Reimagining Public Safety: Defining “Community” in Participatory Research

Abstract

In the context of a national movement to defund police departments, many American cities are starting to reimagine public safety, as activists demand new practices that maintain safety while minimizing harm, as well as ensuring accountability when harms occur. Drawing on Everyday Peace Indicators methodologies, we argue that “community-centered” measurement, combined with researcher-practitioner partnerships, can help move both researchers and policymakers toward a more meaningful approach to policy design and evaluation. However, the application of community-centered measurement to the context of American policing raises important theoretical and practical concerns—in particular, the question of how community is defined, and who gets to define it. In this article, we ask: how do we define “community” in participatory research contexts where the concept of community is overlapping and contested? Using the example of a recent study carried out in the City of Oakland, we illustrate the complexities of applying a community-centered measurement process to the case of public safety and, more broadly, to police reform in American cities. We conclude with a discussion of both the benefits and limitations of our own approach, as well as a set of considerations for those engaging in participatory research.

The killing of George Floyd by Minneapolis police officers in May of 2020 sparked protests across the country, with Americans taking to the streets calling for justice for Mr. Floyd and the roughly one thousand others killed by officers in the United States each year (Jenkins et al. 2021). The persistence of police violence has eroded the public’s faith in officers tasked with protecting, serving, and fostering well-being in communities (Brenan 2020). Communities impacted by violence, including at the hands of police officers, are calling for meaningful reforms to how policing in America is designed and carried out.

In this broader context, a number of American cities have begun reimagining public safety. As they embark on this work, some are coalescing on a new vision, centered on upfront investments in collective well-being and a set of culturally appropriate wraparound systems. Advocates for this vision want to see effective and equitable policy reforms, resulting in practices that maintain safety while minimizing disproportionate harm to historically marginalized communities—and they want the processes that inform these changes to be both participatory and locally relevant (Black Lives Matter 2020). Yet, the ways in which conceptions of community are constructed, convened, and represented in these processes can be a source of significant complexity.

The term “community” invokes a sense of common purpose with an unquestioned positive valence, yet there is rarely agreement on its precise meaning (Levine 2017). This ambiguity allows the concept’s universal value to be strategically leveraged and manipulated in ways that are both complicated and consequential in the context of policing and public safety reform. This complexity stems largely from the fact that in urban American contexts, people belong simultaneously to multiple, overlapping communities, which are defined variously by identity, geography, ideology, membership, personal experience, and more. At the same time, we know that certain forms of communal belonging, such as racial identification and neighborhood of residence, have an especially significant influence on people’s experiences of policing and public safety (Braga, Brunson, and Drakulich 2019).

In this article, we discuss both practical and theoretical concerns related to defining “community” in participatory, community-based research. Specifically, we describe this issue as it relates to a collaborative research project carried out in the City of Oakland, California, which required us to navigate these issues in a dynamic, urban American setting. This is the first part of a broader project in which we are carrying out focus group discussions and community meetings in sites across California to learn how diverse communities experience and understand safety in their everyday lives. Central to this project is the idea that, in order to actualize meaningful public safety reforms, those most impacted by reforms must play a key role. Individuals in communities that are simultaneously overpoliced and underserved by the current public safety infrastructure, especially people of color and low-income residents who are disproportionately impacted by violence, have distinct experiences and understandings of what being and feeling safe entails. Their “webs of meaning” will be critical to reshaping the system, and to holding a reimagined system accountable to their needs.

The Everyday Peace Indicators (EPI) research process offers a structured mechanism for arriving at local, community-driven definitions of safety (Mac Ginty 2013). EPI is a conceptual approach and field-tested methodology for generating participatory, community-centered indicators. Rather than imposing categories, meanings, and predefined choices about measurement that are conceptualized by researchers or policymakers, EPI allows the construction of meaning to emerge from community members themselves. The EPI process thus helps translate the “insider’s perspective” into measurable indicators. These metrics can then be used to craft a rigorous plan to identify, pilot, and evaluate reforms, facilitating policy changes that fulfill local desires for safety and justice.

In this sense, the EPI process offers a unique opportunity to design a set of indicators for evaluating recent efforts at reimagining policing, allowing cities and their stakeholders to track the success of policy changes as measured by citizens’ own experiences of safety. However, the application of a community-centered model to assess changes in American policing makes explicit several thorny issues that often remain hidden in more traditional, outsider-driven measurement approaches. In particular, questions about how “community” is defined become immediately apparent and especially important in this context. These concerns are often insufficiently addressed, even in traditional policy processes that seek community input (Cheng 2022).

In this article, we explore this key methodological issue in the application of EPI to the case of public safety and, more broadly, to police reform in American cities. We begin by situating our motivating questions in the context of the EPI methodology, providing a brief review of the potential application of the process to the case of police reform in an American city. We then lay out the specific challenges of defining community in this context.

We next explore how “community” has been conceptualized and articulated in the City of Oakland, where we situated a recent EPI project to source community indicators of safety and well-being. To do this, we draw on interviews with key stakeholders who are deeply engaged with issues of policing and public safety in the city. We also draw on descriptive data gathered from all public input to the city’s formal process of “Reimagining Public Safety (RPS),” including transcripts of listening sessions, voicemails, and emails related to the city’s Public Safety Task Force.

Our analysis from Oakland complicates the idea that a community’s own definition of its membership and boundaries can easily serve as the starting point for participatory research. Instead, narratives from our interviews and the city’s RPS process make clear that “community,” as defined and articulated by city residents is both contested and complex. We discuss the implications of these multi-dimensional and overlapping definitions of community for how community-generated indicators of safety might be created in cities like Oakland, and ultimately the important role that conceptions of community play in how reforms can be evaluated and assessed. To conclude, we describe our own decisions about how best to resolve these questions in our Oakland work, and lay out next steps for a longer-term agenda to design new ways of engaging communities in conceptualizing and evaluating public safety.

OPPORTUNITIES AND CHALLENGES FOR PARTICIPATORY POLICE REFORM

Over the past decade, a national conversation has re-emerged about police reform. Evolving from the Black Lives Matter movement and building on a long history of struggle for racial justice and equity in America, activists around the country are demanding recognition of the historical roots of policing in slave patrols, a reconciling of whether state power through policing is fundamentally at odds with restorative justice, and a wholescale reconceptualization of community safety and the role of police (Lebron 2017). These efforts remain nascent, but have cohered around a set of “asks”: broadly speaking, these include reductions in funding for armed police and a transition of police resources and responsibilities to alternative, health-centered agencies and organizations (Black Lives Matter 2020).

As researchers, we seek to better understand how these reforms are experienced by relevant stakeholders, as well as how reforms might impact a range of outcomes. To begin building an evidence base about the effectiveness of specific reforms, we must evaluate how these reforms are designed and implemented, as well as their short- and longer-term effects. More immediately, however, we must also begin to develop a theoretical and empirical language that takes seriously the conceptual shift activists are calling for, which seeks to move toward a more robust, community-centered, and proactive system of public safety. But how do we evaluate reforms of this kind against the benchmarks that reformers care about? Answering this question will be critical to understanding whether the intentions of reform are realized in their implementation.

In this article, we argue that it is useful to bring a broader conceptual and methodological framework to this work than historically has been the standard for evaluating policing and public safety reform. Existing measures of safety are frequently limited, generally focusing on crime rates, recidivism, or the ability of police to identify and apprehend those suspected of criminal activity. While significant, these indicators are limited in several important ways. First, they largely capture the absence of a negative (crime), rather than the presence of a positive (safety). Second, they are narrowly focused on the activities of the state in response to crime, rather than the experiences of residents when safety is present, or the community-level factors that might signal crime or reflect conditions of safety. In these ways, existing metrics are out of sync with the emerging goals of recent and broader community-led calls for reform.

Other indicators used to evaluate safety get closer to the types of outcomes being emphasized by community advocates, but by themselves are likewise inadequate to the task of measuring deeper and more systemic reforms. One such metric is the degree to which residents trust police, frequently captured through either snapshots or longitudinal surveys of residents. Like crime rates, this measure centers police activity as the primary indicator of safety, a focus that the current reform efforts aim to de-emphasize. A second subjective measure is more directly community-centered, usually taking the form of a survey question like, “How safe or unsafe do you feel in your neighborhood at night?” or “How safe, if at all, would you say your local community is from crime?” While these measures begin to shift the focus toward the needs of community members, they still capture only a narrow conception of safety, which fails to account for the wide range of ways residents might assess their own safety, well-being, and health.

The relatively limited scope of these metrics becomes that much more important when we consider their utility for evaluating specific reforms. In particular, if we limit the focus to crime rates or fear of crime, or trust in police only, we miss an opportunity to think about how new policies and practices can build toward a richer social well-being, as called for in the public health approach to violence prevention (Butts et al. 2015). As Sally Engle Merry notes, “The turn to indicators has the effect of defining [outcomes] narrowly in terms of specific accomplishments rather than as structural change…. Broad goals, such as ‘access to justice for all,’ are measured by narrow and limited measures which fail to do justice to the conceptions behind the goals” (2019, 146). Even metrics that are ostensibly sourced from community members but are constrained by existing norms and power dynamics, such as those emerging from policymaker-sponsored listening sessions or police department-initiated town halls, can fail to reflect the wide range of alternatives to traditional policing that some residents might desire (Cheng 2022).

In contrast to these traditional metrics, the EPI process offers a well-defined way to source alternative indicators that are potentially better suited to measuring progress toward systemic, structural, and holistic reforms. EPI emerged over the last decade as a response to discontent with existing measurement systems amongst policymakers, researchers, and practitioners working on peace and conflict, who recognized the substantial limitations of existing ways of understanding and measuring complex concepts related to peace and conflict (Firchow and Mac Ginty 2017, 6–27). Specifically, scholars in these areas began to question the potentially problematic role of elite frames and top-down metrics that historically dominated political and policy debates. In response to these concerns, scholars and practitioners have devoted energy to developing community-sourced methodologies in peacebuilding, stabilization, and humanitarian contexts (Khan and Nyborg 2013). These efforts are mindful that top-down data-gathering often lacks the conceptual clarity, nuance, and granularity required to fully understand the experiences of others (Chabal 2012).

The EPI approach draws from these efforts to develop a deeper, ethnographic understanding of multiform problems. At the same time, it also recognizes the value of quantitative methods in generating representative and policy-relevant knowledge. Inspired by critical environmental studies (Nordstrom 1997; Miller 2005), ethnographic work associating “local voice” with authenticity and accuracy, and development studies using participatory research methods (Krimerman 2001; McIntyre 2007), the EPI project establishes a bridge between interpretivist, qualitative work on local measurement and quantitative work in the tradition of “participatory numbers” (Chambers 2007).

The result is not only an alternative to conventional research design, but a robust methodology that offers the rigor and replicability of quantitative measurement while incorporating the nuance, context, and local concept development typifying qualitative, participatory approaches. Where traditional indicators can reinforce the subaltern position of researched communities, the EPI approach affords participants a voice and role in the research process itself. ¹ Where universal measurements are limited in their recognition of subtle differences at small-scale levels, the EPI methodology is designed to flexibly assess differences at the meso-level, whether between villages, neighborhoods, civic groups, or other sorts of “communities” relevant to research and reform. To date, EPI has been used as both a research and evaluation tool, drawing on local, participatory understandings to assess diverse interventions, policies, and projects (Firchow 2017, 2018).

CONCEPTUALIZING COMMUNITY IN POLICE REFORM

EPI offers a compelling resource for public safety reforms, but also raises difficult conceptual questions, particularly in regards to how communities are defined, convened, and represented in research and policy processes. EPI emerged in the field of peace and conflict studies, and has been traditionally applied in largely rural areas in countries with histories of violent armed conflict, such as Colombia, Sri Lanka, Afghanistan, and Uganda (Mac Ginty 2014; Firchow 2018; Firchow and Urwin 2022; Dixon and Firchow 2022). In such contexts, the relevant communities are defined in coordination with research subjects often as villages or village centers (a “geographic” community), whether corresponding with official designations or following informal definitions. In these rural contexts, the identification of the salient community is often relatively straightforward because the boundaries of the community are commonly understood and agreed-upon by all. A representative sample of this community is then invited to participate in the indicator generation process.

In dense urban settings, however, defining community poses a set of thornier decision points. In diverse American cities, identity groups are frequently loose and overlapping, and social capital is often low and inconsistent, complicating the very concept of “community.” Indeed, one of the pernicious effects of violence in urban settings can be the breakdown of communal bonds. Urban areas marked by low socioeconomic status, high rates of residential mobility, and substantial resident heterogeneity are particularly likely to experience difficulty sustaining and enforcing widely shared community norms (Sampson 1986). More broadly, some forms of social capital have declined in the United States as a whole, particularly with respect to membership in traditional associational organizations (Putnam 2000). Likewise, over the past several decades American public opinion has shown consistently declining levels of trust in institutions, government, leaders, and other people (Rainie and Perrin 2019). In the United States today, people’s identities (and the various communities to which they belong) are often multifaceted: different identities hold different levels of salience; the salience of identity groups can change over time and across contexts; and multiple identities can often be comfortably maintained.

The multifaceted and overlapping nature of identity groups in American cities complicates the clear delineation of the boundaries of any particular community. Numerous scholars have suggested that externally defined geographic, economic, and cultural markers of community obfuscate many critical social and political nuances experienced by community members, especially among marginalized and less traditionally structured groups, and often do not reflect the ways in which individuals conceive of their communities (Li 1996; Jewkes and Murcott 1998; Merzel and D’Afflitti 2003; Mathie and Cunningham 2003). Still, policymakers and ordinary citizens alike frequently invoke the term “community” in arguments related to public policy. As Levine notes, when invoked in this way, the term “community” represents “more than a place or a group of people; it signifies the common good, a valued entity” (2017, 1156). While most agree on the value of community, however, its bounds are rarely clearly defined or consensual. As such, the term can paradoxically sap community members of political power when its ambiguity is leveraged to obscure political conflict, mask exclusion, or facilitate claims of representation.

In discussing the importance of defining community to the project of police reform, Daniel Flynn writes:

The police definition of what constitutes a community must conform to parameters within which each police department must function. These parameters include jurisdictional boundaries, division of labor within the department, service demand and reporting requirements. As a result, over time, the police have solidified a paradigm of community that generally is limited to residential and business/residential neighborhoods. Shifting that paradigm to a more generic paradigm of community allows the police to develop new applications of the community policing strategy in non-traditional communities, those with shared geography, character or identity and common concerns or problems. By recognizing that non-traditional communities need not be primarily residential or permanent, police departments can derive the full benefits of the community policing strategy in the community structures that make sense for each jurisdiction (1998, 18).

At the same time, it is by now well established that certain racial communities bear the brunt of police surveillance. African Americans are more likely to be stopped and questioned by police (Edwards, Lee, and Esposito 2019; DeVylder et al. 2022). Black and Latino youth in particular are significantly more likely to report having involuntary contact with law enforcement (Crutchfield et al. 2009). However, racial disproportionality in policing is also tied to racial segregation and racialized poverty; in cities across the country, areas of concentrated disadvantage are frequently more heavily policed. In these areas, police are more prevalent and more adversarial (Smith 1986; Kane 2002; Terrill and Reisig 2003; Weitzer and Tuch 2006; Roh and Robinson 2009).

The issue around defining community is equally complex when establishing methods for community-based participatory research. Finding good representatives from the community requires definitions of both “good representative” and “the community,” which runs the risk of oversimplifying the community, biasing participation, and leaving out crucial groups of people that may be less vocal and particularly marginalized (Jewkes and Murcott 1998). The notion that there is a “good representative” that is a “typical specimen” of a community relies on the idea that a community is homogeneous (Jewkes and Murcott 1998). Yet even in clearly delineated communities, disagreement is common, and high levels of consensus cannot be assumed (Levine 2021).

The participatory research tradition defines community as “a sense of identification and emotional connection to other members, common symbol systems, shared values and norms, mutual—although not necessarily equal—influence, common interests, and commitment to meeting shared needs” (Israel et al. 1998, 178). Yet community-based research studies tend to define “community” in a functional capacity for each study. This has resulted in a plethora of factors and approaches used to identify different communities, including language (Ugolini 1998), occupations (Tonks 1999), social class (Reid 1999), denominal affiliation (McLaughlin 2002), kinship networks and community organizations (Dikeni, Moorhead, and Scoones 1996; Cooper 1998). Unsurprisingly, many studies define communities based around geography (Luginaah et al. 2001; Merzel and D’Afflitti 2003), but the geographic approach is often unable to reflect the political dimensions of a community. Indeed, especially in urban contexts, researchers often conflate a geographic neighborhood with community without acknowledging that the simple fact of sharing space does not guarantee the presence of a shared sense of community between residents (MacColman and Dickenstein 2022). Moreover, utilizing such simplified notions of community can inadvertently cause harm when creating specific policy recommendations, as many less organized groups, such as immigrants, are easily left out (Li 1996).

When applying EPI to public safety in the American context, we are therefore faced with how to weigh the myriad potentially relevant communities from which we could begin the research process. As a starting point, we return to the community-driven approach that is central to the EPI framework: let research subjects’ own definition of what constitutes community take precedence (Firchow 2018). In practice, this means carefully identifying the spaces within a given community where public safety issues are discussed, and analyzing these discussions to understand how individuals participating in these discussions define the communities relevant to reimagining safety. It also requires paying attention to spaces that are impacted by violence, where the effects of public safety reform will be felt, but where public discussions may not be happening. Yet even this approach can easily become tautological, and might raise more questions than answers. Namely: who is part of the “community” that gets to define the scope and boundaries of the community itself? Public safety reform in the City of Oakland provides an illustration of the relevant challenges.

Conceptualizing Community: Data from the City of Oakland

With a population of over 430,000 residents, the City of Oakland is a vibrant, diverse, and complex city with a rich cultural history. Home to the Black Panther movement of the 1970s and a historically large African American community, the city has struggled since the 1980s with high rates of violence, tensions from urban gentrification, and a deepening housing crisis. The city’s Black population has declined by nearly half in recent decades, from nearly fifty percent at its peak to roughly twenty-eight percent of residents today. Like many cities across the country, Oakland is also confronting longstanding challenges with policing and public safety. The city’s police department has been subject to independent monitoring since 2003 following a federal lawsuit, and in 2009 the transit authority (BART) police were involved in the high-profile officer-involved shooting of Oscar Grant at Oakland’s Fruitvale station.

In July of 2020, the city initiated a formal process of stakeholder engagement centered around the Reimaging Public Safety Task Force (City of Oakland 2021a). The Task Force was designed “to rapidly reimagine and reconstruct the public safety system in Oakland by developing a recommendation for Council consideration to increase community safety through alternative responses to calls for assistance, and investments in programs that address the root causes of violence and crime (such as health services, housing, jobs, etc.), with a goal of a fifty percent reduction in the OPD General Purpose Fund (GPF) budget allocation.”

Throughout its work, the Task Force sought input from members of the Oakland community, through surveys, email, and listening sessions. The Task Force presented their forty-eight recommendations to the Oakland City Council on May 3, 2021 (City of Oakland 2021b). Yet the ensuing budget negotiations between the mayor and the city council ultimately led to only $17.5 million in cuts to OPD’s budget over the next two years, which fell far short of the proposed fifty percent reduction. As a point of reference, Oakland’s proposed 2021–2022 budget allocated roughly $340 million to the city’s police department (City of Oakland n.d.). It is thus unclear how many of the Task Force’s recommendations will be implemented, and on what timeline.

To understand how the concept of “community” is utilized by stakeholders in the context of Oakland’s public safety reforms, we drew on two data sources. First, using an inductive coding process, we analyzed public input to the RPS Task Force, including recordings of four ninety-minute virtual listening sessions that were widely publicized throughout Oakland, as well as six voicemail transcripts and 245 emails sent from the public to the Task Force, all of which were made public on the RPS website. Our research team listened to the RPS listening session recordings and reviewed the emails and voicemail transcripts, taking notes on areas where residents spoke about Oakland’s “communities” in ways that included racial groups, linguistic groups, neighborhoods, and/or the city as a whole.

Second, we drew on semi-structured interviews that we conducted through video calls in the summer of 2020 with eleven individuals who have deep knowledge about the city’s public safety reform efforts. Our goal was to develop a sense of how these individuals define the “communities” in Oakland that are most heavily affected by violence and public safety reform. These individuals were contacted through a snowball sample of people who are actively working on issues of police reform and public safety in Oakland, whether through their professional employment, activism, or research efforts. Of the eleven interview subjects, eight were male and three female. The group was racially diverse: we characterize four as white, four as Black, two as Asian, and three as Latinx.² They also represented a diversity of professional fields, including law, research, public safety, violence prevention and racial justice. Seven worked or volunteered with public offices, five were directly engaged in activism around public safety or racial justice, and three or more had direct lived experience of the criminal justice system (including individuals who had been victims of crime or police harassment, and those who were formerly incarcerated). Our interviews lasted between forty-five and seventy minutes, and were attended by at least two members of our research team, with one taking primary responsibility for note taking. The interview topics ranged from the subject’s prior work in the City of Oakland, the constituents they worked with or represented, their thoughts about the goals and direction of the ongoing RPS process, and the communities they thought would most likely be impacted by any reforms.

Community as Contested, Complex, and Overlapping

Across both our interviews and other sources, we found a complex discussion of overlapping geographic, identity-based, and associational communities. One common usage of community was in reference to the entire populace of Oakland as an undifferentiated whole. This was especially common in both emails and community comments in the RPS process where the individual was contacting the Task Force to urge them not to cut OPD’s budget. In our own interviews, stakeholders identified these vocal individuals as a group with the most at stake because they feel “the most preyed upon.” In the RPS process, most of these individuals expressed the belief that more police will mean more safety for the community, which they do not differentiate but instead tend to equate with people like themselves. If there is an “us” and a “them,” in this set of comments, the community is “us” and criminals are “them” and the police are needed to protect “us” from “them.” While many of these people expressed concern about crime in their particular neighborhood, they tended to talk about the entire Oakland community without any differentiation. Many of their messages juxtapose this community, which they see themselves as prototypical members of, against people they do not include in their “circle of we.” Often, these outsiders are referenced simply as “the criminals,” but other references include mention of people who are experiencing homelessness or struggling with mental illness.

The tendency to refer to the community as an extension of the self was not limited to people who want more police presence. Many members of the public on the other side of the police reform issue during the RPS process similarly refer to the Oakland community as an undifferentiated monolith. For example, one person wrote in an email, “we should invest in police alternatives and ways to build up and heal our community…. Police do not keep our community safe.” Here, however, the boundaries of the “circle of we” manifest differently. Now, the community is the “us,” and the police are the “them.” In this view, rather than facilitating safety, the police are either directly threatening community safety or, at the very least, are sapping resources that could be better used to support the community. We see this latter view reflected again in this comment: “Safe communities are whole communities with self-determination. The Task Force should send recommendations to our City Council that invest in alternative and replicable models for public safety, that redirect resources from policing to the people, and that empower the community to take care of ourselves.”

Many of the RPS references to “community” as an undifferentiated whole are coupled with references to the multiplicity of communities that comprise Oakland. Take, for example, this comment: “I urge the Task Force to agree to cut OPD services by 50% to reimagine public safety for our community. When police are dropped into communities to supposedly protect the people living in them, it has led to violence and often leaves the community feeling less safe than before.” The writer of this email recognizes that Oakland is made up of multiple communities, yet suggests that the entire community of Oakland is affected by police violence. Similarly, one person who referred to the Oakland community as an undifferentiated whole later wrote, “I want to re-emphasize my support for recommendations that invest in alternative models for public safety and redirect resources from police to people and communities.” This demonstrates that the person simultaneously thinks of Oakland as a single community and also perceives that it is made up of a variety of communities.

Yet, while we found a large number of RPS references to Oakland as a single, monolithic community, commenters more frequently referred to community in more specific terms. Throughout, we found references to particular communities within Oakland, defined by both geographic and identity-based markers. Geographic references to the various communities that make up Oakland were highly prevalent during the RPS process. In our interviews, our respondents similarly referred to particular communities by district number or other geographic markers. In the emails and public comments, individuals often located themselves geographically within Oakland prior to discussing their community. Many of these used specific neighborhood names, such as Eastlake or Oakmore, but they also often referenced the administrative units in which they live, like the district number (e.g. D2 or D7) or even named the specific police beat (e.g. 17x or 26y). Individuals also used geographical references to specifically identify other Oakland residents’ neighborhoods. Throughout the city’s conversation about reform, we found references to the differing lived experience of people in the more affluent and whiter “hills” and of those living in “the flats.” One of our interviewees specifically mentioned Highway 580 as the dividing line, and one speaker at a listening session likewise referred to this division when asserting that there are “two Oaklands.” She explained that “there is a side that is fully resourced, and fully valued, and there is another part of Oakland that is under-resourced and under-valued.”

Definitions of community based on racial or ethnic identities tended to be invoked when referencing groups that are negatively impacted by the police. This was especially true in our own interviews when we asked respondents who is most likely to be impacted by any reforms. In answering this question, most of our interview respondents referenced communities of color, such as “Black and Brown males who are more likely to be stopped by police.” In the RPS data, one email described OPD as “an entity that, by design, actively harms our most vulnerable and historically marginalized neighbors.” In doing so, this person references particular identity groups who constitute the community whose safety needs must be considered. Another email called for cultural competency from the people charged with providing public safety. It asked for “neighborhood patrols with people who actually live there and speak the languages of the area and are respectful of the culture.” Here, we see an even more explicit recognition of the importance of cultural identities in defining the bounds of community.

Some of these identity-based discussions in the RPS process referenced divisions between identity groups within Oakland. Several of these pointed to recent targeting of Asian-Americans. One writer noted “important nuances when it comes to race-based violence between Asians and other races.” In discussing this conflict, the writer claimed that the RPS process itself inflamed tensions between the groups: “Suggesting more investment in the black (sic) community as the solution to anti-Asian violence betrays a lack of compassion and understanding. It is especially inappropriate considering Yahya Muslim, a Black man, was recently charged with attacking three Asian seniors in Oakland’s Chinatown.” Another example of references to inter-community tensions noted its long-standing nature: “Anti-Asian violence is not a new phenomena (sic) in the Eastlake district. Ten years ago there was a group of young African-American youth who were attacking elderly Asian women for there (sic) gold chains. They also engaged in robberies of undocumented day laborers because they were less likely to call the police and usually were paid in cash from their jobsites. One of my neighbors was engaged in this before he was sentenced to Santa Rita. He joked that he was ‘amigo jacking.’”

In this quote, we also see community safety as bounded by the extent to which community members are able to access police services without fearing for their own safety or livelihoods. Here, the “circle of we” excludes members of other identity-based communities as well as the police. In our interviews, one subject referenced the fear that older members of the Black and Brown communities feel toward each other, noting especially that “they are afraid of each other’s children.” This interview respondent expressed hope that these communities might be unified by their shared experiences of police misconduct directed toward their community members, but that this unity is hampered both by the fear they feel toward each other, and the difficulty of speaking about it honestly because it is not politically correct.

Much discussion of community also highlighted the intersection of geographic and identity-based definitions of community. Amongst our interview respondents, we noted a tendency to use geographic references as a stand-in for the racial and ethnic groups that predominantly reside in different parts of the city. For example, “Fruitvale” was a common shorthand for referencing the Latinx community, and “Chinatown” was used to reference Oakland’s Asian residents, while “West Oakland” and “East Oakland” were largely used to reference Oakland’s Black communities, and mention of “the hills” was a stand-in for the white community. Similarly among the RPS responses, a handful of emails from the RPS process highlight the intersection of geography and identity, and the historical role of racist policy, such as red-lining, which constructed Oakland’s racially segregated neighborhoods.

This nuanced understanding is evidenced in an email that urged the Task Force to recognize that “neighborhoods like East Oakland share a commonality of a history of racial discrimination, stigmatization, and lack of investment, largely because of the practice of red-lining, and this has long-term physical and mental consequences. When communities are invested in and trusted rather than policed and profiled, they’re safer and healthier. Please listen to this call and reinvest 50% of the OPD budget into community-led programs and resources. Let people who are impacted by racism and police violence be the ones to take care of themselves and heal, or else we will only perpetuate a cycle of pain, animosity, and lost life.” Here, the community that is invoked is clearly racially defined, but it is also geographically bounded.

Yet there is also evidence that rapid gentrification in formerly red-lined neighborhoods in Oakland has given rise to a more complex intersection of identity and geography in these areas. In the RPS process, sometimes this is referred to as the distinction between “new Oakland” and “old Oakland,” though it is rarely stated explicitly as such in the Task Force’s public documents. This division is evidenced primarily by writers’ and speakers’ references to the depth of their roots in the community. Sometimes, speakers in the RPS process referred to themselves as “born and raised” in a particular Oakland neighborhood, while others noted the length of time prior generations of their family had lived in the city. When speakers and writers mark themselves as belonging to “old Oakland” in this way, they tend to leave the contrasting group unstated. In some RPS emails or public comments, it is possible to discern that the person might be classified as “new Oakland,” usually through references to their recent arrival in a neighborhood, yet these commenters rarely acknowledge the division between themselves and the longstanding community in their neighborhood. In our interviews, one subject described the explicit distrust some community members expressed when interacting with Task Force members without deep roots in Oakland.

Finally, in addition to the geographic and identity-based definitions of community, there are clear ideological divides that manifest in individuals’ associations with community organizations that are engaged in political mobilization. While these positions are often simplified to being pro- or anti-police, a number of our interviewees cautioned us not to see the dividing lines that simply. Within the groups that are seeking changes in the delivery of public safety, for example, some are striving to defund OPD in order to completely restructure the delivery of public safety, while others are working to enact reforms to improve OPD’s delivery of public safety.

Another important complication is the ways that these associational communities intersect with geographic and identity-based groupings. As one of our interviewees put it, “people are coalescing around ideology more than race.” While a large number of public voices during the RPS process calling to defund the police are coming from those who are disproportionately affected by the current system, there is a sizable contingent of voices against cuts to the police budget coming from Black people who are concerned about a range of issues, from petty crime to gun violence. One contributor to the RPS process, who identified herself as a Black resident of a predominantly Black neighborhood, wrote: “If funding for OPD is reduced in an attempt to combat police brutality to the point that my life is in danger because there are no officers on the street to provide the public safety protection I require as a resident of District 6, that is extremely short-sighted, unjust, and unfair.” In the RPS community listening sessions, there was a chorus of similar voices who spoke of their frustration with a lack of police engagement in their community. They complained about police allowing drug trafficking to take place in broad daylight, of rising homicide rates, and of police that do not show up even when they are called. As one of our interviewees put it, these people want someone to respond when they call for help, but “they don’t want to be brutalized.”

PUTTING “COMMUNITY” INTO PRACTICE

As discussions in the City of Oakland make clear, researchers confront significant challenges to defining “community” in the urban American context. In particular, implementation of an EPI project or other form of participatory research is made difficult due to the complex and interconnected geographic, racial, socioeconomic, and associational identities within Oakland’s population. In seeking to carry out a community-centered research process in this context, we are thus keenly aware of the ramifications of adopting any one particular definition of “community.” When there are complicated and overlapping communities, any bounded definition of community necessarily excludes members of other, related communities. Given this, how should we go about selecting communities with which to work?

One possible starting point would be to pick a representative sample of residents and workers in the city, accounting for as many demographic and geographic characteristics as is feasible. However, we are quickly faced with the fact that these demographic and geographic classifications are externally defined. How useful are these groupings in defining the relevant community for this purpose, if individuals in these groups do not see themselves as part of the same community, do not have a strong sense of shared identity or belonging, or do not interact in socially or politically meaningful ways? Will a representative sample of this kind actually get us closer to establishing metrics that can help build and sustain a new system of public safety? Or would defining communities using a traditional approach of this kind merely replicate the issues associated with standard, top-down forms of conceptualization and measurement? As we have noted, externally defined markers of community can obscure key differences and vary from individuals’ own conceptions of community. We do not suggest doing away with demographic categories. Rather, we used demography as a data point while seeking out potential civil society partners, and we then worked with these groups to define specific communities within the populations they serve or represent.

Our decisions about how to identify partners were ultimately focused on two basic principles. First, we sought partners for whom the implications of our key concepts, “safety” and “well-being,” were likely to be most deeply consequential: those who have had disproportionate and frequently negative experiences with local law enforcement, who have been most directly harmed by the city’s existing public safety infrastructure, and who experience disproportionate rates of violence and crime. This approach draws from theories like the “curb cut effect,” which posit that systems designed specifically to address the needs of the most vulnerable can result in positive impacts on society as a whole. Angela Glover-Blackwell describes this phenomenon with respect to the Americans with Disabilities Act, which required changes to the built environment in order to ensure access and mobility for those with different physical abilities. As she notes, when features like curb cuts were introduced,

everybody benefited—not only people in wheelchairs. Parents pushing strollers headed straight for curb cuts. So did workers pushing heavy carts, business travelers wheeling luggage, even runners and skateboarders. A study of pedestrian behavior at a Sarasota, Fla., shopping mall revealed that nine out of ten “unencumbered pedestrians” go out of their way to use a curb cut. As journalist Frank Greve has noted, the barricades stormed by disabled advocates in Berkeley 40 years ago were a few inches high, “yet today millions of Americans pass daily through the breaches” (Blackwell 2017).

Second and related, our selection of partners was guided by a belief that those most impacted by policy reforms should have a voice in their design. As such, we sought Oakland residents who are not typically “in the room,” whether in formal town halls or city council meetings, at the ballot box or with direct access to policymakers, or even through leadership positions in the civil society groups that most often represent communities in public discourse (Táíwò 2020).

While we aimed to carry out our selection of participants for the indicator generation process according to these basic principles, we did not have a pre-defined set of community partners. Rather, we strove to identify local civil society organizations most interested in working with us and then to collaborate with these organizations to recruit the individuals they serve, according to their own definitions of community. To do so, we compiled a list of relevant organizations, using both desk research and snowball sampling, beginning with organizations suggested by our interviewees. In total, we reached out to forty-two of these organizations and introduced ourselves via email as looking for a partner along three lines.

First, we sought partners with a strategic and/or programmatic interest in our work and the sorts of activities and research we would carry out. Second, we sought partners who serve or represent their constituencies at a local level, so they could help us understand the community or communities with which they work in detailed ways. Finally, we sought partners with the logistical capacity and know-how to help us organize both small focus groups and larger town-hall meetings. This included helping us select and mobilize participants, organize food, and rent or find spaces.

The majority of our initial conversations with potential partners took place via video calls. In these meetings, we clarified that we were seeking partnerships that would be mutually beneficial and productive for both sides. We explained that we were seeking to enter into formal partnerships in which we could provide financial support for partners’ time and effort, including participation incentives for community members who participated in our research activities. In addition, we offered a research process that could help engage community members around significant, complex issues related to community safety and well-being; a set of data and research findings that could help organizations with their own advocacy and programmatic efforts; and, where there was interest, we could offer partnerships for writing, presentations, and other public communications.

Ultimately, out of the forty-two organizations we contacted about a potential partnership, we received invitations for further conversation from thirty-two of them and eventually partnered with six. While some of our eventual partners expressed immediate interest in working with us, others required several conversations to understand our team and the project’s broader goals in more detail. Our team’s racial composition (primarily white and Latinx), backgrounds (from higher education), funding sources (from a private funder), and research goals (to publish and widely share data and results) were all topics of discussion. In these discussions, we clarified that we were working independently from the City of Oakland, but that we planned to share our results with policymakers. We also discussed both co-ownership of intellectual property and the co-creation of research design, which we assured partners we were wholly committed to. Reasons for declining the partnership primarily included insufficient capacity and, relatedly, concern that the partnership would detract from organizations’ core work. In a few cases, organizations also expressed concern about our team’s lack of diversity and about the potential of our research to distract from more established advocacy efforts in the city.

As we began securing initial partnerships, we became more purposive in our outreach to additional potential partners, to ensure that we would reach a wide range of relevant demographic and geographic groups in Oakland. In particular, none of our earliest confirmed partners specifically served Oakland’s Asian or Latinx communities, so we actively sought partners who could reach those populations. Again, though, we used these pre-identified descriptive groups only as a starting point to identify civil society organizations. Once identified, we worked out the specific contours of community with each organization, according to each organization’s own definitions. This led to novel constructions of community that we do not typically reach in our more traditional, geographically or demographically bounded research, such as communities constructed around shared experiences of violence and contact with the criminal justice system.

Ultimately, we established six partnerships with a diverse set of local organizations (see Table 1). These organizations then helped us reach ten different self-defined communities of Oakland residents, from whom we could gather slices of everyday life in the city. Crucially, these communities were defined as groups of people who saw themselves as having some kind of shared everyday that is informed by common experiences, common environments, or other collective realities. Our Oakland partnerships led to communities that were defined across geography (neighborhoods, housing developments, proximity to parks), identity (shared race, ethnicity and/or language), association (membership in institutions like churches or civil society organizations), and shared experiences (incarceration, experienes with criminal justice, or victimization).

Table 1. Oakland Community Partners

The resulting set of organizations do not constitute a representative sample of Oakland residents, but are instead a subset of groups who are directly impacted by violence and whose voices are not typically heard in policy conversations or public discourse. This is by design, as our goal in this broader EPI project was to create a set of indicators that could measure the everyday experiences of safety and well-being for those who are both most vulnerable and historically underserved. Returning to Blackwell’s curb cut effect, we hope that by focusing on policy reforms that address these groups’ needs, we can strive to have wider-ranging positive impacts for the population as a whole:

There’s an ingrained societal suspicion that intentionally supporting one group hurts another. That equity is a zero sum game. In fact, when the nation targets support where it is needed most—when we create the circumstances that allow those who have been left behind to participate and contribute fully—everyone wins. The corollary is also true: when we ignore the challenges faced by the most vulnerable among us, those challenges, magnified many times over, become a drag on economic growth, prosperity, and national well-being (2017).

CONCLUSION

Our work in the City of Oakland sheds important light on the potential complexities—but also the enormous potential—of using community-centered, everyday indicators as a tool for evaluating policy reforms, both in policing and in other domains. First, it is clear that how researchers define community, as well as how they ultimately choose representatives with whom to engage, has consequences for what conclusions can be drawn and whose voices and experiences are prioritized. This is especially important when it comes to community indicators as part of a broader strategy for evaluating policy change. Different conceptions of community are likely to produce different indicators. In turn, these indicators will point toward different visions of what constitutes success. A public safety policy change should in theory create a greater sense of safety among community residents, based on their particular experiences of safety.

Given that the experience of safety is likely to vary across diverse, overlapping, and intersecting communities, the use of community-based metrics for policy evaluation should be approached with at least two caveats. First, the utility of such metrics is also one of its limitations: that they reflect a very local and specific measure of how the participants from a given community experience the concepts of interest, like “safety” or “well-being.” While this is important for all of the reasons we have already discussed, this aspect of the resulting measures should be emphasized whenever they are employed for broader evaluation. Second, everyday metrics are not the only way to assess policy outcomes, nor should they be used alone. Rather, the daily experience of safety among community members can be one additional tool for evaluation or policy design that helps expand the scope of how we think about, measure, and benchmark policy outcomes—in tandem with more traditional metrics like crime rates and arrest rates.

Moreover, as already noted, this necessarily leaves some aspects of “community” unaddressed. While our open-ended approach allowed for a more organic definition of community, it also precluded us from including every underrepresented voice in our process; numerous impacted communities whose voices also count were left out of our inquiry. More broadly, we are still quickly confronted with questions like: how do we account for members of other, potentially overlapping communities described by geography and identity, who might also have disproportionately negative experiences of safety and policing? For instance, what about middle- or higher-income people of color who live in geographically distinct areas of the city, but are also more likely to be targeted by police? What about low-income white residents, who reside within heavily policed neighborhoods experiencing high levels of unaddressed harm? Should people who work, but do not live, in the city be included? What about those who frequently patronize businesses in the city? And how might we think about people experiencing residential instability or homelessness, whose experiences of safety might be distinct as they move through different geographic areas?

If we are able to resolve these questions, we might still need to consider how best to weigh other geographic and identity communities that are vocal and, in some cases, powerful stakeholders in the implementation of a public safety infrastructure. This could include residents of gentrifying areas, where police are often called upon to monitor physical space and reduce so-called “quality of life” crimes (Beckett and Herbert 2009; MacLeod 2011; Walby and Lippert 2012; Beck 2020); white residents in majority white areas, where racial minorities have historically experienced disproportionately high rates of police contact (Capers 2016; Feldman et al. 2019; Bass 2001; Meehan and Ponder 2002); and high-income areas that wield significant influence over local politics and policy outcomes (Hajnal and Trounstine 2014; Trounstine 2016).

These caveats aside, the EPI process offers an important, and in many ways more expansive, way of assessing policy innovation and reform. As Davis, Kingsbury, and Merry cogently argue:

Indicators set standards. The standard against which performance is to be measured is often suggested by the name of the indicator—corruption, protection of human rights, respect for the rule of law, and so on. To the extent that an indicator is used to evaluate performance against one standard rather than another, the use of that indicator embodies a theoretical claim about the appropriate standards for evaluating actors’ conduct. Indicators often have embedded within them, or are placeholders for, a much further-reaching theory—which some might call an ideology—of what a good society is, or how governance should ideally be conducted to achieve the best possible approximation of a good society or a good policy (2012, 77).

By focusing attention on the ongoing and community-level experiences of policy change, EPI indicators provide an alternative to top-down metrics of what constitutes “success.” This affords a way to prioritize the lived experiences of those most affected by policies, drawing attention to the ways that individuals within impacted communities directly perceive the benefits of policy change. In this sense, it requires policymakers to think not just about long-term, broad policy impacts but also the extent to which policies can have direct and visceral impacts on individuals’ daily lives.

If we aim to generate a set of indicators that holds a new system accountable to providing safety for all, it is critically important that we acknowledge the complexity of these incongruous experiences and power dynamics (Davis, Kingsbury, and Merry 2012). In applying a community-centered measurement process in the context of urban American police reform, it is clear that a thoughtful and nuanced approach to defining “community” is required. Most critically, the multiple and overlapping use of the concept of community demonstrates that the outcomes of the EPI process will be shaped by whether and how these varied conceptions of community can be reconciled.