Loading...
The URL can be used to link to this page
Your browser does not support the video tag.
Home
My WebLink
About
views_of_the_police_and_neighborhood_condition
Jocelyn Fontaine, Sino Esthappan, Nancy La Vigne, and Jesse Jannetta with Anamika Dwivedi, Mathew Lynch, Ellen Paddock, and Carla Vásquez-Noriega August 2019 This brief focuses on the perceptions of residents living in the six cities that participated in the National Initiative for Building Community Trust and Justice: Birmingham, Alabama; Fort Worth, Texas; Gary, Indiana; Minneapolis, Minnesota; Pittsburgh, Pennsylvania; and Stockton, California. The goal of the National Initiative was to improve relationships and increase trust between police and community residents, particularly those living in neighborhoods with the most fraught relationships with law enforcement. As part of the Urban Institute’s evaluation of the National Initiative, residents in neighborhoods with high rates of crime and concentrated disadvantage were asked about their perceptions of the police and police-community relationships and their perceptions of neighborhood conditions. This brief examines the degree to which perceptions improved over the National Initiative implementation period. Justice system contact occurs most frequently in underresourced neighborhoods—which often comprise communities of color—and particularly among young black men and boys. This is largely because of the high rates of violent crime that persist in these communities, and the heavy police presence that often accompanies them. Conventional responses to crime (i.e., arrest, prosecution, and incarceration) have driven the overrepresentation of people from low-income communities of color at every stage of the criminal justice system and fractured relationships between police and the communities they are tasked with serving. Meanwhile, although rates of violent crime have declined substantially nationwide since the 1990s, crime rates have remained relatively stable in places of JUSTICE POLICY CENTER Views of the Police and Neighborhood Conditions Evidence of Change in Six Cities Participating in the National Initiative for Building Community Trust and Justice 2 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES concentrated disadvantage and among communities of color, where police-community relationships are most fraught. In response to these realities, the National Initiative for Building Community Trust and Justice was established through a cooperative agreement from the US Department of Justice in 2014 to strengthen relationships between police and community members and to build safer communities in the US. Led by John Jay College of Criminal Justice’s National Network for Safe Communities, and in partnership with the Center for Policing Equity, Yale Law School, and the Urban Institute, the National Initiative leveraged the expertise of practitioners and researchers to design and implement a suite of evidence- supported interventions that included the following: (1) training and technical assistance for police officers on how to engage with residents in a procedurally just manner, (2) trainings that encouraged officers to understand and mitigate implicit biases, (3) the development of model police department policies and review of extant policies to identify key changes to make, and (4) reconciliation discussions, during which police officers and community members had authentic conversations aimed at acknowledging historic tensions, harms, and misconceptions and repairing relationships. Inviting and soliciting the perspectives of the people closest to violence, disorder, and neighborhood police—victims, survivors, perpetrators, and residents of neighborhoods where crime most often occurs—is a unique departure from focusing solely on official reports or surveys representative of entire cities to determine the effectiveness of public safety strategies. People with lived experiences are critical stakeholders whose overlooked views represent a missed opportunity to generate valuable community safety strategies. These voices can inform efforts to bridge gaps in trust between communities and the police and supplement official administrative reports. Studies that have solicited resident perspectives have often been limited by selection biases that overrepresent the views of older, affluent, and educated residents (see Blumberg and Luke 2007) and overlook the people for whom issues of safety, violence, and justice involvement are most salient (de Leeuw and de Heer 2002). This brief presents the results of two waves of surveys of residents in the six National Initiative pilot communities. The first survey found that residents across the six cities held negative views about their local police departments and their neighborhood conditions (La Vigne, Fontaine, and Dwivedi 2017). This is perhaps unsurprising, given that we sampled residents of the areas with the highest rates of crime and disadvantage in each city. The second wave of surveys, featured in this brief and administered after a period of sustained National Initiative implementation, yielded similarly negative perceptions of local police departments and neighborhood conditions, but those views were markedly more positive than at baseline. However, there was notable variation among respondents across the six cities. In some cities, residents’ perceptions of the police grew significantly more positive, while in others they were unchanged or grew more negative. This is likely because of the cities’ varying experiences implementing the initiative as well as differing local contexts. Finally, we also observed notable variation in perceptions among key sociodemographic groups, notably among Black respondents, whose perceptions became considerably more positive. This brief is organized into five sections. We first describe the study’s research questions and methodology, followed by a description of the survey sample. We then examine variations in EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 3 perceptions by site and key demographic groups. We conclude by interpreting the findings in the context of National Initiative activities and implementation fidelity, and discuss the implications of the findings for policy and practice. Research Questions and Methods The surveys were intended to document community members’ views of and experiences with their neighborhood conditions in the six National Initiative cities. The community survey was designed to answer the following two research questions: Was the National Initiative associated with changes in residents’ perceptions of the police and police-community interactions and relationships? Was the National Initiative associated with changes in residents’ neighborhood conditions, specifically their victimization experiences and perceptions of safety and disorder? Because the National Initiative activities were intended to improve relationships between police and communities that experience the most frequent contact with police (Jannetta et al. 2019), the survey sample comprises people living in neighborhoods with high concentrations of disadvantage and crime. Our purposive sampling methodology was intended to represent residents living in communities with high rates of crime and victimization and that have the most fraught relationships with police. We assumed that any measurable impact of National Initiative activities on resident perceptions and experiences would occur in these neighborhoods. In partnership with local organizations, we conducted in-person surveys in the six National Initiative cities. To identify and locate the intended sample, we took the following steps: We created a sample of residences in each of the six cities. We requested address-level crime data from the six National Initiative police departments and combined these data with publicly available census data on measures of poverty and disadvantage to create a composite index of concentrated crime and poverty/disadvantage for each street segment in each city. We identified the street segments with the highest indices of concentrated crime and poverty in each city (defined as those in the top 10 percent). These street segments accounted for 39 to 50 percent of all reported crimes on residential streets and made up our final sampling frame of 6,336 households across the six cities. We mailed letters to the 6,336 households in the sampling frame to inform them about their eligibility to participate in the survey, the purpose, benefits, and risks of participation, and when researchers would be in their neighborhood. We partnered with community organizations to recruit and hire local residents to administer the surveys. Residents were trained and supervised by the Urban research team. Throughout the survey effort, survey teams of approximately 8 to 12 people attempted to contact one adult occupant at each 4 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES randomly selected household. Surveys were administered in person in English or Spanish using a paper- and-pencil instrument and were designed to take approximately 15 minutes to complete. Some respondents completed the survey over the phone with a member of the Urban research team if they were unavailable when the survey team initially knocked on their door. We endeavored to survey at least 200 residents in each city, by wave. Our final sample of 1,278 was accomplished by spending approximately two weeks in each city (including at least one weekend day) in wave 1 (after knocking on 3,947 doors). Our second sample of 1,202 was accomplished over a similar time frame in wave 2 (after knocking on 4,916 doors). As shown in table 1, the baseline survey was fielded between September 2015 and January 2016 and the follow-up survey was fielded between July 2017 and October 2017. TABLE 1 Response Rate by Site Wave 1 Wave 2 Date Completed Doors knocked Response rate Date Completed Doors knocked Response rate Minneapolis 10/2015 208 712 29.2% 7/2017 192 696 27.6% Gary 9/2015 269 620 43.4% 7/2017 219 806 27.2% Pittsburgh 9/2015 209 595 35.1% 8/2017 194 890 21.8% Stockton 10/2015 195 764 25.5% 9/2017 198 935 21.2% Birmingham 11/2015 203 612 33.2% 9/2017 218 723 30.2% Fort Worth 1/2016 194 644 30.1% 10/2017 181 866 20.9% Sample Demographics Table 2 shows the sample demographics across the six cities at both waves. The same sampling frame and sampling strategy were used at both waves; accordingly, there were few significant differences between the wave 1 and wave 2 samples. Recall, our sample consisted entirely of residents living in neighborhoods characterized by high levels of disadvantage and crime. On average, our respondents were in their late 40s at the time of the wave 1 and wave 2 surveys and most respondents identified as female (59 percent). Approximately two-thirds of respondents identified as Black/African American and roughly one-tenth identified as white and Latinx/Hispanic at each wave. Overall, the sample skewed toward the lower end of the income scale. Most respondents reported a household income of $20,000 or less. Relatedly, fewer than half of all respondents reported being employed at the time of the survey, and among those respondents, roughly three-quarters reported being employed full time. At wave 2, our sample includes a significantly higher percentage of employed respondents, and wave 2 respondents’ household incomes were significantly different than those surveyed at wave 1 (differences are most pronounced among the lower levels of the household income scale). EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 5 TABLE 2 Sample Characteristics Notes: Statistically significant differences were assessed using t-tests and chi-square tests, as appropriate. * p < 0.05, ** p < 0.01. Other^ category includes respondents who selected Native American/American Indian and Native Hawaiian/Pacific Islander, other, or more than one race. The findings presented in this brief focus on survey items relating to residents’ perceptions of and experiences with the police and police-community relationships, as well as items relating to residents’ perceptions of their neighborhood conditions. The survey items in each category were generated from previous studies directed by the research team and drawn from the literature. We first use the aggregated sample, combining results across the six sites to describe residents’ views and perceptions on dozens of items related to the local police and their neighborhood conditions at wave 1 and wave 2. We then focus on 14 scaled domains and the degree to which they differ from wave 1 to wave 2 in each city. Finally, we detail perceptions of key demographic groups on the scaled domains at wave 1 and wave 2. Perceptions of the Police and Police-Community Relationships The primary goal of the National Initiative was to improve relations between police and community residents. The survey included several items focused on the degree to which residents believed the police department used practices associated with procedural justice, considered the department to be legitimate, viewed police officers as engaging in biased activities, and perceived officers as accountable and responsive to community members and their needs. These items were included in the community Wave 1 n=1,278) Wave 2 n=1,202) Average age (years) 47 48 Age group (percent) 30 or younger 22.2 20.0 31 or older 77.8 80.0 Race and ethnicity (percent) Asian American 2.0 4.0 Black/African American 66.0 64.1 Latinx/Hispanic 10.2 11.1 White 11.9 10.8 Other^ 9.9 10.1 Gender (percent) Male 41.5 41.5 Female 58.6 58.5 Employed (percent)** 38.4 43.9 Employed full time 72.1 75.5 Annual household income (percent)* 10,000 or below 35.8 40.1 10,001–20,000 26.2 19.8 20,001–30,000 15.6 15.3 30,001–40,000 10.9 11.7 40,001–50,000 4.1 5.6 50,001 or above 7.5 7.6 6 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES survey specifically because a core component of the National Initiative was to strengthen the police departments’ relationships with their communities through policy change, community engagement, and officer training on implicit bias and procedural justice (see Jannetta et al. 2019). The survey also included several items designed to gauge residents’ agreement with and support of the rule of law. These items were included because a department’s effectiveness in neighborhood crime prevention is related not only to its perceived legitimacy and strong community relationships, but also to residents’ belief in the rule of law and whether residents are cynical about the justice system, its rules, and its effectiveness. Several survey items were also posed to gauge residents’ willingness to partner with police on neighborhood crime-control efforts. The following seven valid and reliable domains were used to assess residents’ perceptions of the police and police-community relationships: procedural justice police department legitimacy police bias community-focused policing perceptions of the law relatability of police willingness to partner with the police Procedural Justice, Police Legitimacy, and Police Bias Overall, at the time of the wave 1 survey (2015), sampled residents in the six cities reported that they did not perceive their local law enforcement officials to be engaging in procedurally just behaviors figure 1) and that those officials lacked legitimacy (figure 2). Approximately half of the sampled community residents perceived their police department as behaving in a biased manner (figure 3) at wave 1. By the time of the wave 2 survey (fielded in 2017), residents’ perceptions of nearly all items were significantly more favorable. For example, on a 5-point scale that ranged from “almost never” (1) to almost always” (5), 30 percent of residents agreed (selecting “4” or “5”) with the statement, “police treat people with dignity and respect” at wave 1. By wave 2, 38 percent of sampled residents selected 4” or “5” on the same 5-point scale (figure 1). Similarly, on a 5-point scale that ranged from “strongly disagree” (1) to “strongly agree” (5), 43 percent of residents agreed (selecting “4” or “5”) with the statement, “police in your community are legitimate authorities” at wave 1, and by wave 2, the share of residents who agreed with that statement was 51 percent. Notably, in 2015, 28 percent of sampled residents agreed with the statement , “when police deal with people, they almost always behave according to the law,” and by 2017, 38 percent agreed with this statement (figure 2). EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 7 As shown in figure 3, although perceptions of officer bias were lower at wave 2 than at wave 1, near majorities agreed with statements that officers exercise bias in their interactions with community members. Although residents expressed more favorable perceptions of police officers’ engagement in procedural justice, legitimacy, and bias at the time of the wave 2 surveys, there is considerable room for improvement, particularly with regard to perceptions of police bias. FIGURE 1 Perceptions of Procedural Justice Percentage of residents who agreed or strongly agreed that police… Notes: Response options range from 1 (almost never) to 5 (almost always). Data represent share of respondents who selected 4 and 5. Valid N = 2,326. Cronbach’s alpha = 0.9. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. 41% 39% 39% 38% 37% 37% 35% 31% 34% 34% 33% 30% 30% 30% 29% 26% Try to help people that they deal with** Try to do what is best for the people they are dealing with** Explain their decisions and actions in ways that people can understand** Treat people with dignity and respect** Give people a chance to tell their side of the story before they decide what to do** Respect people's rights** Make decisions based on the law and not their personal opinions or biases* Make fair and impartial decisions with the cases they deal with* 2015 2017 8 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES FIGURE 2 Perceptions of Police Legitimacy Percentage of residents who agreed or strongly agreed that… Notes: Response options ranged from 1 (strongly disagree) to 5 (strongly agree). Data represent percentage of respondents who selected 4 and 5. Valid N = 2,340. Cronbach’s alpha = 0.946. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. One survey item in this scale, “the police often arrest people for no good reason,” was reverse-coded. 51% 48% 44% 43% 42% 42% 41% 41% 38% 38% 37% 36% 43% 41% 38% 36% 35% 33% 39% 33% 32% 28% 32% 31% Police in your community are legitimate authorities** You and the police want the same things for your community** Police sincerely try to help people like yourself* Police generally have the same sense of right and wrong that you do** The laws that the police enforce represent the moral values of people like yourself** You generally support how the police act in your community** When the police arrest a person, there is good reason to believe that person has done something wrong Values of most of the police officers who work in your community are similar to your own** Police stand up for values that are important to you** When police deal with people, they almost always behave according to the law** The police often arrest people for no good reason The police usually act in ways consistent with your own ideas about what is right and wrong** 2015 2017 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 9 FIGURE 3 Perceptions of Police Bias Percentage of residents who agreed or strongly agreed that… Notes: Response options ranged from 1 (strongly disagree) to 5 (strongly agree). Data represent percentage of respondents who selected 4 and 5. Valid N = 2,376. Cronbach’s alpha = 0.961. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. Community-Focused Policing On a 5-point scale that ranged from “strongly disagree” (1) to “strongly agree” (5), approximately one- quarter of sampled residents selected “4” or “5” when considering whether the police department held officers accountable, whether the police department prioritized problems important to the community, or whether the police department was responsive to community concerns at the time of the wave 1 survey (figure 4). Although a significantly higher percentage of sampled residents agreed or strongly agreed with these statements at wave 2, the level of agreement for these items was still relatively low. 49% 49% 47% 47% 46% 43% 56% 54% 50% 51% 49% 47% Police officers will treat you differently because of your race/ethnicity** Police officers will judge you based on your race/ethnicity** Something you do might be misinterpreted as criminal by the police due to your race/ethnicity The police act based on personal prejudices or biases** Something you say might be misinterpreted as criminal by the police due to your race/ethnicity* The police suspect you of being a criminal because of your race/ethnicity 2015 2017 10 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES FIGURE 4 Perceptions of Community-Focused Policing Percentage of residents who agreed or strongly agreed that… Notes: Response options ranged from 1 (strongly disagree) to 5 (strongly agree). Data represent percentage of respondents who selected 4 and 5. Valid N = 2,376. Cronbach’s alpha = 0.892. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. Perceptions of the Law Overall, as shown in figure 5, a majority of respondents expressed support for the law itself, despite their less than favorable views of police legitimacy at the time of both the wave 1 and wave 2 surveys. Although there was sizable variation depending on the survey item, a near or overwhelming majority of respondents expressed support for the law. Notably, on a 5-point scale, nearly three-quarters of sampled residents agreed (“4”) or strongly agreed (“5”) that “all laws should be strictly obeyed,” that obeying the law ultimately benefits everyone in the community,” and that “people should do what the law says.” Perceptions of the law were significantly better at the time of the wave 2 surveys. 35% 33% 29% 28% 28% 24% The police department is responsive to community concerns** The police department prioritizes problems most important to your community** The police department holds officers accountable for wrong or inappropriate conduct in the community* 2015 2017 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 11 FIGURE 5 Perceptions of the Law Percentage of residents who agreed or strongly agreed that… Notes: Response options ranged from 1 (strongly disagree) to 5 (strongly agree). Data represent percentage of respondents who selected 4 and 5. Valid N = 2,363. Cronbach’s alpha = 0.880. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. Relatability to Police and Willingness to Partner with Police Several survey items assessed residents’ perceptions of police-community relations, including the degree to which respondents related to the police and their willingness to partner with police on various crime-control efforts. Police “relatability” questions included items about the degree to which residents viewed police as honest, trustworthy, and part of the community. Figure 6 shows that, overall, residents did not perceive police to be very relatable. At wave 1, on a 5-point scale from “strongly disagree” (1) to strongly agree” (5), only 24 percent of those sampled agreed (“4”) or strongly agreed (“5”) that “police are honest,” and roughly one-third agreed that “police are part of [their] neighborhood,” and that “[they] feel safe” and “comfortable around the police.” Notably, 43 percent of sampled residents said they could imagine being friends with a police officer at wave 1, which increased to a near majority at wave 2. Overall, perceptions of police relatability were uniformly higher at wave 2, though there is clear room for improvement in this domain. 79% 78% 76% 63% 60% 58% 58% 74% 74% 73% 56% 54% 49% 52% All laws should be strictly obeyed** People should do what the law says** Obeying the law ultimately benefits everyone in the community** The laws in your community are consistent with your own intuitions about what is right and wrong** A person who disobeys laws is a danger to others in the community** The laws of our system are generally consistent with the views of the people in your community about what is right and wrong** It is hard to break the law and keep your self- respect** 2015 2017 12 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES As shown in figure 7, there was wide variation in residents’ willingness to partner with police at the time of the wave 1 and wave 2 surveys. On a 5-point scale ranging from “very unlikely” (1) to “very likely” (5), more residents suggested they were likely to call the police to report a crime (71 percent) than were likely to patrol the street as part of an organized group (41 percent) or to volunteer their time to help police (47 percent) at the time of the wave 1 survey. Although more sampled residents reported being likely to partner on crime-control efforts at wave 2 than wave 1, the differences were not very pronounced and were only significantly better at wave 2 on two questions (providing information to help police find a subject and calling the police to report a crime). FIGURE 6 Perceptions of Police Relatability Percentage of residents who agreed or strongly agreed that… Notes: Response options ranged from 1 (strongly disagree) to 5 (strongly agree). Data represent percentage of respondents who selected 4 and 5. Valid N = 2,360. Cronbach’s alpha = 0.870. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. 48% 44% 43% 41% 35% 27% 43% 36% 38% 36% 30% 24% You can imagine being friends with a police officer* You feel comfortable around the police** You feel safe around the police** The police are a part of your neighborhood You personally trust the police** The police are honest** 2015 2017 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 13 FIGURE 7 Willingness to Partner with the Police Percentage of residents who reported being likely or very likely to… Notes: Response options ranged from 1 (very unlikely) to 5 (very likely). Data represent percentage of respondents who selected 4 and 5. Valid N = 2,439. Cronbach’s alpha = 0.888. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. Perceptions of Neighborhood Conditions As a complement to survey items about their perceptions of police and police-community relationships, we asked residents about the condition of their neighborhoods, their experiences with victimization, and their perceptions of safety. A department’s ability to improve relationships and increase trust among residents may result in better neighborhood conditions and public safety. Conversely, departmental efforts to improve neighborhood conditions could improve residents’ perceptions of police. The following seven valid and reliable domains were included to assess residents’ perceptions of their neighborhood conditions: neighborhood safety neighborhood disorder neighborhood crime personal victimization experiences vicarious victimization experiences concerns about victimization precautionary behaviors 75% 73% 69% 54% 49% 40% 71% 69% 64% 53% 47% 41% Call the police to report a crime* Report suspicious activity near your home to the police Provide information to the police to help find a suspect* Attend a community meeting with the police to discuss crime in your neighborhood Volunteer your time to help the police solve a crime or find a suspect Patrol the streets as part of an organized community group 2015 2017 14 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES Neighborhood Safety, Disorder, and Crime Overall, respondents perceived their neighborhoods as moderately safe places (figure 8) with some signs of physical and social disorder (figure 9) where violent crime is not uncommon (figure 10). On a 5- point scale ranging from “dangerous” (1) to “safe” (5), 39 percent of respondents perceived their neighborhood to be relatively safe (“4” or “5”) and roughly 50 percent perceived their block to be relatively safe. A near majority or majority of respondents felt their neighborhood landmarks were relatively safe at the time of the wave 2 survey, which was significantly higher than at the time of the wave 1 survey. Residents responded to several survey items about neighborhood problems commonly associated with physical and social disorder. At wave 1, on a 5-point scale ranging from “not a problem” (1) to “a big problem” (5), the share of respondents reporting that various aspects of their neighborhood’s physical and social conditions were a problem (“4” or “5”) ranged from 30 to 48 percent (figure 9). At the time of the wave 2 survey, a significantly smaller percentage of respondents perceived their neighborhood’s physical and social conditions to be a problem. Notably, 48 percent of sampled residents perceived people using or selling drugs” to be a problem in 2015, compared with 37 percent of sampled residents in 2017. We also asked residents how often crime occurs in their neighborhoods. With the exception of shooting and shooting attempts, a relatively small percentage of respondents reported that crime occurred at least weekly in their neighborhoods at the time of both the wave 1 and wave 2 surveys figure 10). A significantly smaller share of respondents reported that crime occurred at least weekly at the time of the wave 2 surveys as compared with the wave 1 surveys. Some types of crime were reported to occur more frequently than others. For example, rapes and sexual assaults were reported as occurring less frequently than other, more publicly visible forms of crime such as shootings or shooting attempts or assaults. Significantly, there was a 10-point drop in the share of residents reporting shootings and shooting attempts—the most violent crime type—as occurring at least weekly (35 percent of respondents at wave 1 and 25 percent of respondents at wave 2). EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 15 FIGURE 8 Perceptions of Neighborhood Safety Percentage of respondents who rated the following landmarks as relatively safe Notes: Response options ranged from 1 (dangerous) to 5 (safe). Data represent percentage of respondents who selected 4 and 5. Valid N = 2,359. Cronbach’s alpha = 0.918. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.0,5 **p < 0.01. FIGURE 9 Perceptions of Neighborhood Disorder Percentage of respondents who reported the following to be a problem in their neighborhood Notes: Response options ranged from 1 (not a problem) to 5 (a big problem). Data represent percentage of respondents who selected 4 and 5. Valid N = 2,403. Cronbach’s alpha = 0.881. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. 37% 40% 33% 29% 26% 24% 48% 46% 44% 40% 34% 30% People selling or using drugs** Litter, broken glass, or trash on the sidewalks and streets** Vacant or deserted houses, buildings, or storefronts** Teenagers hanging around on the streets** People being drunk or rowdy in public places** Vandalism or graffiti on buildings or cars** 2015 2017 59% 54% 52% 49% 48% 50% 50% 48% 45% 42% 41% 39% Your block** Your closest local public transportation station or stop** The way to and from the local public transportation station** Neighborhood stores** Neighborhood parks** Your neighborhood** 2015 2017 16 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES FIGURE 10 Perceptions of the Frequency of Neighborhood Violence Percentage of respondents who indicated that the following problems occurred weekly or daily Notes: Response options ranged from 1 (never) to 5 (daily). Data represent percentage of respondents who selected 4 (weekly) and 5 (daily). Valid N = 2,409. Cronbach’s alpha = 0.931. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. Personal and Vicarious Victimization Experiences To complement the survey items about residents’ perceptions of the safety of their neighborhoods, we asked residents about their personal and victimization experiences with property and violent crime. In particular, we asked residents whether they knew someone in their neighborhood who had experienced crime in the past year, as well as whether they had been personally victimized in the past year. As shown in figure 11, a sizable share of sampled residents knew someone who had been victimized in the past year at the time of the wave 1 and wave 2 surveys. Notably, 44 percent of sampled residents knew a victim of a recent shooting or shooting attempt at wave 1. Similar shares of neighborhood residents knew victims of violent fights, assaults, and home burglaries at wave 1. Of the sampled residents at wave 1, 16 percent had recently been victimized by at least one of the eight crime types we asked about, which was not significantly different from the 14 percent who reported being personally victimized at wave 2. Across the eight crime types, a significantly smaller percentage of sampled residents reported knowing someone who had been victimized at wave 2 as compared with wave 1. Nonetheless, 31 percent of wave 2 respondents knew someone who had been shot or shot at (a 13 percent reduction from the 44 percent of wave 1 respondents). 25% 12% 11% 10% 9% 10% 9% 5% 35% 19% 15% 15% 15% 14% 12% 6% Shooting or shooting attempts** Assaults** Violent fights between neighbors** Muggings or robberies** Home burglaries** Car thefts** Stabbings or stabbing attempts** Rapes or sexual assaults* 2015 2017 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 17 FIGURE 11 Personal and Vicarious Victimization Experiences Percentage of respondents who knew someone who experienced the following in the prior year Notes: Response options included yes or no. Data represent the percentage of respondents who selected yes. Valid N = 2,430. Cronbach’s alpha = 0.880. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. Concerns about Victimization and Precautionary Behavior To further understand residents’ perceptions of safety in their neighborhoods, we asked them to indicate their concerns about being victimized and their engagement in precautionary behaviors out of fear for their own safety. Using a 5-point scale ranging from “not concerned” (1) to “very concerned” (5), we found that nearly 4 in 10 residents were concerned (“4” or “5”) about being shot or shot at, being mugged or robbed, and having their homes being broken into at wave 1 (figure 12). At wave 2, a smaller percentage of respondents were concerned about being violently victimized or having their property broken into or stolen. Consistent with their reported concerns about personal victimization, a sizable percentage of respondents reported they were likely to engage in precautionary behaviors. Unlike the responses in other domains, respondents’ answers were remarkably similar at wave 1 and wave 2 on these items. On a 5-point scale ranging from “very unlikely” (1) to “very likely” (5), approximately one- half of respondents reported being likely (“4” or “5”) to stay in at night at both wave 1 and wave 2 and one-quarter reported being likely to carry a weapon at wave 1 and wave 2. Although residents’ responses were fairly similar at both waves, a significantly smaller percentage of wave 2 respondents reported being likely or very likely to avoid certain streets and buildings, not travel alone, and stay in at night as compared with wave 1 respondents. 14% 31% 25% 21% 22% 23% 20% 14% 8% 16% 44% 35% 34% 30% 29% 29% 23% 11% In the past year, have you been a victim of any of the above crimes in your neighborhood Shooting or shooting attempts** Home burglaries** Assaults** Violent fights between neighbors** Car thefts** Muggings or robberies** Stabbings or stabbing attempts** Rapes or sexual assaults** 2015 2017 18 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES FIGURE 12 Concerns about Various Violent and Property Crimes Percentage of respondents who reported being concerned about… Notes: Responses ranged from 1 (not concerned) to 5 (very concerned). Data represent the percentage of respondents who selected 4 and 5. Valid N = 2,389. Cronbach’s alpha = 0.954. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05, **p < 0.01. FIGURE 13 Precautionary Behaviors Percentage of respondents who were likely to do the following out of concern for their own safety Notes: Responses ranged from 1 (very unlikely) to 5 (very likely). Data presented are percentage likely and very likely. Valid N = 2,426. Cronbach’s alpha = 0.763. Statistically significant differences represent differences in the total mean score and were assessed using t-tests; *p < 0.05 **p < 0.01. 33% 32% 32% 27% 26% 23% 21% 41% 39% 39% 35% 33% 29% 28% Having your home broken into** Being mugged or robbed** Being shot or shot at** Having your car stolen** Being physically assaulted** Being stabbed** Being sexually assaulted** 2015 2017 48% 37% 37% 35% 25% 51% 40% 42% 39% 25% 0%20%40%60% Stay in at night* Not travel alone** Avoid certain buildings* Avoid certain streets Carry a weapon 2015 2017 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 19 In summary, perceptions of the police and police-community relationships and perceptions of neighborhood conditions were more favorable at the time of the wave 2 survey (2017) than they were at wave 1 (2015). To better assess differences in responses between waves, we created 14 scaled domains. These 14 domains were created following a factor analysis that assessed the validity and reliability of the concepts (domains were created by averaging the ratings of the individual questions). Table 3 shows the average ratings and percentages among these domains at wave 1 and wave 2. Based on t-tests of the difference in means, with one exception, perceptions are significantly more favorable at wave 2. These findings are notable given the few significant sociodemographic differences between the wave 1 and wave 2 samples, as shown in table 2. TABLE 3 Scale Differences by Wave Perceptions of the police and neighborhood conditions Wave 1 n=1,278) Wave 2 n=1,202) Perceptions of police and police-community relationships Procedural justice 2.88 3.04** Police legitimacy 2.83 3.01** Police bias 3.35 3.20** Community-focused policing 2.61 2.80** The law 3.80 4.00** Relatability to the police 2.83 3.01** Willingness to partner with police 3.54 3.61 Perceptions of neighborhood conditions Neighborhood safety 3.15 3.41** Neighborhood disorder 2.94 2.62** Frequency of neighborhood crime 2.10 1.87** Personal victimization experience~ 0.16 0.14** Vicarious victimization experience~ 0.29 0.20** Concerns about various property and violent crimes 2.60 2.36** Precautionary behavior 2.82 2.70* Note: Statistically significant differences were assessed using t-tests; *p < 0.05 **p < 0.01; this measure is the percentage of respondents reporting “yes.” Variation within and across Cities As Jannetta and colleagues (2019) detail, the six National initiative cities varied by geographic region, population and police department size, historical crime trends, and sociodemographic characteristics. Each city also varied in its approach to and execution of the National Initiative activities and fidelity to the model. As such, our analyses also focus on whether residents’ perceptions varied by city. Table 4 demonstrates that the aggregated findings mask significant variation in perceptions by city at wave 2 compared with wave 1. Residents in Stockton and Minneapolis had more favorable perceptions of the police and police-community relationships at wave 2 relative to wave 1, while Fort 20 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES Worth residents’ perceptions of police and police-community relationships were not significantly different at wave 2. Gary and Birmingham residents’ perceptions of police and police-community relationships were more favorable on some domains at wave 2 than wave 1. Meanwhile, residents in Pittsburgh sampled at wave 2 had less favorable perceptions of police than those sampled at wave 1. On the domains assessing perceptions of neighborhood conditions, Fort Worth residents sampled at wave 2 had more favorable perceptions on all of the survey domains than those sampled at wave 1. Residents in Stockton and Pittsburgh perceived better neighborhood conditions at wave 2 on several domains, while Gary residents sampled at wave 2 had no significantly different perceptions of neighborhood conditions than those sampled at wave 1. Birmingham and Minneapolis residents perceived a handful of the neighborhood conditions to be better at wave 2 than those sampled at wave 1. To determine whether differences in the sociodemographic characteristics of the respondents might account for differences in perceptions at the two points in time, we analyzed the variation in the sociodemographic characteristics by city. We found a handful of significant differences (shown in the appendix). We found that the wave 2 samples differed from wave 1 in the following ways: In Birmingham, there were significantly more employed respondents at wave 2 (p < 0.01). In Fort Worth, the wave 2 sample included a greater proportion of residents in the middle- income categories (p < 0.05). In Gary, there were significantly more older respondents at wave 2 (p < 0.05). In Minneapolis, the wave 2 sample was not significantly different. In Pittsburgh, the wave 2 sample included a greater proportion of residents in the lower-income categories (p < 0.01). In Stockton, in wave 2, the racial/ethnic breakdown of respondents was significantly different (p 0.05), with a notably higher proportion of Asian respondents and significantly more employed respondents (p < 0.01). Although these differences are worth noting and provide some context for the observed differences in the measured domains between wave 1 and wave 2, we feel confident the differences in the measured domains are not chiefly because of differences in the samples’ sociodemographic characteristics. EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 21 TABLE 4 Scale Differences across Sites Views on police and victimization Stockton Fort Worth Pittsburgh Birmingham Gary Minneapolis W1 W2 W1 W2 W1 W2 W1 W2 W1 W2 W1 W2 Perceptions of police/police- community relationships Procedural justice 2.8 3.1** 3.5 3.5 2.7 2.6 3.0 3.1 2.8 3.1** 2.5 2.9** Police legitimacy 2.7 3.0** 3.4 3.3 2.8 2.6* 3.0 3.1 2.7 3.1** 2.5 2.8* Police bias 3.5 3.2* 2.9 3.1 3.6 3.5 3.1 3.2 3.3 2.9 3.6 3.2 Community-focused policing 2.4 2.7* 3.1 3.2 2.5 2.3 2.8 2.9 2.6 3.1** 2.3 2.5 The law 3.9 4.0 4.0 4.2 3.9 3.7* 3.8 4.3** 3.8 4.0** 3.4 3.7** Relatability to police 2.5 3.1** 3.3 3.4 2.9 2.6* 2.9 3.0 2.8 3.0 2.6 3.0** Willingness to partner with police 3.2 3.5* 3.8 4.0 3.8 3.3** 3.5 3.9** 3.7 3.5 3.2 3.5* Perceptions of neighborhood conditions Neighborhood safety 2.9 3.4** 2.9 3.3** 3.3 3.6** 3.2 3.3 3.6 3.4 3.0 3.4** Neighborhood disorder 2.9 2.5** 3.2 2.6** 3.0 2.6** 3.0 2.8 2.6 2.5 2.8 2.7 Frequency of neighborhood violence 2.1 1.8** 2.1 1.8** 2.0 1.9 2.4 2.2* 1.8 1.7 2.2 2.0* Personal victimization experiences* 0.20 0.14 0.17 0.13 0.11 0.13 0.14 0.17 0.11 0.06* 0.22 0.21 Vicarious victimization experiences* 0.33 0.20** 0.20 0.15 0.25 0.20 0.38 0.28** 0.26 0.16** 0.33 0.22** Concerns about victimization 2.6 2.4 2.9 2.5** 2.4 2.0** 3.1 2.8 2.3 2.4 2.2 2.0 Precautionary behavior 2.8 2.8 3.3 3.0** 2.6 2.3** 3.1 3.0 2.5 2.5 2.5 2.5 Notes: Statistically significant differences were assessed using t-tests; *p < 0.05 **p < 0.01.; this measure is the percentage of respondents reporting “yes.” Results highlighted blue indicate significant increases and results highlighted yellow indicate significant decreases. 22 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES Variation by Key Subgroups The National Initiative was designed to improve relationships and increase trust between the justice system and communities, particularly communities of color. The history of police brutality toward African Americans was an important impetus for the conceptualization of the National Initiative and its key components reconciliation, training, and policy change). Black residents, and men and boys in particular, have particularly negative perceptions of police (Nadal et al. 2017; Peck 2015; Schafer, Huebner, and Bynum 2003). To determine whether the National Initiative implementation period was associated with differences in perceptions among key demographic groups, we assessed the degree to which differences in perceptions were evident by race/ethnicity and by age. We focused these analyses on the entire sample, because city- specific analyses do not have sufficient samples for subgroup analyses. Racial/ethnic differences at wave 1 and wave 2 are shown in table 5, focusing on Black/African American respondents, Latinx respondents, and those falling in the “other” category. There were many significant differences by race/ethnicity on perceptions of police and police-community relationships and some differences, though fewer, on perceptions of neighborhood conditions. At wave 1, Black respondents’ perceptions of police were significantly worse than those of Latinx and other racial/ethnic respondents. By wave 2, these differences persisted. There were fewer racial/ethnic differences in perceptions of neighborhood conditions. At wave 1 and wave 2, Black respondents perceived more neighborhood disorder than Latinx and other respondents. Yet, a smaller percentage of Black respondents reported being personally victimized at wave 1 than Latinx and other respondents and at wave 2, more Latinx respondents reported being personally victimized than Black and other respondents. Black respondents reported engaging in more precautionary behaviors at wave 1 than Latinx and other respondents, and this difference was not evident in wave 2. Overall, these results demonstrate some improvements in Black respondents’ perceptions of neighborhood conditions compared with Latinx and other respondents. Table 6 and table 7 show differences in perceptions of Black and Latinx respondents, respectively, at wave 1 relative to wave 2. With three exceptions, Black respondents’ perceptions of police and police- community relationships and neighborhood conditions were significantly more favorable at wave 2 than at wave 1. The size of the percentage differences is notable. There were fewer differences in perceptions among Latinx respondents at wave 2 relative to wave 1; however, those differences followed the expected direction over time (that is, wave 2 Latinx respondents had better perceptions of some aspects of police and police-community relationships and the neighborhood conditions than wave 1 Latinx respondents). As shown in table 8, at both wave 1 and wave 2, older respondents (ages 31 and older) had better perceptions of the police and police-community relationships than younger respondents (ages 30 and younger). The same pattern exists when focusing on perceptions of neighborhood conditions, though there were fewer significant differences in residents’ perceptions by age. Notably, at wave 2, there were fewer significant differences in perceptions of neighborhood conditions among those ages 31 and older and those ages 30 and younger. This may be an indication that efforts to increase neighborhood conditions and safety over time were particularly felt or evidenced among young adults. EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 23 TABLE 5 Scale Differences by Race/Ethnicity Notes: Statistically significant differences among racial and ethnic groups were assessed using analysis of variance; *p < 0.05 **p < 0.01. TABLE 6 Scale Differences among Black/African-American Respondents by Wave Views on police and neighborhood conditions Wave 1 n=843) Wave 2 n=770) Perceptions of police and police-community relationships Procedural justice 2.80 2.91 Police legitimacy 2.81 2.92* Police bias 3.55 3.35** Community-focused policing 2.51 2.71** The law 3.80 3.98** Relatability to the police 2.72 2.85* Willingness to partner with police 3.56 3.53 Perceptions of neighborhood Conditions Neighborhood safety 3.19 3.42** Neighborhood disorder 3.02 2.66** Frequency of neighborhood violence 2.12 1.89** Personal victimization experience~ 0.13 0.12 Vicarious victimization experience~ 0.28 0.21** Concerns about various property and violent crimes 2.66 2.39** Precautionary behavior 2.91 2.74** Notes: Statistically significant differences were assessed using t-tests; *p < 0.05 **p < 0.01; ~ percentage of respondents reporting yes.” Results highlighted blue indicate significant increases and results highlighted yellow indicate significant decreases. Wave 1 Wave 2 Black n=843) Latinx n=130) Other n=305) Black n=770) Latinx n=133) Other n=299) Perceptions of police and police- community relationships Procedural justice 2.80 3.10 3.01** 2.91 3.49 3.20** Police legitimacy 2.81 3.17 2.76** 2.92 3.41 3.05** Police bias 3.55 3.12 2.84** 3.35 2.97 2.88** Community-focused policing 2.51 2.97 2.75** 2.71 3.13 2.86** The law 3.80 4.01 3.71* 3.98 4.24 3.93** Relatability to the police 2.72 3.08 3.06** 2.85 3.35 3.29** Willingness to partner with police 3.56 3.45 3.56 3.53 3.84 3.71** Perceptions of neighborhood conditions Neighborhood safety 3.19 3.20 3.03 3.42 3.48 3.34 Neighborhood disorder 3.02 2.76 2.80** 2.66 2.61 2.51 Frequency of neighborhood violence 2.12 2.07 2.09 1.89 1.78 1.86 Personal victimization experiences 0.13 0.18 0.22** 0.12 0.21 0.15* Vicarious victimization experiences 0.28 0.28 0.31 0.21 0.20 0.18 Concerns about various violent and property crimes 2.66 2.55 2.45 2.39 2.58 2.21* Precautionary behavior 2.91 2.64 2.66** 2.74 2.83 2.54* 24 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES TABLE 7 Scale Differences among Latinx Respondents by Wave Views on police and neighborhood conditions Wave 1 n=130) Wave 2 n=133) Perceptions of police and police-community relationships Procedural justice 3.10 3.49** Police legitimacy 3.17 3.41 Police bias 3.12 2.97 Community-focused policing 2.97 3.13 The law 4.01 4.24* Relatability to the police 3.08 3.35 Willingness to partner with police 3.45 3.84* Perceptions of neighborhood conditions Neighborhood safety 3.20 3.48* Neighborhood disorder 2.76 2.61 Frequency of neighborhood violence 2.07 1.78* Personal victimization experience~ 0.18 0.21 Vicarious victimization experience~ 0.28 0.20* Concerns about various property and violent crimes 2.55 2.58 Precautionary behavior 2.64 2.83 Notes: Statistically significant differences were assessed using t-tests; *p < 0.05 **p < 0.01; ~ percentage of respondents reporting yes”. Results highlighted blue indicate significant increases and results highlighted yellow indicate significant decreases. TABLE 8 Scale Differences by Age Notes: Statistically significant differences were assessed using t-tests; *p < 0.05 **p < 0.01. Results highlighted blue indicate significantly higher mean scores among respondents age 31 or older compared with those 30 or younger, and results highlighted yellow indicate significantly lower mean scores among respondents age 31 or older compared with those 30 or younger. Wave 1 Wave 2 30 or younger n=284) 31 or older n=994) 30 or younger n=240) 31 or older n=962) Perceptions of police and police-community relationships Procedural justice 2.62 2.95** 2.89 3.08* Police legitimacy 2.66 2.88** 2.85 3.05* Police bias 3.39 3.34 3.29 3.17 Community-focused policing 2.34 2.69** 2.58 2.85** The law 3.55 3.87** 3.73 4.07** Relatability to police 2.38 2.96** 2.68 3.09** Willingness to partner with police 3.01 3.70** 3.06 3.75** Perceptions of neighborhood conditions Neighborhood safety 3.05 3.18 3.35 3.42 Neighborhood disorder 3.12 2.89** 2.86 2.56** Frequency of neighborhood violence 2.33 2.04** 2.05 1.83** Personal victimization experiences 0.20 0.15* 0.16 0.14 Vicarious victimization experiences 0.35 0.27** 0.23 0.20 Concerns about various violent and property crimes 2.61 2.60 2.50 2.33 Precautionary behavior 2.78 2.83 2.71 2.70 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 25 Limitations As with any research undertaking, this study had a few methodological limitations. First, because we did not set out to interview the same people from wave 1 to wave 2, the second sample of respondents is not necessarily representative of the initial sample. In addition, our sampling methodology was designed to focus on specific types of communities, and the results represent the views of people who we were able to contact and who agreed to participate; findings should not be viewed as representative of all adults living in the sampled neighborhoods. Moreover, the sample sizes from both waves, although sufficient for comparison purposes, were small relative to the number of people residing in the sampled street segments. Nevertheless, the information collected through this self-reported survey provide a critical complement to the information collected through official police data. Conclusion The evaluation finds that perceptions of police-community relationships and neighborhood conditions among a sample of residents living in the National Initiative cities were more positive after a period of sustained implementation (from 2015 to 2017). As expected, residents’ perceptions of policing and police- community relationships were negative overall when the National Initiative launched. Many of these perceptions of police and police-community relationships were on the neutral or positive end of the scale by the time of the second survey. Although there is considerable room for improvement, the National Initiative activities were associated with changes in residents’ perceptions in the intended domains (most notably procedural justice, police legitimacy, and police bias). Yet, the aggregated findings mask significant variation among the six National Initiative cities. In particular, residents’ perceptions and experiences of their local police department, police-community relationships, and their neighborhood conditions improved considerably in Stockton and Minneapolis. In Fort Worth, though residents’ perceptions of their neighborhood conditions improved, there was no measured change in residents’ perceptions of the police or police-community relationships. In Gary and Birmingham, residents perceived some improvement in their neighborhood conditions and the police and police-community relationships. Finally, in Pittsburgh, though residents perceived some improvement in their neighborhood conditions, their perceptions of the police and police-community relationships grew more negative. As discussed in more detail in the companion implementation report, the variation in residents’ perceptions by city tracks broadly with variation in implementation fidelity and dosage. The survey also demonstrated important differences in perceptions among key demographic groups, albeit in mostly expected directions at the timing of the first and second surveys (younger residents had more negative perceptions and experiences than older residents and Black residents had more negative perceptions and experiences than Latinx and other residents). Notably, differences in perceptions of and experiences with neighborhood conditions among Blacks and young adults were more consistent with those of whites and older adults, respectively, at the time of the second wave. This provides some indication that the National Initiative activities may have led to a greater feeling of safety in certain communities over time, as it was designed to do. Critically, and consistent with the logic of the National 26 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES Initiative, Black residents’ perceptions of police, police-community relationships, and neighborhood conditions were notably better at the time of the second survey wave. Latinx residents’ perceptions also somewhat improved. This also provides support for the initiative’s design and implementation approaches. Though the findings show promise and empirical support for the National Initiative effort overall, it is inappropriate to suggest that activities associated with the National Initiative were the sole cause of the measured improvements in residents’ perceptions of police, police-community relationships, or neighborhood conditions. As illustrated in the accompanying implementation and impact evaluation report, the surveys were administered over a two-year period when changes in community conditions and activities (e.g., economic conditions), as well as incidents within the departments and American policing more broadly (e.g., changes in police leadership, critical incidents in the department, national conversations around policing and police accountability), could have influenced both the implementation of the National Initiative and residents’ perceptions. Nevertheless, the results suggest key and notable improvements over a period of sustained National Initiative implementation among the residents represented in this brief. This is particularly evident in the cities that experienced success implementing the department training, community engagement activities, and policy changes the National Initiative endeavored to accomplish. Appendix A TABLE A.1 Sample Characteristics: Birmingham Notes: Statistically significant differences were assessed using t-tests and chi-square tests, as appropriate. *p < 0.05 **p < 0.01.; ^ Other category includes respondents who selected Native American/American Indian and Native Hawaiian/Pacific Islander, other, or more than one race. Wave 1 n=203) Wave 2 n=218) Average age (years) 46 46 Age group (percent) 30 or younger 25.12% 21.56% 31 or older 74.88% 78.44% Race and ethnicity (percent) Asian American 0.99% 0.92% Black/African American 80.30% 87.61% Latinx/Hispanic 0.99% 0.92% White 6.40% 5.50% Other^ 11.33% 5.05% Gender (percent) Male 42.08% 43.44% Female 57.92% 56.67% Employed (percent)** 29.12% 42.72% Employed full time 84.91% 70.79% Annual household income (percent) 10,000 or lower 44.44% 38.81% 10,001–20,000 20.00% 19.90% 20,001–30,000 16.11% 20.90% 30,001–40,000 9.44% 8.46% 40,001–50,000 2.78% 5.47% 50,001 or above 7.22% 6.47% EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 27 TABLE A.2 Sample Characteristics: Fort Worth Wave 1 n=194) Wave 2 n=181) Average age (years) 47 47 Age group (percent) 30 or younger 23.71% 20.44% 31 or older 76.29% 79.56% Race and ethnicity (percent) Asian American 1.03% 1.10% Black/African American 53.61% 45.86% Latinx/Hispanic 24.74% 25.97% White 11.86% 11.05% Other^ 8.76% 16.02% Gender (percent) Male 36.84% 41.71% Female 63.16% 58.29% Employed (percent) 41.36% 45.98% Employed full time 82.28% 81.48% Annual household income (percent)* 10,000 or lower 34.83% 28.95% 10,001–20,000 28.65% 23.03% 20,001–30,000 17.42% 19.08% 30,001–40,000 6.74% 14.47% 40,001–50,000 2.25% 7.89% 50,001 or above 10.11% 6.58% Notes: Statistically significant differences were assessed using t-tests and chi-square tests, as appropriate. * p < 0.05 ** p < 0.01.; ^ Other category includes respondents who selected Native American/American Indian and Native Hawaiian/Pacific Islander, other, or more than one race. 28 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES TABLE A.3 Sample Characteristics: Gary Notes: Statistically significant differences were assessed using t-tests and chi-square tests, as appropriate. * p < 0.05 ** p < 0.01.; Other^ category includes respondents who selected Native American/American Indian and Native Hawaiian/Pacific Islander, other, or more than one race. Wave 1 n=269) Wave 2 n=219) Average age (years)* 48 54 Age group (percent) 30 or younger 14.50% 13.24% 31 or older 85.50% 86.76% Race and ethnicity (percent) Asian American 0.00% 0.46% Black/African American 93.31% 89.95% Latinx/Hispanic 0.37% 2.28% White 2.97% 2.74% Other^ 3.35% 4.57% Gender (percent) Male 42.21% 39.44% Female 57.79% 60.56% Employed (percent) 35.09% 35.07% Employed full time 76.09% 81.58% Annual household income (percent) 10,000 or lower 50.53% 55.92% 10,001–20,000 26.06% 18.01% 20,001–30,000 10.11% 9.95% 30,001–40,000 7.98% 9.00% 40,001–50,000 1.60% 2.37% 50,001 or above 3.72% 4.74% EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 29 TABLE A.4 Sample Characteristics: Minneapolis Notes: Statistically significant differences were assessed using t-tests and chi-square tests, as appropriate. * p < 0.05 ** p < 0.01.; Other^ category includes respondents who selected Native American/American Indian and Native Hawaiian/Pacific Islander, other, or more than one race. Wave 1 n=208) Wave 2 n=192) Average age (years) 43 44 Age group (percent) 30 or younger 29.33% 25.00% 31 or older 70.67% 75.00% Race and ethnicity (percent) Asian American 4.81% 4.69% Black/African American 54.81% 50.52% Latinx/Hispanic 5.29% 7.81% White 19.23% 21.88% Other^ 15.87% 15.10% Gender (percent) Male 49.75% 43.92% Female 50.25% 56.08% Employed (percent) 53.47% 56.61% Employed full time 51.40% 64.42% Annual household income (percent) 10,000 or lower 31.18% 31.76% 10,001–20,000 22.58% 20.00% 20,001–30,000 13.98% 10.59% 30,001–40,000 18.28% 15.88% 40,001–50,000 5.91% 5.88% 50,001 or above 8.06% 15.88% 30 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES TABLE A.5 Sample Characteristics: Pittsburgh Notes: Statistically significant differences were assessed using t-tests and chi-square tests, as appropriate. * p < 0.05 ** p < 0.01.; Other^ category includes respondents who selected Native American/American Indian and Native Hawaiian/Pacific Islander, other, or more than one race. Wave 1 n=209) Wave 2 n=194) Average age (years) 51 51 Age group (percent) 30 or younger 15.31% 15.46% 31 or older 84.69% 84.54% Race and ethnicity (percent) Asian American 0.00% 1.03% Black/African American 73.68% 76.29% Latinx/Hispanic 0.00% 0.00% White 20.10% 13.92% Other^ 6.22% 8.76% Gender (percent) Male 30.58% 37.37% Female 69.42% 62.63% Employed (percent) 38.54% 37.10% Employed full time 76.32% 80.28% Annual household income (percent)** 10,000 or lower 31.50% 46.45% 10,001–20,000 32.00% 21.31% 20,001–30,000 12.50% 12.02% 30,001–40,000 9.00% 11.48% 40,001–50,000 4.50% 4.92% 50,001 or above 10.50% 3.83% EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 31 TABLE A.6 Sample Characteristics: Stockton Notes: Statistically significant differences were assessed using t-tests and chi-square tests, as appropriate. * p < 0.05 ** p < 0.01.; ^ Other category includes respondents who selected Native American/American Indian and Native Hawaiian/Pacific Islander, other, or more than one race. Wave 1 n=195) Wave 2 n=198) Average age (years) 43 45 Age group (percent) 30 or younger 28.21% 24.75% 31 or older 71.79% 75.25% Race and ethnicity (percent)* Asian American 6.15% 16.16% Black/African American 29.23% 27.27% Latinx/Hispanic 34.87% 32.32% White 13.33% 11.62% Other^ 16.41% 12.63% Gender (percent) Male 47.57% 43.23% Female 52.43% 56.77% Employed (percent)** 32.79% 47.12% Employed full time 72.88% 78.89% Annual household income (percent) 10,000 or lower 21.97% 33.33% 10,001–20,000 27.17% 16.97% 20,001–30,000 24.28% 20.00% 30,001–40,000 13.87% 12.73% 40,001–50,000 7.51% 7.88% 50,001 or above 5.20% 9.09% 32 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES Appendix B TABLE B.1 Mean Scores on Perceptions of Police and Police-Community Relationships Wave 1 Wave 2 N Mean SD N Mean SD Procedural justice Try to help people that they deal with** 1,226 2.97 1.35 1,168 3.15 1.34 Try to do what is best for the people they are dealing with** 1,228 2.96 1.36 1,164 3.10 1.32 Explain their decisions and actions in ways that people can understand** 1,226 2.88 1.37 1,161 3.05 1.37 Treat people with dignity and respect** 1,224 2.84 1.35 1,161 3.06 1.34 Give people a chance to tell their side of the story before they decide what to do** 1,218 2.86 1.36 1,147 3.04 1.35 Respect people's rights** Make decisions based on the law and not their personal opinions or biases* 1,224 2.82 1.36 1,166 3.02 1.35 Make fair and impartial decisions with the cases they deal with* 1,204 2.85 1.32 1,150 2.98 1.32 Police legitimacy Police in your community are legitimate authorities** 1,224 3.19 1.40 1,158 3.46 1.37 You and the police want the same things for your community** 1,222 3.11 1.46 1,151 3.28 1.45 Police sincerely try to help people like yourself* 1,227 3.05 1.38 1,158 3.20 1.41 Police generally have the same sense of right and wrong that you do** 1,227 2.94 1.44 1,156 3.13 1.42 The laws that the police enforce represent the moral values of people like yourself** 1,220 2.95 1.39 1,152 3.13 1.38 You generally support how the police act in your community** 1,222 2.86 1.42 1,158 3.12 1.41 When the police arrest a person, there is good reason to believe that person has done something wrong 1,223 3.07 1.39 1,151 3.17 1.38 Values of most of the police officers who work in your community are similar to your own** 1,197 2.90 1.38 1,143 3.10 1.39 Police stand up for values that are important to you** 1,221 2.85 1.40 1,149 3.02 1.41 When police deal with people, they almost always behave according to the law** 1,217 2.75 1.35 1,151 2.99 1.40 The police often arrest people for no good reason 1,221 3.09 1.39 1,152 3.04 1.42 The police usually act in ways consistent with your own ideas about what is right and wrong** 1,217 2.86 1.35 1,146 3.04 1.35 Community-focused policing The police department is responsive to community concerns** 1,230 2.69 1.45 1,165 2.93 1.41 The police department prioritizes problems most important to your community** 1,219 2.64 1.46 1,157 2.83 1.43 The police department holds officers accountable for wrong or inappropriate conduct in the community* 1,219 2.49 1.44 1,157 2.63 1.45 Relatability to the police You can imagine being friends with a police officer* 1,233 3.05 1.58 1,160 3.21 1.56 You feel comfortable around the police** 1,235 2.90 1.50 1,163 3.14 1.48 You feel safe around the police** 1,229 2.93 1.49 1,158 3.12 1.46 The police are a part of your neighborhood 1,226 2.90 1.53 1,155 3.02 1.52 You personally trust the police** 1,233 2.63 1.48 1,160 2.85 1.45 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 33 Wave 1 Wave 2 N Mean SD N Mean SD The police are honest** 1,229 2.56 1.32 1,158 2.72 1.33 Police bias Police officers will treat you differently because of your race/ethnicity** 1,230 3.46 1.55 1,157 3.28 1.55 Police officers will judge you based on your race/ethnicity** 1,231 3.42 1.53 1,152 3.26 1.55 Something you do might be misinterpreted as criminal by the police due to your race/ethnicity 1,232 3.30 1.56 1,152 3.19 1.55 The police act based on personal prejudices or biases** 1,225 3.45 1.47 1,151 3.25 1.49 Something you say might be misinterpreted as criminal by the police due to your race/ethnicity* 1,229 3.30 1.55 1,160 3.15 1.55 The police suspect you of being a criminal because of your race/ethnicity 1,230 3.17 1.63 1,162 3.08 1.58 The law All laws should be strictly obeyed** 1,226 4.17 1.17 1,154 4.34 1.08 People should do what the law says** 1,224 4.15 1.13 1,151 4.30 1.02 Obeying the law ultimately benefits everyone in the community** 1,225 4.10 1.22 1,153 4.27 1.08 The laws in your community are consistent with your own intuitions about what is right and wrong** 1,219 3.64 1.33 1,145 3.88 1.22 A person who disobeys laws is a danger to others in the community** 1,221 3.63 1.38 1,150 3.80 1.31 The laws of our system are generally consistent with the views of the people in your community about what is right and wrong** 1,218 3.45 1.37 1,146 3.75 1.25 It is hard to break the law and keep your self-respect** 1,223 3.46 1.48 1,140 3.63 1.43 Willingness to partner with police 1,262 4.02 1.41 1,189 4.14 1.34 Call the police to report a crime* 1,262 3.96 1.43 1,185 4.07 1.38 Report suspicious activity near your home to the police 1,259 3.77 1.53 1,181 3.90 1.50 Provide information to the police to help find a suspect* 1,257 3.37 1.64 1,185 3.42 1.63 Attend a community meeting with the police to discuss crime in your neighborhood 1,258 3.17 1.64 1,181 3.23 1.65 Volunteer your time to help the police solve a crime or find a suspect 1,258 2.98 1.64 1,181 2.91 1.66 Patrol the streets as part of an organized community group 1,262 4.02 1.41 1,189 4.14 1.34 Notes: SD = Standard deviation; statistically significant differences were assessed using t-tests. * p < 0.05 ** p < 0.01. 34 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES TABLE B.2 Mean Scores on Perceptions of Neighborhood Conditions Wave 1 Wave 2 N Mean SD N Mean SD Neighborhood disorder People selling or using drugs** 1,240 3.27 1.56 1,163 2.89 1.56 Litter, broken glass, or trash on the sidewalks/streets** 1,271 3.16 1.55 1,198 2.97 1.55 Vacant or deserted houses, buildings, or storefronts** 1,261 3.13 1.59 1,187 2.71 1.53 Teenagers hanging around on the streets** 1,271 2.96 1.56 1,192 2.54 1.51 People being drunk or rowdy in public places** 1,265 2.75 1.56 1,192 2.46 1.48 Vandalism or graffiti on buildings or cars** 1,266 2.60 1.49 1,190 2.35 1.46 Frequency of neighborhood violence Shooting or shooting attempts** 1,263 2.77 1.49 1,185 2.45 1.42 Assaults** 1,256 2.25 1.31 1,183 1.94 1.18 Violent fights between neighbors** 1,269 2.12 1.18 1,191 1.89 1.08 Muggings or robberies** 1,257 2.17 1.21 1,180 1.90 1.11 Home burglaries** 1,257 2.19 1.18 1,181 1.93 1.08 Car thefts** 1,251 2.08 1.18 1,183 1.90 1.08 Stabbings or stabbing attempts** 1,246 1.92 1.16 1,185 2.45 1.42 Rapes or sexual assaults* 1,239 1.62 0.97 1,183 1.94 1.18 Personal and vicarious victimization experience Shooting or shooting attempts** 1,251 0.44 0.50 1,183 0.31 0.46 Home burglaries** 1,254 0.35 0.48 1,182 0.25 0.43 Assaults** 1,256 2.25 1.31 1,183 1.94 1.18 Violent fights between neighbors** 1,255 0.30 0.46 1,185 0.22 0.41 Car thefts** 1,252 0.29 0.45 1,184 0.23 0.42 Muggings or robberies** 1,255 0.29 0.45 1,185 0.20 0.40 Stabbings or stabbing attempts** 1,255 0.23 0.42 1,181 0.14 0.35 Rapes or sexual assaults** 1,252 0.11 0.32 1,178 0.08 0.27 In the past year, have you been a victim of any of the above crimes in your neighborhood? 1,174 0.16 0.37 1,114 0.14 0.35 Concerns about various property and violent crimes Having your home broken into** 1,255 2.94 1.66 1,184 2.67 1.60 Being mugged or robbed** 1,252 2.88 1.66 1,184 2.64 1.61 Being shot or shot at** 1,253 2.82 1.70 1,183 2.52 1.66 Having your car stolen** 1,221 2.66 1.66 1,168 2.39 1.58 Being physically assaulted** 1,249 2.61 1.64 1,179 2.37 1.55 Being stabbed** 1,248 2.43 1.65 1,182 2.19 1.56 Being sexually assaulted** 1,252 2.32 1.65 1,179 2.07 1.51 Neighborhood safety Your block** 1,253 3.44 1.37 1,188 3.72 1.23 Your closest local public transportation station or stop** 1,222 3.39 1.37 1,145 3.62 1.27 The way to/from the local public transportation station** 1,218 3.34 1.36 1,141 3.55 1.28 Neighborhood stores** 1,227 3.21 1.37 1,165 3.41 1.34 Neighborhood parks** 1,205 3.11 1.42 1,162 3.34 1.37 Your neighborhood** 1,250 3.16 1.35 1,187 3.45 1.30 Precautionary behavior Stay in at night* 1,260 3.35 1.63 1,190 3.21 1.64 Not travel alone** 1,258 2.99 1.64 1,182 2.80 1.61 Avoid certain buildings* 1,256 2.95 1.70 1,183 2.80 1.65 Avoid certain streets 1,257 2.84 1.69 1,189 2.76 1.64 Carry a weapon 1,248 2.21 1.62 1,178 2.14 1.64 Notes: SD = Standard deviation; statistically significant differences were assessed using t-tests. * p < 0.05 ** p < 0.01 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES 35 References Blumberg, Stephen J., and Julian V. Luke. 2007. “Coverage Bias in Traditional Telephone Surveys of Low-Income and Young Adults.” Public Opinion Quarterly 71 (5): 734–49. de Leeuw, Edith, and Wim de Heer. 2002. “Trends in Household Survey Nonresponse: A Longitudinal and International Comparison.” In Survey Nonresponse, edited by Robert M. Groves, Don A. Dillman, John L. Eltinge, and Roderick J. A. Little, 41–54. New York: Wiley. Jannetta, Jesse, Sino Esthappan, Jocelyn Fontaine, Mathew Lynch, and Nancy La Vigne. 2019. Learning to Build Trust and Justice: Implementation Assessment Findings from the Evaluation of the National Initiative for Building Community Trust and Justice. Washington, DC: Urban Institute. La Vigne, Nancy, Jocelyn Fontaine, and Anamika Dwivedi. 2017. “How Do People in High-Crime, Low-Income Communities View the Police?” Washington, DC: Urban Institute. Lawrence, Daniel, Nancy La Vigne, Jesse Jannetta, and Jocelyn Fontaine. 2019. Impact of the National Initiative for Building Community Trust and Justice on Police Administrative Outcomes. Washington, DC: Urban Institute. Nadal, Kevin L., Kristin C. Davidoff, Neil Allicock, Christine R. Serpe, and Tanya Erazo. 2017. “Perceptions of Police, Racial Profiling, and Psychological Outcomes: A Mixed Methodological Study.” Journal of Social Issues 73 (4): 808–30. Peck, Jennifer H. 2015. “Minority Perceptions of the Police: A State-of-the-Art Review.” Policing: An International Journal of Police Strategies & Management 38 (1): 173–203. Schafer, Joseph A., Beth M. Huebner, and Timothy S. Bynum. 2003. “Citizen Perceptions of Police Services: Race, Neighborhood Context, and Community Policing.” Police Quarterly 6 (4): 440–68. 36 EVIDENCE OF CHANGE I N NATIONAL INITIATIV E CITIES About the Authors Jocelyn Fontaine is a senior fellow in the Urban Institute’s Justice Policy Center. Her research incorporates quantitative and qualitative methodologies to evaluate the impact of community-based violence reduction and policing initiatives. Sino Esthappan is a research analyst in the Justice Policy Center. His research focuses on policing, juvenile justice, and school discipline. Nancy La Vigne is vice president for justice policy at the Urban Institute. She manages a staff of more than 50 scholars and conducts her own research on policing, criminal justice technologies, and reentry from incarceration. Jesse Jannetta is a senior policy fellow in the Justice Policy Center, where he leads research and technical assistance projects on reentry, justice system change efforts, community-based violence reduction strategies, and the practice of risk assessment in the criminal justice system. Acknowledgments This brief was supported by cooperative agreement number 2014-MU-MU-K051 awarded by the US Department of Justice, Office of Justice Programs, and Office of Juvenile Justice and Delinquency Prevention. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders. Funders do not determine research findings or the insights and recommendations of Urban experts. Further information on the Urban Institute’s funding principles is available at www.urban.org/fundingprinciples. We are grateful for the hard work of the surveyors from community organizations across the six cities: the Birmingham Urban League, One Safe Place, Neighborhood Services Inc., Minneapolis Urban League, Urban League of Greater Pittsburgh, and Fathers & Families of San Joaquin. Survey collection, entry, and analysis were accomplished by a qualified team of fellow researchers at Urban, including, Dan Lawrence, Emma Kurs, Joshua Eisenstat, Hanna Love, Erica Kouka, Isela Banuelos, Cameron Okeke, Shani Buggs, and Laura Golian. We also appreciate each city’s police department and their cooperation in providing data to support this research: Birmingham Police Department, Fort Worth Police Department, Gary Police Department, Minneapolis Police Department, Pittsburgh Bureau of Police, and Stockton Police Department. ABOUT THE URBAN INST ITUTE The nonprofit Urban Institute is a leading research organization dedicated to developing evidence-based insights that improve people’s lives and strengthen communities. For 50 years, Urban has been the trusted source for rigorous analysis of complex social and economic issues; strategic advice to policymakers, philanthropists, and practitioners; and new, promising ideas that expand opportunities for all. Our work inspires effective decisions that advance fairness and enhance the well-being of people and places. Copyright © August 2019. Urban Institute. Permission is granted for reproduction of this file, with attribution to the Urban Institute. 500 L’Enfant Plaza SW Washington, DC 20024 www.urban.org