	0.61*	0.40*	0.52*	0.72*
<30 Years old	[0.57, 0.99]	[0.51, 0.94]	[0.46, 0.79]	[0.29, 0.56]	[0.38, 0.72]	[0.53, 0.98]
Hispanic	0.99	0.71*	0.57*	1.03	0.91	0.81
Hispanic	[0.75, 1.32]	[0.52, 0.97]	[0.43, 0.75]	[0.71, 1.49]	[0.64, 1.29]	[0.59, 1.11]
Black	0.95	0.84	0.67*	0.81	0.79	0.82
Black	[0.68, 1.32]	[0.58, 1.22]	[0.48, 0.93]	[0.54, 1.22]	[0.53, 1.17]	[0.57, 1.18]
<US$30,000 income	0.92	1.22	1.01	0.77	0.95	0.88
<US$30,000 income	[0.73, 1.17]	[0.93, 1.59]	[0.80, 1.28]	[0.57, 1.04]	[0.72, 1.27]	[0.68, 1.15]
College education	0.94	1.37*	1.32*	1.05	1.29	0.93
College education	[0.75, 1.17]	[1.06, 1.78]	[1.05, 1.65]	[0.78, 1.41]	[0.98, 1.71]	[0.72, 1.19]
Democrat	0.93	0.92	0.94	0.66*	0.78	0.75*
Democrat	[0.74, 1.16]	[0.72, 1.19]	[0.75, 1.18]	[0.50, 0.88]	[0.60, 1.03]	[0.59, 0.97]
Resident > 5 years	1.31*	1.46*	1.15	1.32*	1.38*	1.31*
Resident > 5 years	[1.06, 1.61]	[1.15, 1.85]	[0.93, 1.43]	[1.00, 1.73]	[1.06, 1.78]	[1.03, 1.66]
City > 1 million people	1.09	0.76*	0.99	0.70*	0.68*	0.83
City > 1 million people	[0.88, 1.35]	[0.59, 0.98]	[0.79, 1.23]	[0.52, 0.94]	[0.52, 0.90]	[0.65, 1.06]
Southern state	0.86	0.86	1.07	1.37*	1.18	1.15
Southern state	[0.70, 1.06]	[0.67, 1.09]	[0.86, 1.32]	[1.03, 1.82]	[0.90, 1.54]	[0.90, 1.47]
Confidence institutions	1.07	1.00	1.12	0.89	0.88	1.07
Confidence institutions	[0.79, 1.44]	[0.71, 1.39]	[0.83, 1.52]	[0.62, 1.30]	[0.62, 1.27]	[0.77, 1.50]
Personal involvement	1.44*	1.97*	1.34*	1.86*	1.78*	1.43*
Personal involvement	[1.12, 1.85]	[1.47, 2.63]	[1.03, 1.74]	[1.34, 2.60]	[1.30, 2.44]	[1.07, 1.91]
Crime fear/concern	1.19	1.00	1.09	0.97	1.04	1.19
Crime fear/concern	[0.99, 1.43]	[0.57, 1.09]	[0.91, 1.32]	[0.77, 1.23]	[0.84, 1.30]	[0.96, 1.46]
Live in high crime area	1.15	0.79	0.88	0.78	0.95	1.13
Live in high crime area	[0.84, 1.56]	[0.57, 1.09]	[0.65, 1.18]	[0.55, 1.12]	[0.67, 1.34]	[0.80, 1.60]
Poor police experience	0.79	0.71	0.91	0.59*	0.54*	0.79
Poor police experience	[0.57, 1.11]	[0.49, 1.02]	[0.65, 1.27]	[0.39, 0.89]	[0.36, 0.80]	[0.54, 1.15]
Police effectiveness	1.05	1.13	1.05	1.07	1.01	1.04
Police effectiveness	[0.92, 1.20]	[0.98, 1.31]	[0.92, 1.20]	[0.90, 1.27]	[0.86, 1.18]	[0.89, 1.21]
Procedural justice	1.05	1.29*	1.21*	1.03	1.38*	1.24*
Procedural justice	[0.87, 1.26]	[1.05, 1.59]	[1.01, 1.45]	[0.81, 1.30]	[1.11, 1.72]	[1.01, 1.53]
Police legitimacy	1.13	0.81*	1.05	0.87	0.76*	0.84
Police legitimacy	[0.94, 1.35]	[0.66, 0.99]	[0.87, 1.26]	[0.69, 1.10]	[0.60, 0.95]	[0.68, 1.03]
BWC reduce misconduct	2.12*	2.62*	2.08*	3.29*	3.60*	2.53*
BWC reduce misconduct	[1.68, 2.68]	[2.01, 3.41]	[1.64, 2.64]	[2.43, 4.46]	[2.70, 4.80]	[1.94, 3.28]
BWC violates privacy	0.62*	0.51*	0.74*	0.63*	0.58*	0.69*
BWC violates privacy	[0.48, 0.80]	[0.38, 0.67]	[0.57, 0.95]	[0.45,0. 87]	[0.42, 0.80]	[0.52, 0.93]
BWC accurate account	3.32*	4.02*	3.12*	3.96*	3.43*	3.18*
BWC accurate account	[2.64, 5.19]	[3.10, 5.23]	[2.47, 3.95]	[2.94, 5.34]	[2.59, 4.55]	[2.45, 4.12]
Mean support (%)	83.1	86.9	83.7	90.4	89.2	87.8
N	3,051	3,051	3,051	3,051	3,051	3,051
Nagelkerke R ²	.20*	.26*	.20*	.26*	.26*	.19*

Note. CI = confidence interval.

*p < .05.

Based on the relative size of their ORs and significance, individuals’ beliefs of the benefits and social costs of BWCs have the strongest impact on support for this technology across the different areas of police work. In particular, the relative odds of BWCs’ support across policing areas are over 2 times higher when this technology is viewed as either reducing misconduct or providing an accurate account of the events that occurred in police–citizen encounters. In contrast, beliefs that this technology violates the privacy of crime suspects are linked to substantially lower odds of support for BWCs across each domain. Among the other predictor variables in these analyses, individuals with greater personal involvement in social institutions (ORs ranging from 1.34 to 1.97) were found to have significantly higher support for use of this technology across all areas of police work. Across these different areas, persons under 30 years old (ORs ranging from 0.75 to 0.40) were significantly less likely to support BWCs than older persons.

Aside from these significant predictors of support for BWCs, Table 2 also reveals that many sociodemographic attributes and police-related factors do not have significant direct effects on these public attitudes. For example, individuals’ support for BWCs across different domains is largely independent of their gender, income, regional location, confidence in social institutions, level of concern about crime, and whether they live in a high or low crime area. However, some racial and ethnic differences were observed in these public attitudes (e.g., both Black and Hispanic respondents were less supportive than White respondents of BWC usage in local neighborhood patrol). BWCs’ support in these analyses is also typically unaffected by differences in individuals’ experiences with police and views about their effectiveness and police legitimacy. However, in most areas of police work in this study, there was significantly greater support for BWCs among individuals with more positive beliefs about procedural fairness (i.e., significant OR [p < .05] greater than 1.20 for high procedural justice in four of the six areas of police work).

Conjunctive Analysis of Moderating Effects and Contextual Variability

One possible explanation for the largely null effects of individuals’ beliefs about police legitimacy and effectiveness on public support of BWCs is that their impact is context-specific and moderated by other factors. For example, when individuals have high confidence in social institutions and personal involvement in them, the relevance of police-related factors on support for BWCs may be augmented by the more generic feeling of overall institutional trust that emanates from personal attachments to social institutions. In contrast, when institutional confidence and involvement are low, differences in police-related attitudes may have less salience on support for this technology because the impact of these attitudes may be nullified by the individuals’ higher level of institutional and personal anomie. Under this interpretative framework, the presence of institutional confidence and involvement may either enhance or nullify the causal significance of police-related attitudes on support for BWCs.

To empirically evaluate the nature and magnitude of these types of moderating and context-specific effects, a conjunctive analysis of support for BWCs was performed on a subset of the major predictor variables in this study. In particular, we explored the interrelationships between individuals’ ratings on each measure of BWCs’ support and the following predictor variables:² (1) BWC reduces misconduct (0 = no/unsure, 1 = yes), (2) confidence in social institutions (0 = low, 1 = high), (3) personal involvement in social institutions (0 = low, 1 = high), (4) police legitimacy (0 = low, 1 = high), and (5) police effectiveness (0 = low, 1 = high). When these five variables are considered simultaneously, a total of 32 distinct contexts for BWC support are possible (i.e., 2 [levels] × 5 [variables] = 2⁵ = 32 contexts).

Table 3 displays the contextual variability across these 32 different “causal” pathways for predicting the relative prevalence of general support for BWCs in all areas of police work.³ As indicated by the sequencing of the profile numbers, these unique conjunctive profiles are ranked from their highest to lowest levels of support. The shaded areas in this table represent the conjunctive profiles that fall within the top 25% of all observed profiles on BWCs’ support (Profiles #1–8) and the bottom 25% for support (Profiles #24–32).

Table 3.

Conjunctive Profiles of Contextual Variability in Body-Worn Camera (BWC) Support.

Profile Number	Reduce Police Misconduct	High Confidence in Institutions	High Personal Involvement	High Police Legitimacy	High Police Effective	N of Cases	Percent BWC Support in All Areas of Policing (%)
1	1	0	1	1	0	37	96
2	1	1	0	0	1	33	94
3	1	1	0	1	0	43	94
4	1	1	1	1	1	833	93
5	1	1	1	0	1	63	91
6	1	1	0	1	1	270	90
7	1	0	1	0	0	95	89
8	1	0	0	0	1	44	88
9	1	0	1	1	1	218	86
10	1	0	0	1	1	147	85
11	1	1	1	1	0	105	84
12	1	1	0	0	0	23	84
13	1	1	1	0	0	92	84
14	1	0	0	1	0	51	84
15	0	0	1	1	1	56	79
16	1	0	1	0	1	34	79
17	1	0	0	0	0	119	78
18	0	0	0	1	0	32	78
19	0	0	0	1	1	85	78
20	0	1	1	1	1	195	77
21	0	1	0	1	1	89	75
22	0	0	0	0	1	21	75
23	0	0	1	0	1	21	72
24	0	0	1	0	0	52	71
25	0	1	0	1	0	21	67
26	0	1	1	1	0	37	66
27	0	0	1	1	0	10	63
28	0	1	0	0	1	21	59
29	0	0	0	0	0	98	52
30	0	1	1	0	1	34	49
31	0	1	1	0	0	39	49
32	0	1	0	0	0	33	39

Note. Cell entries represent presence (1) or absence (0) of attribute. Shade areas indicate profile within the top 25% and bottom 25% percentile rankings.

A visual inspection of Table 3 reveals several patterns about the contextual variability in support for BWCs. First, although the average support across the entire sample is high (83%), the level of support exhibits substantial variability across these conjunctive profiles. Nearly universal support for BWCs is found in some contexts (e.g., 96% in Profile #1), but support is much lower in other contexts (e.g., 39% in Profile #32). Second, the most frequent contexts for support (i.e., the top 25% of profiles) always involve persons who believe that BWCs reduce police use of excessive force and other misconduct (100% among Profiles #1–8) and individuals with either high confidence in social institutions (62% of these profiles) or high ratings on police effectiveness (62% of these profiles). Among the lowest quartile of contexts for support of this technology (Profiles #24–32), the absence of beliefs about BWCs reducing police use of excessive force and other misconduct was always found (100%) in these profiles, and ratings of low police effectiveness were also present in most of them (75% [six of eight profiles]). Third, the impact of individuals’ level of personal involvement in institutions and their belief about police legitimacy was more evenly dispersed across contexts with different levels of support for BWCs. For these predictor variables, their importance in explaining the level of support was far more context-specific and contingent upon the other set of attributes contained in the particular conjunctive profile.

To evaluate context-specific and moderating effects, conjunctive analysis employs a case-comparative method to assess variability in the main effect of each predictor variable on an outcome variable across different conjunctive profiles (see, for applications, Hart, 2014; Miethe et al., 2008).⁴ The nature of these effects across conjunctive profiles is often visually represented by boxplots (Hart, Miethe, & Regoeczi, 2014). For example, if the impact of police-related attitudes (e.g., beliefs about police legitimacy, police effectiveness) on BWC support is largely invariant across contexts, the boxplot of group differences should be visually represented by a small box width (i.e., small interquartile range), short “whiskers” (i.e., deviations within a 1.5 width of the box), and no extreme outliers falling outside this range. In contrast, the greater the context-specific effects of any variable, the larger the magnitude of these moderating effects, and the wider the area covered by the interquartile range (i.e., the “box”) and the deviations from it. Alternatively, a pattern of strong and invariant main effects would be visually represented by boxplots that (1) have a small box width and no deviations/outliers, (2) do not intersect the positive and negative scale values of the comparative axis, and (3) depart substantially from the 0 point on this axis.

When these principles are empirically assessed in the current study, the boxplots in Figure 1 reveal several patterns about the average effect of each variable and their contextual variability on support for BWCs in the different domains of police work. First, the strongest “main effects” on BWCs’ support across each area of police work involve individuals’ beliefs about BWCs reducing police use of excessive force and other misconduct. For this variable, the average percentage differences across the pairs of conjunctive profiles are relatively large (i.e., median differences in the range of 15–22%), and the interquartile range of their context-specific effects is always positive values (i.e., values greater than 0 for the middle 50% of the profile pairs). Second, the main effects of the other predictor variables are appreciably lower in magnitude and exhibit far greater contextual variability (e.g., their interquartile ranges cover both positive and negative scale values). For these variables, a primary focus on main effects (as shown in Table 2) may seriously distort their actual importance in explaining variation in support for BWCs in police work.

Figure 1.

Visual display of contextual variability in main effects of key variables in BWC support. Force = BWC reduces use of excessive force and other misconduct; iconf = confidence in social institutions; involv = personal involvement in social institutions; legit = ratings of police legitimacy; peffe = ratings of police effectiveness; BWC = body-worn camera. See Table 1 and text for coding and measurement of these variables.

Further investigation of the contextual variability in Figure 1 reveals the primary sources of these moderating effects. In particular, the largest impact of differences in public views about police effectiveness on the prevalence of BWCs’ support is found among individuals with (a) low confidence in social institutions, (b) low personal involvement in them, (c) low ratings of police legitimacy, and (d) negative views about BWCs reducing excessive force. Among persons with high levels of confidence and involvement in these institutions, beliefs about police effectiveness have only a minor impact on support for BWCs in each area of police work. In contrast, differences in views about police legitimacy have the largest impact on increasing BWCs’ support among those with negative attitudes about this technology reducing excessive force. However, high ratings of police legitimacy are linked to lower support for BWCs among individuals with lower ratings of police effectiveness, low confidence in institutions, and high personal involvement in them. Thus, differences in personal involvement and confidence in social institutions moderate the causal significance of police-related factors by either enhancing or mitigating their influence on public support for BWCs.

Discussion and Conclusions

Based on large national samples, the current study investigated two research questions about (1) the sociodemographic attributes and police-related factors associated with public attitudes toward BWCs and (2) the presence of moderating effects that may account for the wide contextual variability in their support. A review of the study’s results, their limitations, and a discussion of their implications for future research is summarized below.

Summary of Findings

Public support for BWCs in this national survey was consistently high over the 2017–2018 study period and different areas of police work. The vast majority (83%) of adult respondents indicated that they supported the BWCs’ usage in all areas of police work, with even greater support found in particular areas (e.g., 90% for use in crime scene investigations, 89% for crowd monitoring). This high level of support for BWCs is consistent with the findings from earlier national surveys (e.g., Sousa et al., 2015).

Based on the estimation of logistic regression models, public support for the use of BWCs in different areas of policing was strongly influenced by specific beliefs about the consequences of their usage (e.g., reducing police misconduct, privacy issues, providing accurate video accounts of incidents) and individuals’ level of personal involvement in social institutions. Most measures of individuals’ sociodemographic characteristic and police-related factors (e.g., prior experiences with police, beliefs about police legitimacy, and effectiveness) had no significant net effect on these public attitudes toward BWCs. However, the conjunctive analysis of a subset of these predictor variables revealed widespread contextual variability in the observed effects of many of these variables. These results provide direct evidence of how the absence of significant main effects for particular variables may be attributed to the substantial moderating effects that both enhance and mitigate their particular influence across different contexts.

Study Limitations

The survey data collected and analyzed in the current study were based on national online sampling frame. While online sources are increasingly used in social science research, there are well-known sources of sampling bias associated with online surveys. These include sampling biases due to the “digital divide” (i.e., the underrepresentation of Black and Hispanic respondents, persons over 65 years old, persons with less formal education) and the tendency for greater participation in surveys by female respondents and those in the midrange of annual household income (Pew, 2015). Although the stratified sampling design and poststratification weighting used in the current study may help equate the sample and population distributions, sampling bias remains a problem with all survey methods because nonrespondents are likely to be different than survey participants on a variety of grounds (e.g., interest in the topic, availability/contact rates).

For studying the moderating effects of confidence and involvement in social institutions, this online sample is also limited because people who may be the most socially disadvantaged and alienated from these institutions (and those with lower incomes and fewer attachments to these institutions) are often not represented in these samples. Due to this suppression of sample variability, both the direct effects and moderating effects of these variables on BWC support and its predictors may have been attenuated.

Implications for Future Research

The results of this study have several implications for future research on public attitudes about BWCs and their usage in contemporary policing activities. These areas of future research are summarized below.

Contextual factors associated with low BWC support

Despite the general finding of wide public support, this study’s conjunctive analysis revealed (1) high levels of contextual variability in the level of support for BWCs and (2) the presence of several contexts for which the majority of respondents do not support the usage of this technology (see Profiles #30–#32 in Table 3). The unique signatures of these particular conjunctive profiles of low BWC support are that they always involve individuals who (1) do not believe that BWCs reduce police misconduct, (2) have low personal involvement in activities that support social institutions, and (3) have low beliefs about police legitimacy. Based on these findings, any police or public policies that are designed to further enhance public support for BWCs would benefit from consideration of these three factors, individually and collectively, because they represent the major sources of current public opposition to BWCs.

Weak direct effects of police-related factors on support for BWCs

Previous research indicates that police-related factors have strong direct effects on public support for visual surveillance technology (e.g., drones) and citizen’s overall cooperation and satisfaction with police (Heen et al., 2017; Hinds & Murphy, 2007; Tyler, 2003, 2005). However, within the current study, measures of these police-related factors (e.g., beliefs about police legitimacy, police effectiveness, personal experiences with police) had little direct impact on the level of public support for BWCs (see Table 2).

As revealed by the conjunctive analysis, the primary explanation for these discrepant findings is that the magnitude of the main effects for measures of police legitimacy and police effectiveness is distorted by various moderating effects. In particular, the impact of these police-related factors is dependent upon individuals’ level of confidence and personal involvement in social institutions in highly contextual ways. For example, among individuals with high confidence and personal involvement in social institutions, high ratings of police legitimacy can be found to either enhance, reduce, or have no discernable effect on BWC support (see Figure 1). These types of context-specific effects are also found for the impact of police effectiveness on BWC support. Given the identification of these types of moderating effects, an important question for future research involves the assessment of the possible explanations for this contextual variability in the salience of police-related factors on public support for the use of this technology.

Perceived benefits and costs of BWC usage

Over the last decade, BWCs have been widely touted as a technological remedy for many issues in contemporary policing (e.g., increasing transparency, enhancing real-time evidence gathering, and moderating citizen and police behavior by reducing use of force, citizen complaints, and lawsuits). Privacy issues and concerns about lower citizen cooperation with recorded conversations represent some of the adverse consequences that may derive from the police use of this technology. However, the high level of public support for the use of BWCs in this study and other national surveys suggests that the perceived benefits of this technology currently exceed their perceived costs.

Given the wide availability of alternative audio–video technology (e.g., high-resolution cell phones, aerial drones) and changing sociopolitical views about the necessity of this technology, whether this benefit-to-cost ratio will remain over the next decade is an important question to monitor in future research. Emerging issues about privacy protections, increased awareness of distortions of images from BWCs (e.g., due to the angle and physical context of the video footage), greater scrutiny of media coverage of this evidence in news accounts, and the uncertainty of the probative value of this technology for litigation are some areas of concern that may weaken public support for BWCs. Once the informative capacity of BWCs becomes better understood by further research and critical discourse, beliefs about their particular benefits/risks may change and alter the public’s general support for this technology. However, answers to these questions will be unavailable without future research to monitor patterns of change and stability in public attitudes about BWCs over time.

Footnotes

Appendix

Table A1.

Steps in the Analysis of Contextual Variability of a Variable’s Main Effects in Conjunctive Analysis.

Column 1	Column 2	Column 3	Column 4	Column 5	Column 6	Column 7
Profile	Predictor Variable 1	Predictor Variable 2	Predictor Variable 3	Focal Predictor 4	Outcome Variable 1 (%)	Difference H–L
1	1	1	1	1	94	3
2	1	1	1	0	91
3	1	1	0	1	84	0
4	1	1	0	0	84
5	1	0	1	1	70	−24
6	1	0	1	0	94
7	1	0	0	1	68	10
8	1	0	0	0	58
9	0	1	1	1	95	15
10	0	1	1	0	80
11	0	1	0	1	88	6
12	0	1	0	0	82
13	0	0	1	1	75	−7
14	0	0	1	0	82
15	0	0	0	1	70	45
16	0	0	0	0	25

Note. (1) Identify a set of predictor (Columns 2–5) and outcome variables (Column 6) to develop unique conjunctive profiles. (2) Aggregate scores on the outcome variable for each profile to derive their respective mean values (Column 6). (3) Sort data matrix by the focal predictor (Column 5) and then successively for the remaining predictor variables. (4) Compute differences between high (1) and low (0) values of focal predictor on outcome variable (Column 7). For example, the differences of −24 in Column 7 is found by subtracting the percentage value on the outcome variable (70%) associated with the high value of the focal predictor (the value “1” in Profile #5, Column 5) from the percentage value (94%) associated with the low value of the focal predictor (the value “0” in Profile #6, Column 5). The numerical value on the outcome variable (−24%) indicates that high values on the focal variable are associated with a 24 percentage point lower prevalence of the outcome variable than is true of the same profile with low values on the focal predictor. (5) These differences across pairs of profiles represent the focal variable’s “main effect” across each context. (6) Repeat Steps 3 and 4 for each of the other predictor variables (Predictors 1–3). (7) Examine the nature of these differences across profiles and visually display their variability in a boxplot. (8) For the inclusion of additional predictor variables, expand the data matrix to include all possible combinations of the levels of these variables considered simultaneously and repeat Steps 2–7.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Support for this research was provided by a grant from the National Science Foundation (NSF #1625808).

Notes

References

Ariel

Farrar

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. doi:10.1007/s10940-014-9236-3

Cato. (2016, January 5). Cato/YouTube Poll: 92% Support police body camera, 55% willing to pay more in taxes to equip local police: Results from the Cato Institute Criminal Justice Survey. Retrieved from https://www.cato.org/survey-reports/catoyougov

Chow

G. C.

(1960). Tests of equality between sets of coefficients in two linear regressions. Econometrica, 28, 591–605.

Crow

M. S.

Snyder

J. A.

Crichlow

V. J.

Smykla

J. O.

(2017). Community perceptions of police body-worn cameras: The impact of views on fairness, fear, performance, and privacy. Criminal Justice and Behavior, 44, 589–610.

Habermas

(1975). Legitimation crisis. Boston, MA: Beacon Press.

Hart

T. C.

(2014). Conjunctive analysis of case configurations. JDiBrief Series. London, England: UCL Jill Dando Institute of Security and Crime Science. Retrieved from www.jdibrief.com

Hart

T. C.

Miethe

T. D.

(2014). Street robbery and public bus stops: A case study of activity nodes and situational risk. Security Journal, 27, 180–193.

Hart

T. C.

Miethe

T. D.

Regoeczi

W. C.

(2014). Contextualizing sentencing disparities: Using conjunctive analysis of case configurations to identify patterns of variability. Criminal Justice Studies, 27, 344–361.

Heen

M. S. J.

Lieberman

J. D.

Miethe

T. D.

(2017). The thin blue line meets the big blue sky: Perceptions of police legitimacy and public attitudes towards aerial drones. Criminal Justice Review, 31, 18–37.

10.

Hinds

Murphy

(2007). Public satisfaction with police: Using procedural justice to improve police legitimacy. Australian & New Zealand Journal of Criminology, 40, 27–42. doi:10.1375/acri.40.1.27

11.

Kim

J. S.

Kaye

Wright

L. K.

(2001). Moderating and mediating effects in causal models. Issues in Mental Health Nursing, 22, 63–75.

12.

LaFree

(1998). Losing legitimacy: Street crime and the decline of social institutions in America. Boulder, CO: Westview Press.

13.

Lawrence

Peterson

Thompson

(2018). Community views of Milwaukee’s body-worn camera program. Washington, DC: Urban Institute.

14.

Merton

R. K.

(1938). Social structure and anomie. American Sociological Review, 3, 672–682.

15.

Messner

S. F.

Rosenfeld

(2013). Crime and the American dream (5th ed.). Belmont, CA: Wadsworth-Centage Learning.

16.

Miethe

T. D.

Hart

T. C.

Regoeczi

W. C.

(2008). The conjunctive analysis of case configurations: An exploratory method for discrete multivariate analyses of crime data. Journal of Quantitative Criminology, 24, 227–241.

17.

Miller

Tolliver

, & Police Executive Research Forum. (2014). Implementing a body-worn camera program: Recommendations and lessons learned. Washington, DC: Office of Community Oriented Policing Services.

18.

Moore

(2015, May 7). Overwhelming support for public body cameras. Retrieved from https://today.yougov.com/topics/politics/articles-reports/2015/05/07/body-cams

19.

Norman

(2016, June 13). American’s confidence in institutions stays low. Retrieved from https://news.gallup.com/poll/192581/americans-confidence-institutions-stays-low.aspx

20.

Paternoster

Brame

Mazeroll

Piquero

(1998). Using the correct statistical test for the equality of regression coefficients. Criminology, 36, 859–866.

21.

Pew Research Center. (2015). American’s internet access: 2002–2015. Retrieved from http://www.pewinternet.org/2015/06/26/americans-internet-access-2000-2015

22.

Pew Research Center. (2017, January). Behind the badge: Amid protests and calls for reform, how police view their jobs, key issues and recent fatal encounters between blacks and police. Retrieved from http://www.pewsocialtrends.org/2017/01/11/behind-the-badge/

23.

Reisig

M. D.

Bratton

Gertz

M. G.

(2007). The construct validity and refinement of process-based policing measures. Criminal Justice and Behavior, 34, 1005–1028.

24.

Rennison

C. M.

DeKeseredy

W. S.

(2017). Situational contexts of rural violence: A comparison of male and female perpetration. Journal of Contemporary Criminal Justice, 33, 189–206. doi:10.1177/1043986216688813

25.

Sakiyama

(2017). The balance between privacy and safety in police UAV use: The power of threat and its effect on people’s receptivity (Unpublished PhD dissertation). Las Vegas, US: Greenspun College of Urban Affairs, University of Nevada.

26.

Sousa

W. H.

Miethe

T. D.

Sakiyama

M. A.

(2015). Body worn cameras on police: Results from a national survey of public attitudes. Las Vegas, NV: Research in Brief, Center for Crime and Justice Policy.

27.

Sousa

W. H.

Miethe

T. D.

Sakiyama

M. A.

(2017). Inconsistencies in public opinion of body-worn cameras on police: Transparency, trust, and improved police-citizen relationships. Policing: A Journal of Policy and Practice, 11, 1–9. doi:10.1093/police/pax015

28.

Tyler

T. R.

(2003). Procedural justice, legitimacy, and the effective rule of law. In Tonry

(Ed.), Crime and justice—A review of research (Vol. 30, pp. 431–505). Chicago, IL: University of Chicago Press.

29.

Tyler

T. R.

(2005). Policing in black and white: Ethnic group differences in trust and confidence in the police. Police Quarterly, 8, 322–342. doi:10.1177/1098611104271105

30.

Venger

(2017). Media freedom and the control of nation’s sociolegal problems of governance: A conjunctive analysis of asymmetric effects and multiple causal pathways. Journal of Contemporary Criminal Justice, 33, 173–188. doi:10.1177/1043986216688821

31.

White

(2014). Police officers body-worn cameras: Assessing the evidence. Washington, DC: Office of Community Oriented Policing Services.

32.

White

M. D.

Todak

Gaub

J. E.

(2017). Assessing citizen perceptions of body-worn cameras after encounters with police. Policing: An International Journal of Police Strategies and Management, 40, 689–703.

Profile Number	Reduce Police Misconduct	High Confidence in Institutions	High Personal Involvement	High Police Legitimacy	High Police Effective	N of Cases	Percent BWC Support in All Areas of Policing (%)
1	1	0	1	1	0	37	96
2	1	1	0	0	1	33	94
3	1	1	0	1	0	43	94
4	1	1	1	1	1	833	93
5	1	1	1	0	1	63	91
6	1	1	0	1	1	270	90
7	1	0	1	0	0	95	89
8	1	0	0	0	1	44	88
9	1	0	1	1	1	218	86
10	1	0	0	1	1	147	85
11	1	1	1	1	0	105	84
12	1	1	0	0	0	23	84
13	1	1	1	0	0	92	84
14	1	0	0	1	0	51	84
15	0	0	1	1	1	56	79
16	1	0	1	0	1	34	79
17	1	0	0	0	0	119	78
18	0	0	0	1	0	32	78
19	0	0	0	1	1	85	78
20	0	1	1	1	1	195	77
21	0	1	0	1	1	89	75
22	0	0	0	0	1	21	75
23	0	0	1	0	1	21	72
24	0	0	1	0	0	52	71
25	0	1	0	1	0	21	67
26	0	1	1	1	0	37	66
27	0	0	1	1	0	10	63
28	0	1	0	0	1	21	59
29	0	0	0	0	0	98	52
30	0	1	1	0	1	34	49
31	0	1	1	0	0	39	49
32	0	1	0	0	0	33	39

Profile Number	Reduce Police Misconduct	High Confidence in Institutions	High Personal Involvement	High Police Legitimacy	High Police Effective	N of Cases	Percent BWC Support in All Areas of Policing (%)
1	1	0	1	1	0	37	96
2	1	1	0	0	1	33	94
3	1	1	0	1	0	43	94
4	1	1	1	1	1	833	93
5	1	1	1	0	1	63	91
6	1	1	0	1	1	270	90
7	1	0	1	0	0	95	89
8	1	0	0	0	1	44	88
9	1	0	1	1	1	218	86
10	1	0	0	1	1	147	85
11	1	1	1	1	0	105	84
12	1	1	0	0	0	23	84
13	1	1	1	0	0	92	84
14	1	0	0	1	0	51	84
15	0	0	1	1	1	56	79
16	1	0	1	0	1	34	79
17	1	0	0	0	0	119	78
18	0	0	0	1	0	32	78
19	0	0	0	1	1	85	78
20	0	1	1	1	1	195	77
21	0	1	0	1	1	89	75
22	0	0	0	0	1	21	75
23	0	0	1	0	1	21	72
24	0	0	1	0	0	52	71
25	0	1	0	1	0	21	67
26	0	1	1	1	0	37	66
27	0	0	1	1	0	10	63
28	0	1	0	0	1	21	59
29	0	0	0	0	0	98	52
30	0	1	1	0	1	34	49
31	0	1	1	0	0	39	49
32	0	1	0	0	0	33	39

Profile Number	Reduce Police Misconduct	High Confidence in Institutions	High Personal Involvement	High Police Legitimacy	High Police Effective	N of Cases	Percent BWC Support in All Areas of Policing (%)
1	1	0	1	1	0	37	96
2	1	1	0	0	1	33	94
3	1	1	0	1	0	43	94
4	1	1	1	1	1	833	93
5	1	1	1	0	1	63	91
6	1	1	0	1	1	270	90
7	1	0	1	0	0	95	89
8	1	0	0	0	1	44	88
9	1	0	1	1	1	218	86
10	1	0	0	1	1	147	85
11	1	1	1	1	0	105	84
12	1	1	0	0	0	23	84
13	1	1	1	0	0	92	84
14	1	0	0	1	0	51	84
15	0	0	1	1	1	56	79
16	1	0	1	0	1	34	79
17	1	0	0	0	0	119	78
18	0	0	0	1	0	32	78
19	0	0	0	1	1	85	78
20	0	1	1	1	1	195	77
21	0	1	0	1	1	89	75
22	0	0	0	0	1	21	75
23	0	0	1	0	1	21	72
24	0	0	1	0	0	52	71
25	0	1	0	1	0	21	67
26	0	1	1	1	0	37	66
27	0	0	1	1	0	10	63
28	0	1	0	0	1	21	59
29	0	0	0	0	0	98	52
30	0	1	1	0	1	34	49
31	0	1	1	0	0	39	49
32	0	1	0	0	0	33	39