How police agency diversity, policies, and outcomes shape citizen trust and willingness to engage

Abstract

While widespread agreement that policing in the United States needs to be reformed arose in the summer of 2020, little consensus about specific reforms was reached. A common theme that arose, however, is a general lack of trust in the police. One response has been to increase agency diversity in terms of both officer race and officer gender. However, important questions exist about when – and what type(s) of – diversity shape citizen trust in and willingness to cooperate with the police – especially when considered in conjunction with agency performance and policies. To answer these questions, we use two conjoint experiments to evaluate whether citizens consider diversity when evaluating police agencies. We find that while both racial and gender diversity can influence public evaluations sometimes, these effects tend to emerge in the context of only the most (least) diverse institutions and are muted compared to the effects of agency policies and performance.

Full text

1Department of Political Science, University of
South Carolina, Columbia, South Carolina, USA
2School of Public Affairs and Administration,
Rutgers University - Newark, Newark,
New Jersey, USA
3School of Public Policy, University of
Massachusetts, Amherst, Massachusetts, USA
Correspondence
Kelsey Shoub, University of Massachusetts,
Amherst, MA, USA.
Email: kshoub@umass.edu; kelsey.shoub@gmail.com
Abstract
While widespread agreement that policing in the United States
needs to be reformed arose in the summer of 2020, little consen-
sus about specific reforms was reached. A common theme that
arose, however, is a general lack of trust in the police. One
response has been to increase agency diversity in terms of both
officer race and officer gender. However, important questions
exist about when – and what type(s) of – diversity shape citizen
trust in and willingness to cooperate with the police – especially
when considered in conjunction with agency performance and
policies. To answer these questions, we use two conjoint experi-
ments to evaluate whether citizens consider diversity when eval-
uating police agencies. We find that while both racial and gender
diversity can influence public evaluations sometimes, these
effects tend to emerge in the context of only the most (least)
diverse institutions and are muted compared to the effects of
agency policies and performance.
KEYWORDS
descriptive representation, policing, representative bureaucracy, symbolic
representation
ORIGINAL ARTICLE
How police agency diversity, policies, and outcomes
shape citizen trust and willingness to engage
Katelyn E. Stauffer1 | Miyeon Song2 | Kelsey Shoub3
DOI: 10.1111/psj.12479
Received: 30 March 2021 Revised: 29 June 2022 Accepted: 23 July 2022
How citizens view the police is vital for both democratic accountability generally and perceptions of
police legitimacy specifically. Yet, distrust of the police in the United States seems to be increasing –
or at least be more visible: we have seen the spread of protests against police violence, the growth of
movements advocating for structural change, and new lows in Americans' trust of the police. Further,
public perceptions of the police are even bleaker among individuals belonging to historically marginalized
communities (Gill, 2020; Horowitz et al., 2019; Oritz, 2020). The public's distrust of the police has serious
consequences and presents a crisis of legitimacy for both law enforcement and the broader justice system.
In contexts where trust is low, community members are less willing to engage or cooperate with the police
and are less likely to initiate contact in cases where law enforcement may be needed (e.g., Baumgartner
et al., 2018; Epp et al., 2014), which in turn hampers the ability of officers and agencies to carry out their
duties (Horowitz et al., 2019). Despite widespread agreement that the relationship between the public and
the police is in a state of disarray, there is no such agreement on how to repair this relationship.
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and
reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
© 2022 The Authors. Policy Studies Journal published by Wiley Periodicals LLC on behalf of Policy Studies Organization.
Policy Stud J. 2023;51:929–950. wileyonlinelibrary.com/journal/psj 929

While there is not consensus on the specific changes needed, increased diversity on police forces has
been frequently proposed as one way to reduce discrimination, produce better outcomes “by enhanc-
ing the legitimacy of the police force within communities” (Nicholson-Crotty et al., 2017, p. 206), and
increase trust in the police. Here, we examine whether increased diversity in police forces changes the
ways that citizens perceive and interact with these institutions. Building on theories of representation
from political science and public administration, we test whether increasing descriptive representation on
the force (i.e., the proportion of officers who belong to historically underrepresented groups) can influ-
ence how citizens – particularly those sharing these characteristics – view the agency.
While the literature on representative bureaucracy has implications for substantive representation in
policing – i.e., whether officers who are racial and ethnic minorities or women will do their jobs differ-
ently from White or male officers in a way that reduces existing discrimination and inequity (Baumgartner
et al., 2021; Meier & Nicholson-Crotty, 2006; Shoub et al., 2021; Wilkins & Williams, 2008, 2009) –
we focus our attention in this study on symbolic representation, or the feelings and behaviors evoked
among members of the public in response to the inclusion of women and racial and ethnic minorities
(Lawless, 2004; Pitkin, 1967; Schwindt-Bayer & Mishler, 2005). Specifically, we are interested in whether
descriptive representation – i.e., the extent to which the bureaucracy mirrors the society that it serves
– leads to increased trust and willingness to engage with the police (e.g., Riccucci et al., 2016, 2018;
Theobald & Haider-Markel, 2009).
Previous studies provide a range of insights on officer diversity. Important questions, however, remain
about what types of diversity influence citizen levels of trust in the police and willingness to cooperate,
how changes in levels of diversity compare to agency performance, and how historically dominant groups
may respond to varying levels of diversity. To address this, we examine these different factors jointly
and contribute to the existing literature in three ways. First, while previous studies have advanced our
understanding by focusing on one dimension of diversity at a time and drawing on either the literature
in political science or public administration, this article pulls together work from both fields to enhance
our understanding of representation by considering collective racial and gender diversity in tandem. An
important caveat to this is the small group of work focused on the intersectionality of identities to
understand the phenomena of representation (e.g., Fay et al., 2021; Gaynor & Blessett, 2022; Wright &
Headley, 2020). Here we do not adopt an intersectional approach as much reporting by governments,
organizations, and the media focus on one aspect of diversity at a time or multiple aspects in parallel.
Second, this study contributes to our understanding of representative bureaucracy theory by explic-
itly considering how historically dominant groups may respond to varying levels of racial and gender
representation (see also Keiser et al., 2021), which is important as these groups often play a critical role
in the policy process. More specifically, building upon existing theories such social dominance theory and
racial ambivalence theory, we suggest three possible scenarios for the dominant group and test how their
perceptions toward diversity differ from that of marginalized groups.
Third, this study connects the literature on representative bureaucracy to research on performance
information, as we consider the role of diversity in the context of performance metrics. While many
scholars highlight the critical role of performance information and contextual factors in shaping citizen
perceptions of accountability and satisfaction (DeHoog et al., 1990; Olsen, 2017; Porumbescu et al., 2021),
previous experiments on representation have been conducted in isolation from other agency performance
and context considerations. As citizens consider multiple pieces of information when evaluating agencies,
it is important to account for performance when testing responses to differing levels of diversity. Here
we do that, offering a more holistic assessment of how citizens view agency legitimacy and accountability.
Using two conjoint experiments, we test how racial diversity, gender diversity, the local context of the
force, department performance (e.g., felony solve rate), and department policy (e.g., resources allocated
to community policing), influence public support for the police. We find that in some contexts diversity
increases support for the police, but that these effects often operate at the margins. Women appear to only
view the police more favorably at the very highest levels of diversity. Similarly, while Black (and some-
times White) respondents reward high levels of racial diversity and punish low levels of racial diversity,
this occurs only in the context of the most and least diverse forces. Moreover, we note that compared
STAUFFER ET Al.
930

to agency performance, diversity of any kind has a relatively modest effect on opinion, indicating that
diversifying police forces alone will not be enough to restore public faith in the police.
In doing this, we contribute to the ongoing conversation about what policy changes may allow police
departments to rebuild trust in their communities. Additionally, we push toward a more nuanced under-
standing of what higher levels of diversity on the force mean for public opinion and behavior.
DESCRIPTIVE AND SYMBOLIC REPRESENTATION
Does greater diversity on police forces engender greater feelings of trust and legitimacy between these
agencies and the communities they interact with? To ground our theoretical expectations, we draw on
the representation literature from political science and the theory of representative bureaucracy from
public administration. Before continuing, we first make an important conceptual note. Where politi-
cal science has adopted the term “descriptive representation” to describe the demographic similarities
between representatives and those they represent, public administration has adopted the term “passive
representation” (Mosher, 1982). Likewise, where political science refers to the actions undertaken by
representatives on behalf of their constituents as “substantive representation,” public administration
has adopted “active representation” to describe these behaviors (Mosher, 1982). Both fields use the term
“symbolic representation” to refer to the feelings, attitudes, and behaviors that representatives evoke
from the represented (e.g., Lawless, 2004; Pitkin, 1967; Schwindt-Bayer & Mishler, 2005; Theobald &
Haider-Markel, 2009). Despite these differences in terminology, the concepts outlined across the two
fields are quite similar. In this article, we use descriptive and substantive representation.
Political science research has long been interested in how the inclusion of historically marginalized
groups in positions of power influences the way citizens view and interact with institutions (e.g., Atkeson
& Carrillo, 2007; Gay, 2002; Mansbridge, 1999; Philpot & Walton, 2007; Scherer & Curry, 2010). Who
is (and is not) included in institutions sends important signals about who “belongs” in positions of
power, how institutions will treat particular groups, and whose interests are being represented (Celis &
Mazur, 2012; Dovi, 2002; Phillips, 1995). Thus, the inclusion of historically marginalized groups – particu-
larly women and racial and ethnic minorities – is argued to foster greater levels of trust, legitimacy, and
a willingness to engage with institutions that may have previously seemed distant, indifferent, or perhaps
hostile to certain segments of the population (Clark, 2019; Clayton et al., 2018; Hayes & Hibbing, 2017;
Mansbridge, 1999; Rocha et al., 2010). In some cases, the mere belief that minoritized people are included
can be enough to evoke these feelings (Stauffer, 2021; Tate, 2003). While this research has most frequently
focused on descriptive representation among elected officials, increasingly scholars are examining these
relationships in the context of non-elected political actors as well (Badas & Stauffer, 2018, 2019; Barnes
& Taylor-Robinson, 2018; Liu & Banaszak, 2017).
Literature from public administration on representative bureaucracy likewise argues that the inclu-
sion of historically marginalized groups in bureaucratic agencies, such as the police, can engender higher
levels of support for, and satisfaction in, these agencies (e.g., Riccucci et al., 2014, 2018; Theobald &
Haider-Markel, 2009). Scholars also argue that the identities of street-level bureaucrats, such as rank
and file police officers, are critically important, as these actors are the ones who most directly engage
with citizens and play a pivotal role in shaping policy implementation on the ground (Lipsky, 2010). The
inclusion of historically marginalized groups in these positions is thought to be important for at least two
reasons. First, the inclusion of historically marginalized groups can change the outcomes produced by
agencies and their priorities (e.g., Andrews & Johnston Miller, 2013; Baumgartner et al., 2021; Hong, 2021;
Schuck, 2018; Shoub et al., 2021), particularly outcomes that affect fellow group members or are prior-
itized by them (e.g., Bradbury & Kellough, 2011; Headley & Wright, 2020; Hong, 2016, 2017; Meier
& Nicholson-Crotty, 2006; Song, 2018). Second, the inclusion of historically marginalized groups can
shape how the public perceives and interacts with these institutions (e.g., Riccucci et al., 2014; Riccucci
& Van Ryzin, 2017; Theobald & Haider-Markel, 2009). As Riccucci et al. (2018, p. 507) note the social
origins of bureaucrats can induce certain attitudes or behaviors on the part of citizens or clients with-
POlICE DIVERSITY, POlICIES, AND OUTCOMES 931

out the bureaucrat taking any action. In other words, descriptive representation can produce symbolic
outcomes not because of what the bureaucrat does, but rather who the bureaucrat is. We focus on this
aspect of representative bureaucracy here, asking whether increasing representation among police officers
induces the symbolic outcomes discussed.
WHO RESPONDS TO REPRESENTATION?
While scholars and practitioners argue that descriptive representation can foster higher levels of trust in
institutions, an equally important question is which citizens derive symbolic benefits from inclusion. In
other words, if police forces make attempts to diversify their memberships in a bid for public support,
which citizens will respond, and how? On this question, there are two broad group types to be considered:
the historically marginalized group and the historically dominant group. In the context of gender, this
would be women and men respectively, and, in the context of race, this would be non-Whites and Whites,
respectively. We argue that if members of the traditionally marginalized groups see greater representation
in a police agency then they will view that agency as more legitimate and likely to be held accountable. We
then present competing hypotheses as to how the dominant group may respond.
Response by the historically marginalized group
Historically, both the literature in political science and in public administration have examined the effects
of descriptive representation for marginalized groups. Both argue that women and racial and ethnic minor-
ities should be especially receptive to cues provided when other women or co-racial or co-ethnic indi-
viduals are included, as they are two groups historically excluded from power (Atkeson & Carrillo, 2007;
Barnes & Burchard, 2012; Lawless, 2004; Overby et al., 2005; Reingold & Harrell, 2010; Scherer &
Curry, 2010; Tate, 2003). Numerous scholars find that women's engagement with politics increases when
they are represented by other women, as do to their levels of trust and efficacy (Atkeson, 2003; Atkeson &
Carrillo, 2007; Burns et al., 2001; Clayton et al., 2018; Fridkin & Kenney, 2014). Likewise, research on race
and inclusion finds that racial and ethnic minorities (at least in the United States) are more likely to trust
institutions and are more likely to political engage when their group is represented (Broockman, 2014;
Clark, 2019; Gay, 2001; Hayes & Hibbing, 2017; Rocha et al., 2010; Scherer & Curry, 2010; Tate, 2003).
Representative bureaucracy scholars have similarly emphasized how underrepresented groups respond
to descriptive representation. Past research on police forces specifically suggests that inclusion may be a
viable strategy to increase trust among previously marginalized groups. Riccucci et al. (2014), for exam-
ple, find that including more women in domestic violence units increases women's confidence in these
units. Schuck et al. (2021) find that when a department is arresting a lower percentage of sexual assault
offenders, increases in women's representation among units and the leadership increases perceptions of
legitimacy and feelings of trust.
Additionally, Riccucci et al. (2016) show that having more women in an agency's leadership can
increase female citizens’ willingness to cooperate with the agency. Research on racial representation and
the police likewise suggests that Black respondents are more likely to view police actions as legitimate
when carried out by Black officers and that higher levels of racial diversity lead to higher levels of trust
and greater perceptions of fairness among Black respondents (Riccucci et al., 2018). Similarly, focusing on
a Brazilian favela, an experimental study by Dantas Cabral et al. (2022) finds that greater representation
in local police departments increased trust in those departments. Recent research, however, suggests that
symbolic representation has limits, and it might not necessarily improve perceptions among marginalized
individuals. Based on their qualitative interviews, Headley et al. (2021) suggest that citizens' perceptions
of the police are largely shaped by their lived experiences and their previous interactions with police
officers, which implies that the symbolic benefits of descriptive representation are likely contingent on
how the police have treated citizens in the past.
STAUFFER ET Al.
932

Based on past research supporting the link between descriptive and symbolic representation, we
expect that members of historically marginalized groups (i.e., those who are women and those who are
racial and ethnic minorities) will evaluate police agencies more favorably if women and non-Whites make
up a higher proportion of the force (Hypothesis 1).
Response by the historically dominant group
While understanding how marginalized groups respond to representation is important, it is also important
to consider how historically dominant groups (i.e., men, with respect to gender, and Whites, with respect
to race) respond to increased diversity. The importance of understanding the dominant group's response is
underscored by recent work showing that White identity and consciousness are inherently ingroup oriented,
affect attitudes toward policies that protect the interests of the ingroup, and that a “White identity” is not
synonymous with views of “White supremacy” (Jardina, 2019). We should note that, while we discuss possi-
ble responses by male and White respondents together, their responses may differ as the mechanism(s) at
work may be distinct in each case. The dominant group may respond in one of three, general ways.
First, the dominant group, here men or Whites, may view increasing diversity as a threat to their
group's dominant status creating a backlash effect (Hypothesis 2a). A number of literatures converge
on the same logic as to why this might be the case, such as critical race theory, social dominance theory,
and the theory of representative bureaucracy. Critical race theory tells us that our laws, institutions, and
culture establish what we perceive as normal, such as race and subsequent racial disparities (e.g., Conyers
& Wright Fields, 2021; Delgado & Stefancic, 2017; Gamal, 2016; Riccucci, 2021; Treviño et al., 2008). As
it applies here, the police have been normalized to be considered by many Whites as a “neutral” institution
that aims to help and protect its citizens and whose modal officer is white (and male). Deviations from
that normalized picture then, might negatively impact evaluations of the police and generate perceptions
of reverse discrimination.
Similarly, social dominance theory suggests that individuals with a high social dominance orientation
have a preference for stable inequality among social groups and may not support institutional changes
that reduce existing discrimination (Pratto et al., 1994; Sidanius & Pratto, 1999). This tendency is often
associated with the objection to affirmative action and leads to backlash to increased gender and racial
diversity (Haley & Sidanius, 2006). Research notes that these identities are likely primed in hierarchical
environments (e.g., male-dominated police agencies); in such an environment, maintaining a high social
dominance orientation can be considered a way of reinforcing social hierarchy and preserving the status
quo (McLaughlin et al., 2021; Sidanius et al., 2006).
Some representative bureaucracy scholars note that active representation of minoritized populations
may lead members of dominant groups to perceive that the policy processes are not likely to be neutral
and that policy outcomes are more likely to be biased in favor of women and/or racial and ethnic minor-
ities, and against the interests of men and/or Whites (Lim, 2006). This point may be particularly potent
in the context of the Weberian notion of bureaucracy that holds that bureaucracies should be neutral and
depoliticized (Weber, 1958). In this perspective, bureaucratic efforts to represent non-White or women
may be seen as adding biases to a neutral process or even generating favoritism in the eyes of White or
male citizens as the status quo process historically favors these dominant groups (Lim, 2006; Yun, 2020).
However, as scholars highlight, these perceptions are born from a “naive view of structures,” which views
government structures as ways to solve social problems and bureaucracies as a neutral policy instrument
(Meier & Morton, 2015, p. 110).
Further, some studies discuss and show partial support of the “privilege maintenance hypothesis”
which suggests that “a bureaucratic encounter that favors disadvantaged groups may be viewed as unfair
by privileged groups” (Van Ryzin, 2021, p. 3). Specifically, research shows that a female teacher helping
a female student was judged to be unfair by Americans, and this negative effect is mainly driven by male
respondents (Van Ryzin, 2021). Similarly, research also finds that White school principals tend to be less
favorable toward a minority-supportive policy than non-White principals (Andersen, 2017) and the effect
POlICE DIVERSITY, POlICIES, AND OUTCOMES 933

of greater Black representation in police forces is largely negative among White citizens, especially when
civilian complaints are increasing (Riccucci et al., 2018). 1
Second, while the dominant group may respond with hostility to the inclusion of historically margin-
alized groups, they may also be ambivalent. In other words, dominant groups may not consider diversity
much when evaluating institutions. If this is the case, we would expect Whites (men) to not be affected
one way or the other by the inclusion of racial and ethnic minorities (women) in police forces (Hypothesis
2b). Racial ambivalence theory may offer a useful theoretical lens here. Racial ambivalence theory suggests
that Whites' attitudes toward Blacks are not always straightforward; rather they are best understood as
ambivalent (Hogg & Levine, 2010). Scholars show that this ambivalence is often based on conflicting
values: Whites' positive attitudes toward Blacks are based on their support for humanitarian-egalitarian
values, whereas the negative attitudes are the result of the endorsement of the Protestant work ethic
(Hogg & Levine, 2010; Katz et al., 1986).
Another possible reason for this ambivalence among the dominant group may be differences in the
degree to which a shared identity in policy spaces is salient to the individual, such as being a woman in a
predominately male space or a Black individual in a predominately White space. Further, differences in
the degree to which the domain is salient to the individual, such as policing and the criminal justice system
for Black Americans (e.g., Gibson & Nelson, 2018). One facet of White (male) privilege in U.S. society is
not having one's racial identity be a salient identity. 2 In other words, White respondents may simply not
register information on racial diversity or may not view increased diversity as threatening, as they may not
have strong or salient ingroup attachments with respect to their race (or gender).
This proposal aligns with the findings of Van Ryzin (2021) concerning the response of White citizens
to racial representation and Riccucci et al. (2016) concerning the response of male citizens to women
in leadership. In particular, Van Ryzin (2021) shows that White participants do not necessarily view the
encounters where a Black officer helps a Black citizen as unfair, whereas Black participants judge the
encounter in which a White officer helps a white citizen as unfair. The statistically insignificant results
among White participants imply that Whites might not consider representation (or diversity) when evalu-
ating agency fairness as much as marginalized groups do.
Finally, the dominant groups may actually respond favorably to increased diversity (Hypothesis 2c).
Indeed, increasing evidence in political science finds that men as well as women place greater confidence
in gender diverse institutions (Clayton et al., 2018; Schwindt-Bayer, 2010; Schwindt-Bayer & Mishler, 2005;
Stauffer, 2021). Likewise, Hayes and Hibbing (2017) find that in the context of race, both Black and
White respondents view institutions as more legitimate when they are racially diverse. One explanation for
these findings is what Schwindt-Bayer and Mishler (2005) refer to as a “socio-tropic model of representa-
tion.” In this framework, the inclusion of marginalized groups sends broader signals about openness and
responsiveness, sparking positive affect from all citizens.
Thus, the literature offers competing expectations for how dominant groups may respond to
increases in diversity on police forces. We adjudicate between these competing perspectives below using
two conjoint experiments.
OTHER DETERMINANTS OF PUBLIC SUPPORT
While descriptive representation may play a role in shaping attitudes toward the police, it might also be true
that local residents care more about policing outcomes and police-citizen interactions. For example, they may
not care about the composition of the force as long as crime rates remain low, or they may not care about the
composition of the force if they feel harassed by the police. Previous research has shown that coverage of
incidents of police use of force or the rates at which force is used (Boudreau et al., 2019; Mullinix et al., 2021),
quality of services provided (Myhill & Bradford, 2012; Weitzer & Tuch, 2005), and incidents and level of
police misconduct (Weitzer, 2002, 2015) are related to public opinion about the police. More specifically,
this research has found that as rates of force decrease, rates of misconduct decrease, and quality of services
improves that evaluations of and trust in the police improve. We expect to observe similar relationships here.
STAUFFER ET Al.
934

In addition to these considerations, two policies have been pointed to as possible avenues to further
rebuild trust. First, there are proposals to expand community policing efforts. Previous research suggests
that community-oriented approaches increase community trust (Gill et al., 2014; Peyton et al., 2019).
Second, there are proposals to decrease police funding and increase funding for other offices/organ-
izations by protesters, organizations, the news media, and academics (e.g., Harris et al., 2020; Shoub
et al., 2020). As such, information on the degree of community policing and how well (or poorly)
resourced a local department is may influence trust in the police and perceptions of accountability.
Most experimental studies examining the question of whether diversity influences public opinion
about the police have not examined how these various components combine to shape public attitudes.
When presented with both performance information and information on diversity, respondents may
respond differently. In the extreme case, a joint consideration may reveal no relationship between agency
diversity and opinion. In the less extreme case, a joint consideration may reveal more muted patterns than
previously found. Alternatively, joint consideration of these different aspects of an agency may have no
impact on evaluations and interactions.
TESTING THE ROLE OF DIVERSITY IN CITIZEN EVALUATIONS
To adjudicate between these considerations and to test whether racial and ethnic diversity and/or gender
diversity influences public perceptions of the police, we designed two conjoint experiments (Hainmueller
& Hopkins, 2015). Though experiments have become an increasingly popular method for testing the
relationship between descriptive representation in bureaucratic agencies and the attitudes citizens hold
toward those institutions, these studies often manipulate only one facet of the agency (i.e., the level of
either gender or racial diversity) at a time. While this affords the researcher a great deal of control, such
designs often do not allow researchers to consider how multiple factors may vary across agencies and how
this variation may influence public opinion. Accounting for the multiple factors that likely influence public
opinion toward police forces is important in this context because we are interested in understanding how
diversity influences attitudes to the police relative to other agency characteristics. It may be the case, for
example, that citizens prefer diverse forces in the absence of full information, but that this preference is
minimal when compared to other considerations (e.g., an agency's felony solve rate). Thus, understanding
citizen preferences relative to other factors is essential for understanding whether increasing diversity is a
viable strategy for repairing trust between citizens and the police.
Given the question at hand and associated concerns, a conjoint design is particularly well suited to
our needs. Conjoint experiments allow researchers to simultaneously randomize multiple attributes of
an experimental treatment. Here, the experimental treatment randomizes various attributes of a police
force, which allows us to evaluate public attitudes toward the police as a multi-dimensional concept and
recover the independent effects of different types of force diversity relative to other factors. Past research
has validated the external validity of conjoint designs, showing that they are often good predictors of
real-world behavior (Hainmueller & Hopkins, 2015). Moreover, recent research using eye-tracking soft-
ware verifies that participants are attentive to information presented in these designs (Jenke et al., 2021).
Our two experiments were identical in design, with the only difference being the dependent varia-
bles. Prior to completing the experimental tasks, we informed respondents that they would be asked to
evaluate police agencies based on several pieces of information. 3 Respondents were informed that the
information provided came from annual report cards issued by the National Association of Policing
and Law Enforcement (NAPLE), a fictitious organization created by the researchers. 4 Each report card
included nine attributes. Attributes included descriptive information (i.e., percentage non-White officers,
percentage women officers, and population served), performance information (i.e., felony solve rate,
civilian complaints of police misconduct, rate of deadly force, and rate of use of force), and organiza-
tion information (i.e., resources allocated to community policing and local budget allocated to the force).
Within each profile, attributes were randomly assigned to a particular level. For example, in our profiles,
POlICE DIVERSITY, POlICIES, AND OUTCOMES 935

“Felony solve rate” was randomly assigned to take on values of “Above Average,” “Average,” and “Below
Average.” 5 Table 1 presents the full list of attributes as well as their possible levels.
In his experimental work, Olsen (2017) examines the extent to which citizens respond to “episodic”
or “statistical” based information. The former refers to specific examples of individual experiences
with an agency while the latter refers to more abstracted information about overall performance. As
our report-card style treatments fall under the category of statistical based information, Olsen (2017)
provides insights into the generalizability and believability of our results. First, when given the choice, citi-
zens prefer to receive statistical information about agency performance compared to episodic informa-
tion, which is how our treatments provide information. Additionally, while not always the case, episodic
information often evokes stronger emotional responses from respondents, and respondents are more
likely to recall this information long-term relative to statistical information. Because our study is related
to policing, we would expect this to be the type of context where episodic information elicits a stronger
effect on individuals. Seeing videos or images of the police engaging in overly aggressive uses of force, for
example, is likely to evoke stronger responses than statistical information. To the extent that these differ-
ences influence our findings – particularly with regard to the performance attributes – we would expect
our results to be conservative. This design, however, should still allow us to evaluate the relative weight
that citizens give to various factors of agency performance and how they (potentially) shape opinion.
To test the link between force diversity and citizen evaluations, we are particularly interested in two
attributes: “% Female officers” and “% Non-White officers.” 6 The randomized values for percentage
female and non-White officers were selected to capture a wide range of force diversity and inclusion.
While we are interested in evaluating how a high degree of gender or racial inclusion (or exclusion) might
influence attitudes toward police forces, we also wanted to maintain the relative realism of the experimen-
tal treatments when possible. Because police forces remain primarily White, male institutions, we wanted
to ensure that the proportions of women and non-White officers included (particularly at the higher end)
in our profiles were believable to respondents. To that end, in fall 2020, we asked 200 Mechanical Turk
workers to estimate the percentage of women and non-White officers in both their local police force and
in a “typical” police force. Overall, respondents believed police forces were more racially diverse than
gender diverse, but in both cases estimates were wide-ranging. 7 For both our percent female and percent
non-White officers attributes, we include values falling below the 75th percentile of respondent estimates.
Keeping with the patterns observed from our MTurk workers, we select slightly lower values for the
percent women attribute. We should also note that the values chosen are within the range of observed
values for both the percentage of women on a force and the percentage of non-white officers based on
responses to the 2016 Law Enforcement and Agency Management Statistics survey. For more informa-
tion on this and for the associated density plots, see the Appendix.
If diversity fosters more positive evaluations, such as greater feelings of trust, then we should see
respondents giving more positive evaluations to agencies with high levels of diversity and lower ratings to
STAUFFER ET Al.
936
Attributes Level
Felony solve rate Above average, Average, Below average
Civilian complaints of police misconduct Above average, Average, Below average
Rate of deadly force Above average, Average, Below average
Rate of use of force Above average, Average, Below average
Resources allocated to community policing Above average, Average, Below average
Local budget allocated to force Above average, Average, Below average
Population served Urban, Rural, Suburban
% Non-White officers 4%, 16%, 29%, 41%, 54%
% Women officers 6%, 17%, 28%, 37%, 46%
Note: Indicators fully randomized between participants in experiment. Attributes related to agency demographic diversity in bold.
TABLE 1 Conjoint experiment summary

agencies with low levels of diversity. Moreover, if these effects are specific to marginalized groups, then
we should see women responding favorably to gender diversity and non-White respondents respond-
ing favorably to information about racial and ethnic diversity. Examining how (and if) White (male)
respondents adjust their opinions in response to this information allows us to evaluate our second set of
hypotheses. Because we include other agency attributes in our profiles, we also evaluate the relative weight
respondents give to diversity compared to other factors.
In total, each respondent evaluated five, randomly-generated profiles. The order of attributes was
randomized between participants to prevent ordering effects (Hainmueller et al., 2014). Figure 1 presents
an example of the profiles as seen by participants.
Using this design, we conducted two experiments in September 2020 with identical designs. 8 The
only difference between the experiments is the dependent variables, which test how diversity may influ-
ence evaluations of legitimacy and accountability. In Study 1, we measure what we call the “perceived
legitimacy” of police agencies. This concept captures the degree to which citizens trust the agency and
believe they will be treated fairly. Study 2 examines what we call “perceived accountability,” a concept
that captures citizen perceptions of agencies' interactions with the community and other local author-
ities. 9 After reading each profile, respondents were asked a few questions about each agency, which are
then used to create a scale capturing our latent concepts. In Study 1, respondents were asked the extent
to which they agreed with the following statements (with five response options ranging from “strongly
disagree” to “strongly agree”):
1. Citizens can trust this police department to do what is right.
2. This police department treats citizens fairly.
3. This police department is an asset to its community.
These statements are similar to other studies also interested in symbolic representation and the police
(e.g., Riccucci et al., 2018).
Using these three items, we create a summated scale, with higher values corresponding to higher levels
of perceived legitimacy. For ease of interpretation, we rescale the resulting measure to range from 0–1.
Cronbach's α for this scale is 0.907, indicating a high degree of internal consistency.
In Study 2, respondents were similarly presented with a series of statements and asked the extent to
which they agreed based on each profile (again with five response options ranging from “strongly disa-
gree” to “strongly agree”). Statements in study two were as follows:
1. If I needed help, I would feel comfortable contacting this agency.
2. This agency is held accountable by its local government.
3. This agency should be given more authority in its community.
While similar to the statements included in Study 1, these questions allow us to gain a better sense of
the extent to which individuals would be comfortable interacting with agencies and appropriate role of the
agency in the community. The first reflects a more holistic assessment of public perceptions of account-
ability and ties it to behavioral consequences. The second directly captures perceptions of accountabil-
ity. The third focuses on the authority arrangements aspect of accountability. As in the first study, we
create a summated scale using responses and rescale the result to a range from 0–1. This is our measure
of “perceived accountability” (Cronbach's α = 0.837). Given the salience of current debates on police
accountability and resource allocation, we view the inclusion of these measures as particularly relevant.
For each study, we recruited 750 participants through the survey firm Lucid. The unit of analysis is
profile evaluations, resulting in an N of roughly 3,750 for each study. Lucid is a company that aggregates
survey respondents from multiple sources to provide convenience samples that roughly match population
benchmarks from the U.S. census. Coppock and McClellan (2019) find that samples procured through
Lucid largely replicate previous experimental findings from probability samples and other convenience
sampling platforms. 10 Because one of our hypotheses concerns comparisons between those of different
POlICE DIVERSITY, POlICIES, AND OUTCOMES 937

races, both studies include an over-sample of Black respondents, which is the largest non-White group in
the United States. The resulting sample is roughly split evenly across White and Black respondents.
RESULTS
To test whether officer diversity influences citizen perceptions of the police, we calculate marginal means
and compare these quantities to the overall mean evaluation for all profiles included in our samples. 11
Marginal means represent the mean evaluation when a particular attribute level is present, independent
STAUFFER ET Al.
938
FIGURE 1 Example report card from study 1

of all other attributes and their levels (Leeper et al., 2020). Because each attribute level is completely
randomized, all attributes are independent. Although our hypotheses predict a conditional relationship
between force diversity and citizen evaluations, we begin by examining relationships for the full sample.
Figure 2 presents the results for perceived legitimacy in Study 1 and Figure 3 presents the results for
perceived accountability in Study 2. 12 In both cases, these figures present the average level of the outcome
variable at various levels of each attribute. The solid line represents the overall mean of perceived legiti-
macy and accountability, respectively, across all profile evaluations.
Turning first to perceived legitimacy. We see that on balance, gender diversity does not appear to
significantly influence public opinion. In no case is the marginal mean for any level of the “% Women”
attribute significantly different from the overall mean. However, racial diversity does appear to positively
influence perceived legitimacy, but only at the highest levels of diversity tested in this study (p = 0.09).
Examining the effects of agency performance, we see that these factors clearly play a strong role in shap-
ing public assessments of legitimacy. Agencies with a higher than the average number of complaints,
above-average use of force (and deadly force), and below-average felony solve rates are punished heavily,
with legitimacy decreases. Meanwhile, a low number of complaints, a below-average use of force (and
deadly force) and an above-average solve rate all lead to increased perceptions of legitimacy.
Turning to perceived accountability, we again see minimal evidence that gender diversity influences
public opinion on average. Racial diversity, in contrast, does appear to play a role. Agencies with high
levels of diversity are perceived as more accountable relative to the overall mean, while agencies with
minimal diversity are perceived as less accountable, relative to the overall mean. However, as before, we
note that these effects only manifest at the extremes, when diversity is at its highest or lowest. As before,
we also see agency performance playing a substantively significant role in shaping public evaluations.
While these results offer some insights into the factors that citizens consider when evaluating police
forces, we posited that to the extent diversity matters in public opinion formation, this is likely condi-
tional. To assess expectations, in the sections that follow, we conduct subgroup analyses, examining how
women respond to gender diversity compared to men and how Black respondents respond to racial and
ethnic diversity compared to White respondents.
Women's (and men's) response to gender diversity
We begin with an examination of gender diversity. We calculate overall means (and deviations from
these means) separately for men and women, rather than comparing whether men and women deviate
from the overall mean. Figure 4 presents these results, with Figure 4a showing the results from Study 1
and Figure 4b showing the results from Study 2. For ease of visualization, we present only results for
the “% Women” attribute. Full results are available in the Appendix. These figures present the deviation
between women's (men's) overall mean evaluations compared to women's (men's) mean evaluations when
a particular level of gender diversity is displayed (see Leeper et al., 2020). 13 Turning first to the results
from Study 1 (Figure 4a), we find some evidence that gender diversity influences perceived legitimacy
among female respondents, but only in limited circumstances. While gender diverse forces appear to
be viewed as more legitimate by women in our sample, we note this effect is only present when women
make up 46% of an agency, indicating the bar is high for gender diversity to shift attitudes. In Study 2
(Figure 4b), we examine respondent perceptions of agency accountability. Like Study 1, we find evidence
that increases in women's presence can lead to higher levels of perceived accountability among women
respondents, but this again is only seen at the highest levels of diversity.
Across both studies and outcome measures, we find no evidence that levels of gender diversity influ-
ence how men evaluate police agencies, in line with Hypothesis 2b. To more thoroughly evaluate differ-
ences between men and women in our sample, we also tested whether the difference in men's deviation
from their group mean was significantly different from women's deviation from their group mean in
the case of agencies with 46% women. In Study 1, this quantity was significantly different (p = 0.071).
However, while we observe a significant difference for women, but not men, in Study 2, these differ-
POlICE DIVERSITY, POlICIES, AND OUTCOMES 939

ences themselves are not significantly different (p = 0.2). Given this mixed evidence, inferring differences
between men and women's response to high level of gender diversity in Study 2 should be done with
caution. At the same time, however, these findings do provide consistent evidence that there is not a
“backlash effect” among men in contexts where women make up a high proportion of the force.
To some extent, this is an encouraging finding. While men do not reward (punish) agencies with high
(low) levels of diversity, the fact that we find no evidence of backlash among men as women increas-
ingly enter the force is notable. While our primary focus in this paper is the extent to which inclusion
influences citizen opinions, increasingly evidence suggests that women officers police differently and that
STAUFFER ET Al.
940
FIGURE 2 Evaluations of hypothetical agencies: Study 1 perceived legitimacy. The solid vertical line represents the overall
average score for all profiles. Points represent mean values when a particular attribute level is displayed. Bars represent 95%
confidence intervals.
6%
17%
28%
37%
46%
(Women Rep.)
4%
16%
29%
41%
54%
(Minority Rep.)
Resources: Ab
ove Average
Resources: Average
Resources: Below Average
(Resources)
Rural
Suburban
Urban
(Population Served)
Budget: Above Average
Budget: Average
Budget: Below Average
(Budget)
Force: Above Average
Force: Average
Force: Below Average
(Use of Force)
Lethal: Above Average
Lethal: Average
Lethal: Below Average
(Lethal Encounters)
Solve: Above Average
Solve: Average
Solve: Below Average
(Solve Rate)
Civ: Above Average
Civ: Average
Civ: Below Average
(Civilian Complaints)
0.50 0.55 0.60 0.65 0.70
Marginal Mean

these differences lead to better outcomes (i.e., Shoub et al., 2021). Our findings here suggest agencies
can increase women's representation and potentially reap these benefits without incurring a cost from a
subsection of the public.
The results presented here indicate that gender diversity can influence how women perceive police
forces in some contexts and that this diversity does not lead to a backlash among men. However, it is
important to contextualize these results as they relate to other agency characteristics and performance
metrics. For both men and women, performance metrics are by far the largest drivers of public opinion.
POlICE DIVERSITY, POlICIES, AND OUTCOMES 941
FIGURE 3 Evaluations of hypothetical agencies: Study 2 perceived accountability. The solid vertical line represents the
overall average score for all profiles. Points represent mean values when a particular attribute level is displayed. Bars represent
95% confidence intervals.
6%
17%
28%
37%
46%
(Women Rep.)
4%
16%
29%
41%
54%
(Minority Rep.)
Resources: Ab
ove Average
Resources: Average
Resources: Below Average
(Resources)
Rural
Suburban
Urban
(Population Served)
Budget: Above Average
Budget: Average
Budget: Below Average
(Budget)
Force: Above Average
Force: Average
Force: Below Average
(Use of Force)
Lethal: Above Average
Lethal: Average
Lethal: Below Average
(Lethal Encounters)
Solve: Above Average
Solve: Average
Solve: Below Average
(Solve Rate)
Civ: Above Average
Civ: Average
Civ: Below Average
(Civilian Complaints)
0.55 0.60 0.65
Marginal Mean

Thus, while high levels of gender diversity may influence women's evaluations of police agencies in some
contexts, reforms aimed at agency performance are likely to pay the highest public opinion dividends.
Black (and white) responses to racial diversity
We next turn our attention to examining how racial diversity shapes attitudes toward the police. Again
because we expect that effects may be conditional on respondent race, we calculate marginal means
(and deviations from these means) for Black and White respondents separately. Results are presented
in Figure 5, with results from Study 1 presented in Figure 5a and results from Study 2 are presented in
Figure 5b. As before, for ease of visualization, we present only the results for the “% Non-white officers”
attribute. Full results are available in the Appendix. Figure 5 shows the deviation between Black (White)
respondents overall mean evaluations compared to their evaluations when a particular level of racial and
ethnic diversity was displayed.
In Study 1, we observe that Black respondents report higher levels of perceived legitimacy when a
larger proportion of a police force is non-White. However, just as was the case with gender, we note that
this effect is only observed for the very highest levels of diversity. Moreover, we do not observe Black
respondents “punishing” agencies in terms of perceived legitimacy relative to the overall mean for lacking
diversity. However, we do see agencies being punished in terms of perceived accountability in Study 2.
Like in Study 1, we observe that Black respondents view agencies more favorably at the highest levels of
diversity. We also note that in Study 2 the lowest levels of diversity are also punished by Black respond-
ents. When agencies comprise just 4% non-white officers, Black respondents expressed lower levels of
perceived accountability.
Just as we found no evidence of backlash among men presented with gender diverse agencies, we
also find no evidence to suggest that White respondents respond negatively (positively) to high (low)
levels of racial diversity on the force. To the contrary, we see White respondents rewarding high levels of
STAUFFER ET Al.
942
FIGURE 4 Evaluations of hypothetical agencies given different levels of force gender diversity by respondent gender and
study. Thin bars indicate 95% confidence intervals; thicker bars indicate 90% confidence intervals.
Study 1: Perceived Legitimacy Study 2: Perceived Accountability
−0.050
−0.025
0.000
0.025
0.050
0.075
−0.050
−0.025
0.000
0.025
0.050
0.075
6%
17%
28%
37%
46%
Difference from Group Mean
Gender Men Women
(a) (b)

diversity. White respondents view agencies with high levels of diversity as more accountable (p < 0.05).
These respondents also view agencies as more legitimate, though the difference between White's mean
evaluation in the highest diversity condition and their grand mean does not reach conventional levels of
significance (p = 0.105). We likewise see White respondents viewing agencies with low levels of diversity
as less accountable (p = 0.089). This set of findings is consistent with Hypothesis 2c. Again, to more
thoroughly evaluate differences in how White and Black respondents respond to racial diversity, we test
whether the differences between mean evaluations at high (low) levels of diversity and group grand means
differ across groups. First looking at perceived accountability. In neither case do we find evidence of a
significant difference-in-difference (p = 0.198 and p = 0.296, respectively). We similarly find no evidence
that the difference between Black respondents evaluations of legitimacy and their group's grand mean
is significantly different than the same quantity for White respondents when diversity is high (p = 0.33).
Findings from this study suggest that at high levels of diversity civilians view the police as more
legitimate and perceive higher levels of accountability and these results hold for both Black and White
respondents. Moreover, perceived accountability appears to suffer in contexts where racial and ethnic
diversity is lacking on the force. However, we once again note that our findings tend to operate at the
extremes. Only the most diverse forces are rewarded, and only the least diverse are punished. Gradients
in between seem to make little difference. Moreover, we again note that, compared to performance char-
acteristics, diversity does exerts a relatively small effect.
DISCUSSION AND CONCLUSION
Recent events highlight the erosion in trust in the police that has occurred. While there is generally wide-
spread agreement that reforms are needed, consensus regarding specific reforms has remained elusive.
One commonly cited reform is for police forces to diversify. In this paper, we address whether diversifi-
cation represents a viable strategy for agencies looking to regain trust with the public they serve. Using
two conjoint experiments, we examine whether (and how) both gender and racial diversity among officers
POlICE DIVERSITY, POlICIES, AND OUTCOMES 943
FIGURE 5 Evaluations of hypothetical agencies given different levels of force racial diversity by respondent race and
study. Thin bars indicate 95% confidence intervals; thicker bars indicate 90% confidence intervals.
Study 1: Perceived Legitimacy Study 2: Perceived Accountability
−0.075
−0.050
−0.025
0.000
0.025
0.050
0.075
−0.075
−0.050
−0.025
0.000
0.025
0.050
0.075
4%
16%
29%
41%
54%
Difference from Group Mean
Race BlackWhite
(a) (b)

shape public support for the police. This design allows us to examine whether diversity meaningfully
shapes opinion when other forms of information are available and how the effects of diversity compare
to other agency characteristics.
Our analyses indicate that diversity can influence support for the police independent of other factors
but that these effects operate at the extremes. In the case of women's representation, we found that
women perceived agencies as more legitimate and accountable but only at the highest levels of diversity
tested. We likewise observe that racial diversity (or a lack thereof) can engender (detract) support for the
police among Black Americans. Across both studies, Black respondents only punished (rewarded) agen-
cies with the very lowest (highest) levels of diversity; outside of these contexts, inclusion seemed to do
little to move the needle of public opinion. These findings indicate that increased diversity can influence
perceptions of the police among women and Black Americans – but only in the most extreme scenar-
ios tested. Thus, while diversifying may help forces at the lowest end of the spectrum, more generally
anything less than substantial gains in diversity may do little to increase trust.
Further, we examined how diversity influences support among historically dominant groups (i.e., men
and White Americans). We found no evidence that these groups respond negatively to the inclusion of
marginalized groups. Indeed, in the case of White respondents, we found some evidence that racial diver-
sity can lead to more favorable attitudes toward police agencies. However, we did not find a similar effect
for male respondents and gender diversity. Divergence of these findings may indicate that distinct mech-
anisms are at work with respect to racial and gender diversity and attitudes of the dominant group. For
example, it may be that increased racial, but not gender, diversity may be perceived to alter the degree to
which racism is a problem in a given department, as non-White, and specifically Black, officers may be
perceived as not (or less) racist. While we cannot untangle these distinct mechanisms here, we view this
as a fruitful avenue for future research.
Moving forward, additional research should examine how the inclusion of other historically margin-
alized groups influences trust among community members. In this article we focused on how women
(and men) and Black (and White) respondents reacted to police diversity. Yet, it is also important to
consider how other racial and ethnic minority groups – such as Latinx or Asian communities – respond to
increased representation. Though we would expect similar relationships to emerge among these groups,
further investigation would be needed to confirm this expectation. It is also crucial to consider how racial
and ethnic minority groups respond to increased inclusion of their group and to the increased inclusion
of other groups, which we are unable to do in this study. This is especially important as some research
reveals conflict between non-White communities. For example, a Cuban American may respond differ-
ently to diversity increasing due to the hiring of more Black officers and vice versa.
While it is not without limitations, this study contributes to the literature on representative bureau-
cracy and symbolic representation and to the literature on policing in three important ways. First, we
examined multiple forms of diversity in parallel to each other, which builds on insights from prior studies
by putting into conversation multiple literatures from both political science and public administration.
Given that prior research has mainly focused on one dimension of diversity at a time, our study offers
a more comprehensive approach to tackling the role of agency diversity in shaping citizen perceptions.
However, our study does not adopt an intersectional approach, as we provide information on isolated
identities mimicking how such information is typically presented and discussed. Future studies should
push this avenue of research even further by doing so.
Second, we centered our theoretical discussion on both the historically marginalized group, as is
often done in studies of symbolic representation, and the historically dominant group, which is often
the accidental focus in general studies of policing, as Whites make up the majority of the population in
the United States. In doing so, we highlight that both groups respond differently to changing levels of
diversity, finding little evidence of backlash to rising levels of diversity. Understanding the heterogeneous
nature of responses is important, as the dominant group's view is important because they often strongly
influence the public policy process. If the dominant group views diversity favorably, policymakers may
be more likely to push institutional changes that reduce racism and sexism, while if increased diversity
produces backlash, leading to severe racism and sexism, then such changes are less likely. Either way,
STAUFFER ET Al.
944

theorizing and understanding how these individuals respond to diversity is crucial; this article advances
our understanding of the dominant groups views as well as the marginalized group's views of diversity.
Third, we simultaneously examined agency diversity and varied other salient agency characteristics,
which connects the literature on representative bureaucracy to research on performance information.
When evaluating agencies, citizens consider multiple pieces of information about an agency simultane-
ously – not only diversity but also agency performance (Porumbescu et al., 2021), thus it is important
to account for these factors when studying public opinion toward police forces. While many scholars
highlight the critical role of performance information and contextual factors in shaping citizen percep-
tions of accountability and satisfaction (DeHoog et al., 1990; Olsen, 2017), previous experiments on
symbolic representation have been conducted in isolation from these other considerations. Theoretically,
this allows for a more robust conversation about the impact of descriptive representation on symbolic
representation.
At the same time, it is important to note that the present study relies on providing respondents with
agency-level statistical performance information. As such, we are unable to evaluate how specific encoun-
ters with the police influence respondent opinion, which likely evoke more intense responses. Future
research should continue to probe how more visceral experiences shape public support for the police and
how this differs from exposure to statistical performance information.
Finally, it is important to acknowledge the potentially unique time frame of our study. Coming on
the heels of the protests of summer 2020, it is possible that the unique climate surrounding these events
influenced our results in some way, though as we note in the Appendix it appears that public attention
had receded by the time our survey was fielded. Regardless, our findings point to the limits of descrip-
tive representation in restoring public trust in political institutions. If our results are a spillover from
the events of summer 2020, this suggests that in contexts where calls for reform are strongest, the
inclusion of historically marginalized groups is insufficient. If, as we suspect, our results are more akin
to periods where police reform discourse is more muted, this indicates that including underrepresented
groups in police forces is likely to do little more than maintain the status quo. In this sense, future work
on representation in both public administration and political science should continue to examine the
limits of when descriptive representation translates into symbolic representation. Moreover, these find-
ings should serve as a reminder to practitioners that simply adding marginalized groups to the police is
not a cure-all. Such changes must be accompanied by systemic changes to agency structures and practices.
While our findings indicate that diversity can influence public support for and trust in the police in
some instances, an important caveat is that performance metrics, such as the felony solve rate, rate of
civilian complaints, and rate of use of (deadly) force, played a far larger role in shaping opinion than
diversity. Thus, our findings indicate that while diversity may help increase public support for the police
in some instances, it is by no means a comprehensive solution. However, increasing diversity may affect
performance outcomes, thus indirectly informing opinion. Instead, agencies seeking to regain public trust
should look to implement reforms that will alter agency outcomes and performance in conjunction with
diversification to regain the trust that has been lost with the public, and they should question to what
extent diversification may help outcomes/performance. Further, this implies that for some policies (e.g.,
community policing) examining performance metrics is equally as important as capturing the public's
response to them, as programs may indirectly affect evaluations of the police by producing better policing
outcomes.
ACKNOWLEDGMENTS
We would like to thank participants at the 2020 Gender and Political Psychology conference, with a
special thanks to David Peterson, Jennifer Merolla, Heather Ondercin, Tessa Ditonto and participants on
our 2021 panel at the annual meeting of the Southern Political Science Association, with a special thanks
to our discussant Tiffany Barnes, for helpful comments on earlier drafts of this paper. Additionally, we
would like to thank Leah Christiani for feedback on an earlier draft of this paper.
POlICE DIVERSITY, POlICIES, AND OUTCOMES 945

ORCID
Katelyn E. Stauffer https://orcid.org/0000-0001-7285-0440
Miyeon Song https://orcid.org/0000-0002-2875-9786
Kelsey Shoub https://orcid.org/0000-0003-3370-3311
ENDNOTES
1 The authors note that this negative effect is relatively small compared to the significant gains in Black respondents perceptions
of the police.
2 For broader discussions, see Omi and Winant (2014) or DiAngelo (2018).
3 While many conjoint designs ask respondents to choose between two profiles, this is not a requirement for this design.
4 After completing the study respondents received a debriefing message informing them that NAPLE is a fictitious organization.
5 Due to the attributes included in our design, the terms above or below average and average have slightly different meanings depending
on the attribute. For example, an above average number of civilian complaints indicates poor functioning, whereas an above average
felony solve rate indicates a well-functioning department. One concern is that this might increase the cognitive load of respondents or
otherwise confuse them. In the results, performance indicators perform as expected. This helps assure us that, generally, respondents
read our treatments as intended. To the extent this did not occur, this should lead to a more conservative set of results.
6 We recognize a slight disconnect between this analytic approach and the design of the experiment. While we examine the opinion
of Black respondents, our use of “non-White” as a profile attribute might not cue higher levels of Black representation on the
force (as non-White can be a descriptor applied to Latinx or Asian individuals). Thus, a stronger design may have been to explic-
itly state the percentage of Black officers. However, we believe our design can still provide important insights. While respond-
ents may not think all non-White officers are Black, it is unlikely that they think none are, which means that seeing increases in
non-White representation should indicate some increase in Black representation. Consequently, this means that the results we
present here are likely conservative.
7 Leveraging respondent perceptions of inclusion is a similar approach to Stauffer (2021)’s study of women's representation in
legislative politics.
8 It is important to acknowledge the social-political context in which this study was fielded. The summer of 2020 saw widespread
protests of police violence and systematic racism within the criminal justice system. Some readers might question the extent to which
this affected responses to our treatments. On this point, it is worth noting that the majority of protests ended more than a month
before our survey was fielded. Moreover, evidence from Google trends suggest that attention to the issue of police accountability
spiked dramatically in the summer of 2020 but had receded to close to pre-summer levels by the time our survey went into the field
(see Appendix G). As such, our findings are likely to generalize to periods of relative stability in police reform discourse and are
unlikely an artifact of the political climate of civil unrest from the summer 2020. To understand how our findings would transport
to these contexts – and to more thoroughly evaluate the role that socio-political context plays – more research is needed.
9 One limitation of this approach is that because the two studies were run separately we are unable to fully assess discriminant validity
with regard to our measures. Results from the analyses are fairly consistent across studies – with a few small differences. For future
researchers interested in measurement, additional data and analysis would be required to evaluate the relationship between concepts.
10 Just as there has been increased concern about declining data quality from other online sampling options — most notably
Mechanical Turk — so too have scholars examined ways to ensure data quality from Lucid samples. Here, we implement recom-
mendations from Aronow et al. (2020) to screen out inattentive respondents at the beginning of the study.
11 An alternative approach would be to calculate the average marginal component effects (AMCEs). ACMEs show how the appear-
ance of a particular attribute level changes the dependent variable relative to a reference category. However, as Leeper, Hobolt,
and Tilley (2020, p. 1) note, comparing ACMEs across subgroups to describe group preferences can “be substantially misleading
about the degree of agreement or disagreement between subgroups due the simple, but often forgotten, property that inter-
actions are sensitive to the reference category used in regression analysis.” While our initial examination of the data explores
overall patterns in the data, testing our theoretical expectations ultimately requires a series of subgroup analyses, which appear
later in the text. Following the advice of Leeper et al. (2020), we use marginal means for these analyses; for consistency, we also
use marginal means in our initial examination of the data.
12 Because our sample includes an oversample of Black respondents, we construct survey weights to bring our sample in line with
the racial composition of the U.S. population.
13 For the marginal means by respondent sex and race, please see section E of the Appendix.
STAUFFER ET Al.
946

REFERENCES
Andersen, Simon Calmar. 2017. “From Passive to Active Representation: Experimental Evidence on the Role of Normative Values
in Shaping White and Minority Bureaucrats Policy Attitudes.” Journal of Public Administration Research and Theory 27(3): 400–14.
Andrews, Rhys, and Karen Johnston Miller. 2013. “Representative Bureaucracy, Gender, and Policing: The Case of Domestic
Violence Arrests in England.” Public Administration 91(4): 998–1014.
Aronow, Peter Michael, Joshua Kalla, Lilla Orr, and John Ternovski. 2020. “Evidence of Rising Rates of Inattentiveness on Lucid
in 2020.” SocArXiv.
Atkeson, Lonna Rae. 2003. “Not all Cues Are Created Equal: The Conditional Impact of Female Candidates on Political Engage-
ment.” Journal of Politics 65(4): 1040–61.
Atkeson, Lonna Rae, and Nancy Carrillo. 2007. “More Is Better: The Influence of Collective Female Descriptive Representation on
External Efficacy.” Politics & Gender 3(1): 79–101.
Badas, Alex, and Katelyn E. Stauffer. 2018. “Someone like me: Descriptive Representation and Support for Supreme Court Nomi-
nees.” Political Research Quarterly 71: 127–42.
Badas, Alex, and Katelyn E. Stauffer. 2019. “Michelle Obama as a Political Symbol: Race, Gender, and Public Opinion toward the
First Lady.” Politics & Gender 15: 431–59.
Barnes, Tiffany D., and Stephanie M. Burchard. 2012. “Engendering Politics the Impact of Descriptive Representation on Women's
Political Engagement in Sub-Saharan Africa.” Comparative Political Studies 46: 767–90.
Barnes, Tiffany D., and Michelle M. Taylor-Robinson. 2018. “Women Cabinet Ministers in Highly Visible Posts and Empowerment
of Women: Are the Two Related?” In Measuring Women's Political Empowerment across the Globe 229–55. Cham: Springer.
Baumgartner, Frank R., Kate Bell, Luke Beyer, Tara Boldrin, Libby Doyle, Lindsey Govan, Jack Halpert, et al. 2021. “Intersectional
Encounters, Representative Bureaucracy, and the Routine Traffic Stop.” Policy Studies Journal 49(3): 860–86.
Baumgartner, Frank R., Derek A. Epp, and Kelsey Shoub. 2018. Suspect Citizens: What 20 Million Traffic Stops Tell us about Policing and
Race. Cambridge: Cambridge University Press.
Boudreau, Cheryl, Scott A. MacKenzie, and Daniel J. Simmons. 2019. “Police Violence and Public Perceptions: An Experimental
Study of how Information and Endorsements Affect Support for Law Enforcement.” The Journal of Politics 81(3): 1101–10.
Bradbury, Mark J., and Edward Kellough. 2011. “Representative Bureaucracy: Assessing the Evidence on Active Representation.”
The American Review of Public Administration 41(2): 157–67.
Broockman, David E. 2014. “Distorted Communication, Unequal Representation: Constituents Communicate Less to Representa-
tives Not of their Race.” American Journal of Political Science 58(2): 307–21.
Burns, Nancy, Kay Lehman Schlozman, and Sidney Verba. 2001. The Private Roots of Public Action. Cambridge, MA: Harvard Univer-
sity Press.
Celis, Karen, and Amy G. Mazur. 2012. “Hanna Pitkin's ‘Concept of Representation’ Revisited Introduction.” Politics & Gender 8(4):
508–12.
Clark, Christopher J. 2019. Gaining Voice: The Causes and Consequences of Black Representation in the American States. New York: Oxford
University Press.
Clayton, Amanda, Diana Z. O'Brien, and Jennifer M. Piscopo. 2018. “All Male Panels? Represen- Tation and Democratic Legiti-
macy.” American Journal of Political Science 63: 113–29.
Conyers, Addrain, and Christina Wright Fields. 2021. “A Critical Race Theory Analysis of Institutional Racial Paralysis in Organiza-
tional Culture.” Public Integrity 23(5): 484–95.
Coppock, Alexander, and Oliver A. McClellan. 2019. “Validating the Demographic, Political, Psy- Chological, and Experimental
Results Obtained from a New Source of Online Survey Respondents.” Research & Politics 6(1): 2053168018822174.
Dantas Cabral, André, Alketa Peci, and Gregg G. Van Ryzin. 2022. “Representation, Reputation and Expectations towards Bureau-
cracy: Experimental Findings from a Favela in Brazil.” Public Management Review 24(9): 1452–77.
DeHoog, Ruth Hoogland, David Lowery, and William E. Lyons. 1990. “Citizen Satisfaction with Local Governance: A Test of
Individual, Jurisdictional, and City-Specific Explanations.” The Journal of Politics 52(3): 807–37.
Delgado, Richard, and Jean Stefancic. 2017. Critical Race Theory. New York: New York University Press.
DiAngelo, Robin. 2018. White Fragility: Why it's So Hard for White People to Talk about Racism. Boston, MA: Beacon Press.
Dovi, Suzanne. 2002. “Preferable Descriptive Representatives: Will Just any Woman, Black, or Latino Do?” American Political Science
Review 96(4): 729–43.
Epp, Charles R., Steven Maynard-Moody, and Donald P. Haider-Markel. 2014. Pulled over: How Police Stops Define Race and Citizenship.
Chicago, IL: University of Chicago Press.
Fay, Daniel L., Alisa Hicklin Fryar, Kenneth J. Meier, and Vicky Wilkins. 2021. “Intersectionality and Equity: Dynamic Bureaucratic
Representation in Higher Education.” Public Administration 99(2): 335–52.
Fridkin, Kim L., and Patrick J. Kenney. 2014. “How the Gender of US Senators Influences Peoples Understanding and Engagement
in Politics.” The Journal of Politics 76(4): 1017–31.
Gamal, Fanna. 2016. “The Racial Politics of Protection: A Critical Race Examination of Police Militarization.” California Law Review
104: 979.
Gay, Claudine. 2001. “The Effect of Black Congressional Representation on Political Participation.” The American Political Science
Review 95: 589–602.
POlICE DIVERSITY, POlICIES, AND OUTCOMES 947

Gay, Claudine. 2002. “Spirals of Trust? The Effect of Descriptive Representation on the Relationship between Citizens and their
Government.” American Journal of Political Science 46: 717–32.
Gaynor, Tia Sherèe, and Brandi Blessett. 2022. “Predatory Policing, Intersectional Subjec- Tion, and the Experiences of LGBTQ
People of Color in New Orleans.” Urban Affairs Review 58(5): 1305–39.
Gibson, James L., and Michael J. Nelson. 2018. Blacks and Blue: How and why African Americans Judge the American Legal System. New
York: Oxford University Press.
Gill, Charlotte, David Weisburd, Cody W. Telep, Zoe Vitter, and Trevor Bennett. 2014. “Community-Oriented Policing to Reduce
Crime, Disorder and Fear and Increase Satisfaction and Legitimacy among Citizens: A Systematic Review.” Journal of Experi-
mental Criminology 10(4): 399–428.
Gill, Rebecca D. 2020. “Black and Blue: How African Americans Judge the US Legal System. By James L. Gibson and Michael J.
Nelson. New York: Oxford University Press, 2018. 220p. 99.95cloth, 29.95 Paper.” Perspectives on Politics 18(1): 260–2.
Hainmueller, Jens, and Daniel J. Hopkins. 2015. “The Hidden American Immigration Consensus: A Conjoint Analysis of Attitudes
toward Immigrants.” American Journal of Political Science 59(3): 529–48.
Hainmueller, Jens, Daniel J. Hopkins, and Teppei Yamamoto. 2014. “Causal Inference in Conjoint Analysis: Understanding Multidi-
mensional Choices Via Stated Preference Experiments.” Political Analysis 22(1): 1–30.
Haley, Hillary, and Jim Sidanius. 2006. “The Positive and Negative Framing of Affirmative Action: A Group Dominance Perspec-
tive.” Personality and Social Psychology Bulletin 32(5): 656–68.
Harris, Allison P., Elliott Ash, and Jeffrey Fagan. 2020. “Fiscal Pressures and Discriminatory Policing: Evidence from Traffic Stops
in Missouri.” Journal of Race, Ethnicity, and Politics 5(3): 450–80.
Hayes, Matthew, and Matthew V. Hibbing. 2017. “The Symbolic Benefits of Descriptive and Sub- Stantive Representation.” Political
Behavior 39(1): 31–50.
Headley, Andrea M., and James E. Wright. 2020. “Is Representation Enough? Racial Disparities in Levels of Force and Arrests by
Police.” Public Administration Review 80(6): 1051–62.
Headley, Andrea M., James E. Wright, and Kenneth J. Meier. 2021. “Bureaucracy, Democracy and Race: The Limits of Symbolic
Representation.” Public Administration Review 81(6): 1033–43.
Hogg, Michael A., and John M. Levine. 2010. Encyclopedia of Group Processes & Intergroup Relations. Thousand Oaks, CA: Sage
Publications.
Hong, Sounman. 2016. “Representative Bureaucracy, Organizational Integrity, and Citizen Copro- Duction: Does an Increase in
Police Ethnic Representativeness Reduce Crime?” Journal of Policy Analysis and Management 35(1): 11–33.
Hong, Sounman. 2017. “Black in Blue: Racial Profiling and Representative Bureaucracy in Policing Revisited.” Journal of Public
Administration Research and Theory 27(4): 547–61.
Hong, Sounman. 2021. “Representative Bureaucracy and Hierarchy: Interactions among Leadership, Middle-Level, and Street-Level
Bureaucracy.” Public Management Review 23: 1317–48.
Horowitz, J., Anna Brown, and Kiana Cox. 2019. Race in America 2019. Jardina: Pew Research Center.
Jardina, Ashley. 2019. White Identity Politics. Cambridge: Cambridge University Press.
Jenke, Libby, Kirk Bansak, Jens Hainmueller, and Dominik Hangartner. 2021. “Using Eye-Tracking to Understand Decision-Making
in Conjoint Experiments.” Political Analysis 29(1): 75–101.
Katz, Irwin, Joyce Wackenhut, and R. Glen Hass. 1986. “Racial Ambivalence, Value Duality, and Behavior.” In Prejudice, Discrimina-
tion, and Racism, edited by J.F. Dovidio and S.L. Gaertner, 35–59. San Diego, CA: Academic Press.
Keiser, Lael R., Donald P. Haider-Markel, and Rajeev Darolia. 2021. “Race, Representation, and Policy Attitudes in US Public
Schools.” Policy Studies Journal. https://doi.org/10.1111/psj.12443
Lawless, Jennifer L. 2004. “Politics of Presence? Congresswomen and Symbolic Representation.” Political Research Quarterly 57(1):
81–99.
Leeper, Thomas J., Sara B. Hobolt, and James Tilley. 2020. “Measuring Subgroup Preferences in Conjoint Experiments.” Political
Analysis 28(2): 207–21.
Lim, Hong-Hai. 2006. “Representative Bureaucracy: Rethinking Substantive Effects and Active Representation.” Public Administra-
tion Review 66(2): 193–204.
Lipsky, Michael. 2010. Street-Level Bureaucracy: Dilemmas of the Individual in Public Service. New York: Russell Sage Foundation.
Liu, Shan-Jan Sarah, and Lee Ann Banaszak. 2017. “Do Government Positions Held by Women Matter? A Cross-National Exami-
nation of Female Ministers' Impacts on Women's Political Participation.” Politics & Gender 13(1): 132–62.
Mansbridge, Jane. 1999. “Should Blacks Represent Blacks and Women Represent Women? A Contingent ‘Yes’.” The Journal of Politics
61(3): 628–57.
McLaughlin, Danielle M., Jack M. Mewhirter, James E. Wright, and Richard Feiock. 2021. “The Perceived Effectiveness of Collabo-
rative Approaches to Address Domestic Violence: The Role of Representation, Reverse-Representation, Embeddedness, and
Resources.” Public Management Review 23(12): 1808–32.
Meier, Kenneth J., and Tabitha S. M. Morton. 2015. “Representative Bureaucracy in a Cross-National Context: Politics, Identity,
Structure and Discretion.” In Politics of Representative Bureaucracy, edited by B. Guy Peters, Patrick von Maravić, and Eckhard
Schröter. Cheltenham: Edward Elgar Publishing.
Meier, Kenneth J., and Jill Nicholson-Crotty. 2006. “Gender, Representative Bureaucracy, and Law Enforcement: The Case of
Sexual Assault.” Public Administration Review 66(6): 850–60.
Mosher, Frederick C. 1982. Democracy and the Public Service. New York: Oxford University Press.
STAUFFER ET Al.
948

Mullinix, Kevin J., Toby Bolsen, and Robert J. Norris. 2021. “The Feedback Effects of Controversial Police Use of Force.” Political
Behavior 43: 881–98.
Myhill, Andy, and Ben Bradford. 2012. “Can Police Enhance Public Confidence by Improving Quality of Ser vice? Results from Two
Surveys in England and Wales.” Policing and Society 22(4): 397–425.
Nicholson-Crotty, S., J. Nicholson-Crotty, and S. Fernandez. 2017. “Will More Black Cops Matter? Officer Race and Police-involved
Homicides of Black Citizens.” Public Administration Review 77 (2): 206–16.
Olsen, Asmus Leth. 2017. “Human Interest or Hard Numbers? Experiments on Citizens Selection, Exposure, and Recall of Perfor-
mance Information.” Public Administration Review 77(3): 408–20.
Omi, Michael, and Howard Winant. 2014. Racial Formation in the United States. London: Routledge.
Oritz, Aimee. 2020. “Confidence in Police Is at Record Low, Gallup Survey Finds.” New York Times https://www.nytimes.
com/2020/08/12/us/gallup-poll-police.html.
Overby, L. Marvin, Robert D. Brown, John M. Bruce, Charles E. Smith, Jr., and John W. Winkle, III. 2005. “Race, Political Empow-
erment, and Minority Perceptions of Judicial Fairness.” Social Science Quarterly 86(2): 444–62.
Peyton, Kyle, Michael Sierra-Arévalo, and David G. Rand. 2019. “A Field Experiment on Community Policing and Police Legiti-
macy.” Proceedings of the National Academy of Sciences 116(40): 19894–8.
Phillips, Anne. 1995. The Politics of Presence. Oxford: Clarendon Press.
Philpot, Tasha S., and Hanes Walton. 2007. “One of our Own: Black Female Candidates and the Voters Who Support Them.”
American Journal of Political Science 51(1): 49–62.
Pitkin, Hanna F. 1967. The Concept of Representation, Vol 75. Berkeley: University of California Press.
Porumbescu, Gregory A., Suzanne J. Piotrowski, and Vincent Mabillard. 2021. “Performance Infor- Mation, Racial Bias, and Citizen
Evaluations of Government: Evidence from Two Studies.” Journal of Public Administration Research and Theory 31(3): 523–41.
Pratto, Felicia, Jim Sidanius, Lisa M. Stallworth, and Bertram F. Malle. 1994. “Social Dominance Ori- Entation: A Personality Varia-
ble Predicting Social and Political Attitudes.” Journal of Personality and Social Psychology 67(4): 741–63.
Reingold, Beth, and Jessica Harrell. 2010. “The Impact of Descriptive Representation on Womens Political Engagement: Does
Party Matter?” Political Research Quarterly 63(2): 280–94.
Riccucci, Norma M. 2021. “Applying Critical Race Theory to Public Administration Scholarship.” Perspectives on Public Management
and Governance 4(4): 324–38.
Riccucci, Norma M., and Gregg G. Van Ryzin. 2017. “Representative Bureaucracy: A Lever to Enhance Social Equity, Coproduc-
tion, and Democracy.” Public Administration Review 77(1): 21–30.
Riccucci, Norma M., Gregg G. Van Ryzin, and Karima Jackson. 2018. “Representative Bureaucracy, Race, and Policing: A Survey
Experiment.” Journal of Public Administration Research and Theory 28(4): 506–18.
Riccucci, Norma M., Gregg G. Van Ryzin, and Cecilia F. Lavena. 2014. “Representative Bureaucracy in Policing: Does it Increase
Perceived Legitimacy?” Journal of Public Administration Research and Theory 24(3): 537–51.
Riccucci, Norma M., Gregg G. Van Ryzin, and Huafang Li. 2016. “Representative Bureaucracy and the Willingness to Coproduce:
An Experimental Study.” Public Administration Review 76(1): 121–30.
Rocha, Rene R., Caroline J. Tolbert, Daniel C. Bowen, and Christopher J. Clark. 2010. “Race and Turnout: Does Descriptive
Representation in State Legislatures Increase Minority Voting?” Political Research Quarterly 63(4): 890–907.
Scherer, Nancy, and Brett Curry. 2010. “Does Descriptive Race Representation Enhance Institutional Legitimacy? The Case of the
US Courts.” The Journal of Politics 72(1): 90–104.
Schuck, Amie M. 2018. “Women in Policing and the Response to Rape: Representative Bureaucracy and Organizational Change.”
Feminist Criminology 13(3): 237–59.
Schuck, Amie M., Paola Baldo, and Christopher Powell. 2021. “Women in Policing and Legitimacy: A Vignette-Based Study of
Symbolic Representation.” Women & Criminal Justice 31: 1–18.
Schwindt-Bayer, Leslie A. 2010. Political Power and women's Representation in Latin America. New York: Oxford University Press.
Schwindt-Bayer, Leslie A., and William Mishler. 2005. “An Integrated Model of women's Represen- Tation.” Journal of Politics 67(2):
407–28.
Shoub, Kelsey, Leah Christiani, Frank R. Baumgartner, Derek A. Epp, and Kevin Roach. 2020. “Fines, Fees, Forfeitures, and
Disparities: A Link between Municipal Reliance on Fines and Racial Disparities in Policing.” Policy Studies Journal 49: 835–59.
Shoub, Kelsey, Katelyn E. Stauffer, and Miyeon Song. 2021. “Do Female Officers Police Differently? Evidence from Traffic Stops.”
American Journal of Political Science 65(3): 755–69.
Sidanius, Jim, and Felicia Pratto. 1999. Social Dominance: An Intergroup Theory of Social Hierarchy and Oppression. New York: Cambridge
University Press.
Sidanius, Jim, Stacey Sinclair, and Felicia Pratto. 2006. “Social Dominance Orientation, Gender, and Increasing Educational Expo-
sure 1.” Journal of Applied Social Psychology 36(7): 1640–53.
Song, Miyeon. 2018. “Gender Representation and Student Performance: Representative Bureaucracy Goes to Korea.” The American
Review of Public Administration 48(4): 346–58.
Stauffer, Katelyn E. 2021. “Public Perceptions of Women's Inclusion and Feelings of Political Efficacy.” American Political Science
Review 115: 1226–41.
Tate, Katherine. 2003. Black Faces in the Mirror: African-Americans and their Representatives in the US Congress. Princeton, NJ: Princeton
University Press.
POlICE DIVERSITY, POlICIES, AND OUTCOMES 949

STAUFFER ET Al.
950
Theobald, Nick A., and Donald P. Haider-Markel. 2009. “Race, Bureaucracy, and Symbolic Represen- Tation: Interactions between
Citizens and Police.” Journal of Public Administration Research and Theory 19(2): 409–26.
Treviño, A. Javier, Michelle A. Harris, and Derron Wallace. 2008. “What's So Critical about Critical Race Theory?” Contemporary
Justice Review 11(1): 7–10.
Van Ryzin, Gregg G. 2021. “The Perceived Fairness of Active Representation: Evidence from a Survey Experiment.” Public Admin-
istration Review 81(6): 1044–54.
Weber, Max. 1958. “The Three Types of Legitimate Rule.” Berkeley Publications in Society and Institutions 4(1): 1–11.
Weitzer, Ronald. 2002. “Incidents of Police Misconduct and Public Opinion.” Journal of Criminal Justice 30(5): 397–408.
Weitzer, Ronald. 2015. “American Policing under Fire: Misconduct and Reform.” Society 52(5): 475–80.
Weitzer, Ronald, and Steven A. Tuch. 2005. “Determinants of Public Satisfaction with the Police.” Police quarterly 8(3): 279–97.
Wilkins, Vicky M., and Brian N. Williams. 2008. “Black or Blue: Racial Profiling and Representative Bureaucracy.” Public Administra-
tion Review 68(4): 654–64.
Wilkins, Vicky M., and Brian N. Williams. 2009. “Representing Blue: Representative Bureaucracy and Racial Profiling in the Latino
Community.” Administration & Society 40(8): 775–98.
Wright, James E., and Andrea M. Headley. 2020. “Police Use of Force Interactions: Is Race Relevant or Gender Germane?” The
American Review of Public Administration 50(8): 851–64.
Yun, Changgeun. 2020. “Politics of Active Representation: The Trade-off between Organizational Role and Active Representa-
tion.” Review of Public Personnel Administration 40(1): 132–54.
AUTHOR BIOGRAPHIES
SUPPORTING INFORMATION
Additional supporting information can be found online in the Supporting Information section at the end
of this article.
How to cite this article: Stauffer, Katelyn E., Miyeon Song, and Kelsey Shoub. 2023. How
police agency diversity, policies, and outcomes shape citizen trust and willingness to engage. Policy
Studies Journal 51, 929-950. https://doi.org/10.1111/psj.12479
Katelyn E. Stauffer is an Assistant Professor of Political Science at the University of South Carolina.
Her research examines how the inclusion of historically marginalized groups in positions of power
influences public opinion towards political institutions and their members. Her research has been
published in various outlets including the American Political Science Review and the American Journal
of Political Science.
Miyeon Song is an Assistant Professor in the School of Public Affairs and Administration at Rutgers
University - Newark. Her research focuses on public management and policy implementation with an
emphasis on issues of quality and equity in public service provision. Her research has been published
in various journals, including the Journal of Public Administration Research and Theory, Public
Administration Review, and the American Journal of Political Science.
Kelsey Shoub is an Assistant Professor in the School of Public Policy at the University of Massachu-
setts, Amherst. Her research examines two broad questions: How do descriptive identities (e.g., race
and gender) of officials and civilians intersect with context to shape outcomes; and How does language
relate to policy and perceptions of politics? She has been published in Science Advances, the Journal
of Public Administration and Theory, and the American Journal of Political Science, among others.