Community Policing: Least Effective Where Need Is Greatest

Abstract

Little is known about the level or impact of community policing in suburban and rural communities. We surveyed more than 1,300 cities and counties and asked city managers about social cohesion, collective efficacy, and community policing variables. We find no effect of community policing on perception of safety and a positive effect on community participation only in the metro core. For suburbs and rural areas, community policing is only related to youth services. Collective efficacy is positively associated with safety perceptions across all communities but only related to community participation in suburbs and low crime communities. These results raise questions on how to better link collective efficacy and community policing in high crime communities, especially in suburbs and rural areas.

Keywords

collective efficacy community policing families and youth

Crime scholars regularly lament a dearth of research focused outside of major cities (Maguire, Kuhns, Uchida, & Cox, 1997; Wells & Weisheit, 2004). This bias occurs even though most policing agencies are located outside metro core areas. Eighty-seven percent of departments serve populations of less than 25,000 people, and 71% of departments serve populations of less than 10,000 people (Reaves, 2015). In light of the recent riots in Ferguson, Missouri, and the drug/HIV issues experienced by Scott County, Indiana, the need for more research into policing in suburban and rural communities is apparent.

Community policing is an area where this bias is quite visible. Over the past few decades, criminal justice agencies have realized the role they play in building collective efficacy (Sampson & Raudenbush, 1999). Community policing strategies were designed to give residents greater empowerment by allowing them to participate in discussions on policing and work more closely with officers on the street (Community Oriented Policing Services, 2014). Examples of strategies used in community policing include citizen advisory panels, neighborhood watch, and police-run youth programs (Maguire et al., 1997). Literature indicates that law enforcement culture can make citizen empowerment difficult. Critics contend that community policing can frequently be implemented in ways that provide little empowerment, although the focus of this research has been on large cities (Reisig & Parks, 2004).

In this article, we examine how community policing and collective efficacy affect three important aspects of livable communities. In our analysis, we look at whether community policing strategies and measures of collective efficacy are correlated with perceptions of safety, community participation, and service delivery, specifically youth services. We use data obtained from a survey of city managers, an important population, as this group is familiar with both the dynamics of community organizing and building partnerships between citizens and governmental agencies. Unlike many crime studies, our data come from a cross section of communities across the nation with most data coming from rural and suburban communities, enabling a broader view of community policing.

Community Policing

Sampson and Raudenbush (1999) argued that contemporary policing models, which focus largely on removing lawbreakers, are misguided. They contend that these strategies are ineffective because they do not address the social issues which cause crime to occur. Instead, they believe that the best way to fight crime is to increase social cohesion by bringing communities together to address common goals, something they call collective efficacy.

This logic forms the foundation of the community policing model. Important elements of the model include changing police focus from law enforcement to problem solving, making policing less reactive (Moore, 1992). Proponents argue that community policing strategies should make communities feel safer and reduce crime (Goldstein, 1979). Using a holistic approach to crime prevention, community policing advocates that law enforcement be actively involved in the communities they protect by working with vulnerable groups, such as the elderly, youth, and the poor, to link them to services (Cordner, 2015). Providing such assistance ultimately serves to make communities safer by stopping crime before it occurs. Key to community policing is increasing officers’ visibility in their neighborhoods by spending more time interacting with community members (Cordner, 2015).

Community policing encourages departments to increase citizen participation by engaging with community members to define problems and jointly create strategies to address those problems (Kang, 2015). The goal is to get both officers and citizens invested in a community crime prevention strategy. This process is believed to enhance trust levels between officers and community members, so that both groups act as partners (Cordner, 2015).

Community policing has gained increasing popularity. However, many departments experience implementation issues. Given the paramilitary management structure of most police departments, opening up the process to include citizen feedback creates difficulties (Kelling, Wasserman, & Willliams, 1988). In addition, increased emphasis on proactive policing and social service delivery can be difficult for a workforce culturally rooted to be reactive (Chappell & Lanza Kaduce, 2010).

On the citizen side, building trust with policing agencies is often difficult, especially in communities where there are perceptions of unfair, negative interactions with police, and in communities where residents perceive lower levels of collective efficacy (Nix, Wolfe, Rojeck, & Kaminski, 2015). However, some research has shown that when communities trust police, they feel safer (Skogan & Frydll, 2004). Collective efficacy theory recognizes the link between social cohesion, increased services, and crime reduction (Nix et al., 2015; Rukus & Warner, 2013; Sampson & Bartusch, 1998). Social cohesion can be a catalyst to increase attention to the needs of vulnerable groups, such as children and youth (Warner & Rukus, 2013).

Citizen Participation

Citizen participation is a key component of community policing, but creating effective dialogue between community members and government can be challenging. There is no universal model, so police agencies have a great deal of latitude in designing community policing programs (Maguire et al., 1997). Some models focus on issuing citations based on quality-of-life violations, such as loitering (Sousa & Kelling, 2006), while others attempt to develop dialogues with police engaging citizens in community building (Maguire et al., 1997). Research has shown that community policing, which encourages participation, increases confidence in police, whereas actions focused on quality-of-life issues decrease confidence in police (Ren, Cao, Lovrich, & Gaffney, 2005).

A major obstacle is engaging residents in poor communities due to a number of factors, including distrust of police, failure to do sufficient outreach, or fear of reprisal (Grinc, 1998; Kang, 2015; Lyons, 2002). When citizens do participate, they frequently feel their voices are not heard. Describing early experiences, Rosenbaum and Lurigio (1994) noted, “One of the basic problems was . . . community contact for the sake of community contact, rather than a means to solve specific community problems” (p. 303). It appears that law enforcement continues to face similar challenges (Lyons, 2002; Skogan, 2006). In Russell and Gascón’s (2014) study of community policing meetings in Los Angeles, the researchers reported that citizen/police interactions often took the form of petitioning the police for relief, and police responding with what they could and could not do, as opposed to dialoguing with citizens. In certain instances, officers dismissed citizens’ concerns. Studies have found these issues to be less of an obstacle in neighborhoods with more civic engagement (Lyons, 2002; Skogan, 2006).

The limited community policing literature focusing outside large city departments suggests that this dynamic might be different in rural and suburban environments (Connell, Miggans, & McGloin, 2008). Policing in smaller communities involves closer types of citizen interactions, so community policing represents a formal codification of current policing practices. Implicit in the argument is the assumption that increased citizen participation is an element of living in a smaller community. Noted sociologist Tönnies (1887/2001) contended that rural areas are more characterized by informal, personal ties (gemeinsschaft) and urban areas are more characterized by formal, impersonal processes (gesellschaft). But more recent research on rural communities notes the increasing diversity within rural areas, and how economic and technological change has led to the breakdown in community cohesiveness (Shucksmith, Brown, Shortall, Vergunst, & Warner, 2012). This suggests the need for more attention to formal planning, participation, and policing processes in suburban and rural areas.

Perception of Safety

Complicating community policing efforts is the relationship between criminal activity and individual perceptions of safety. One would think that as crime goes down, people’s safety perceptions would increase, but literature frequently shows that actual crime is not correlated with perceptions of safety (Duffy, Wake, Burrows, & Bremmer, 2008; Roberts & Stalan, 1997) unless the data are disaggregated to the neighborhood level (Hipp, 2013). In some instances, crime and safety perceptions are inversely correlated, meaning people in high crime communities perceive them to be safe, whereas people in low crime communities perceive them not to be safe (Parfrement-Hopkins & Green, 2010).

Two groups where the disconnect between perception and reality of crime is greatest are the elderly and parents with children. According to Federal Bureau of Investigation (FBI) victimization surveys, the elderly are least likely to be crime victims, and households with children were less likely to be victimized than those without children (Morgan & Mason, 2014). Yet, these groups are frequently noted as having greater concerns pertaining to public safety (Buffel, Phillipson, & Scharf, 2013; Drakulich, 2015; Tulloch, 2004). Fear for personal safety is consistently correlated with increased isolation. This means that citizen participation in these groups, vital to community policing, may not be occurring due to fear of crime (Buffel et al., 2013; Drakulich, 2015).

Participation of elders and families with children in community planning processes has been shown to have a major impact on government-provided services for these groups, especially in suburban and rural communities (Warner, Homsy, & Morken, 2016; Warner, Xu and Morken, 2017; Warner & Rukus, 2013). Thus, community participation could help shape the neighborhood policing model and lead to increased services. Youth services is one area correlated with a positive impact on lower delinquent behavior. Constructively engaging youth has been shown to decrease crime by helping children get ahead in school (Schweinhart et al., 2005), providing mentorship (Sullivan & Joliffe, 2012), and employment preparation (Heller, 2014). The role of families and youth in the community policing is an especially important element. A gap in the literature is the role family participation plays in community policing, something the current study addresses.

Urban Bias in Police Research

A substantial number of community policing studies focus on large urban departments (Maguire et al., 1997). This is likely because big city crime fighting receives a disproportionate amount of criminal justice resources. While large departments, defined as departments serving 250,000 people or more, represent slightly less than 1% of all local police departments, they employ 41% of the nation’s officers (Reaves, 2015).

However, difficulties arise when researchers attempt to generalize their findings across all agencies because major differences exist between urban, suburban, and rural departments (Falcone, Wells, & Weisheit, 2002). First, there is a large difference in the workforce composition. Officers working for smaller departments are often less educated, paid less, and more likely to be Caucasian and male (Reaves, 2015). Second, structural differences exist. Departments outside major cities have less hierarchical structures, and officers perform a variety of functions as opposed to specializing (Falcone et al., 2002). Third, challenges and results differ. Reported crime outside of large cities is more violent, perhaps because property crime issues tend to be handled informally, and clearance rates, the percentage of cases successfully closed, are higher (Federal Bureau of Investigation, 2015).

Current Study

Using data from a national survey we conducted of city managers, our study explores a number of questions. First, we examine how factors related to community policing relate to perceptions of safety, community participation, and youth services. We give special attention to the role of social cohesion and collective efficacy. Finally, we examine differences across metro status.

Data and Method

The data come from the 2013 Planning Across Generations Survey that one of the authors conducted with the International City County Management Association. The survey was designed based on input from focus groups held around the country in 2012, with city managers and planners, to discuss community activity to promote planning and service delivery that meets the needs of children and seniors. In those focus groups, respondents raised issues about crime, perceptions of safety, and policing approaches. In the final survey, several questions regarding policing—active policing, trust of police, and neighborhood watch—were included. The survey also measured a range of youth services (after-school programs, recreation, mentoring, etc.), as well as levels of participation of families with children, youth, and seniors in community planning processes. In addition, the survey measured community perceptions of safety, social cohesion, and attitudes toward children and seniors both as a resource and as groups that require special services. This survey provides the first national-level sample that enables exploration of how community policing and collective efficacy are related to community perceptions of safety, participation, and services for youth.

The survey was sent to the chief administrative officer of each local government in May 2013 with two follow-up reminders in August and November 2013. The survey was sent to all counties (3,031), all cities more than 25,000 population, a one in three sample of cities between 2,500 and 24,999 population, and a one in 2½ sample of towns and townships more than 2,500. Of the 7,948 municipalities surveyed, 1,478 responded, with a 21% response rate for cities and towns, and a 14% response rate for counties. The sample data are representative of the universe according to a two-sample Kolmogorov–Smirnov test of population for 2012.

One advantage of our data set is the broad coverage of rural and suburban municipalities. We distinguish metropolitan from nonmetropolitan (rural) communities using Office of Management and Budget (OMB) criteria (U.S. Census Bureau, 2013a). Within metropolitan areas, we differentiate principal cities from suburbs using Census Bureau principal city designations (U.S. Census Bureau, 2013b). The sample is representative of metro status categories: principal cities in metropolitan areas (225, 16% vs. 12.5% in universe), suburbs (759, 51% vs. 50% in universe), and rural communities (494, 33% vs. 37.5% in universe). We run separate models for metro core, suburban, and rural communities.

We draw population data from the American Community Survey (2009-2013 rolling averages). Crime rates are obtained from the FBI 2012 Universal Crime Reports. Crime rates are skewed, with the majority of the sample having low crime rates. Thus, we differentiate high crime from low crime communities in our models, as we expect that there may be differences between these communities. We split the sample into low crime communities (below the mean < 0.07 per capita crime, n = 1,072) and high crime communities (above the mean > 0.07 per capita crime, n = 243).

After scrubbing the data and merging with demographic and crime data, we have 1,315 usable responding cities and counties. Table 1 presents descriptive statistics overall. Table 2 presents means by metro status, and high and low crime communities, and provides Scheffe’s tests to show when these subgroup means are significantly different.

Table 1.

Descriptive Statistics of Model Variables.

	Observations	M	SD	Minimum	Maximum
Dependent variables
Perception of safety^a	1,315	7.89	1.62	2	10
Community participation^a	1,315	2.56	1.31	0	6
Services for youth^a	1,315	3.40	1.92	0	7
Independent variables
Community policing
Police trust^a	1,315	1.12	0.91	0	2
Active policing^a	1,315	1.27	0.77	0	2
Eyes on the street^a	1,315	0.81	0.75	0	2
Collective efficacy
Service adequacy^a	1,315	7.52	1.55	2	10
Social cohesion^a	1,315	11.77	1.72	3	15
Socioeconomic variables
Population (log)^b	1,315	9.88	1.34	6.46	14.51
Population growth (2000-2010)^c	1,315	17.09	52.80	−43.06	770.76
Unemployment rate (%)^b	1,315	9.09	3.83	0.17	28.17
Poverty (%)^b	1,315	14.50	8.20	0.29	52.04
Suburban × Poverty	1,315	5.88	7.82	0	46.27
White population (%)^b	1,315	82.74	16.36	13.10	100
Gini^b	1,315	0.43	0.05	0.26	0.61
Per capita income (log)^b	1,315	10.15	0.33	8.62	11.56
Population between 15 and 24 (%)^b	1,315	13.82	5.77	0	71.01
Per capita county crime (%)^d	1,315	0.07	0.18	0	2.95
County (yes = 1)^a	1,315	0.29	0.46	0	1
Metro status
Metro core	1,315	0.16	0.36	0	1
Suburban	1,315	0.52	0.50	0	1
Rural	1,315	0.33	0.47	0	1

Note. FBI = Federal Bureau of Investigation; UCR = Universal Crime Reports.

Planning Across Generations Survey (2013).

American Community Survey (2009-2013).

U.S. Census (2000, 2010), U.S. municipalities and counties.

FBI UCR data.

Table 2.

Means by Metro Status, and High and Low Crime Communities.

	Overall sample
	Metro core	Suburb	Rural	High crime rate	Low crime rate
Dependent variables
Perception of safety^a	7.50²	8.14¹	7.68²	7.91¹	7.89¹
Community participation^a	2.91¹	2.55²	2.41²	2.54¹	2.57¹
Services for youth^a	5.25¹	3.22²	2.80³	2.98²	3.49¹
Independent variables
Community policing
Police trust^a	1.14^1,2	1.19¹	1.01²	1.02¹	1.14¹
Active policing^a	1.44¹	1.34¹	1.08²	1.28¹	1.21¹
Eyes on the street^a	1.09¹	0.91²	0.51³	0.64²	0.84¹
Collective efficacy
Service adequacy^a	7.74¹	7.68¹	7.15²	7.36¹	7.55¹
Social cohesion^a	11.96¹	11.72¹	11.75¹	11.94¹	11.72¹
Socioeconomic variables
Population^b	11.72¹	9.72²	9.27³	9.25²	10.03¹
Population growth (2000-2010)^b	13.18²	26.65¹	3.81²	13.65¹	17.87¹
Unemployment rate (%)^c	10.20¹	8.87²	8.90²	9.03¹	9.10¹
Poverty (%)^c	17.92¹	11.37²	17.82¹	15.10¹	14.36¹
White population (%)^c	73.98³	82.92²	86.65¹	85.92¹	82.01²
Gini^c	0.46¹	0.41³	0.44²	0.43¹	0.43¹
Per capita income^c	10.18²	10.26¹	9.98³	10.11²	10.16¹
Population between 15 and 24 (%)^c	16.60¹	12.85³	14.02²	13.10²	13.98¹
Per capita county crime (%)^d	0.032	0.04²	0.12¹	0.27¹	0.02²
n	206	680	429	243	1,072

Note. Scheffe’s test by metro status and high/low crime: 1 = high, 2 = medium, 3 = low. FBI = Federal Bureau of Investigation; UCR = Universal Crime Reports.

Planning Across Generations Survey (2013).

U.S. Census (2000, 2010), U.S. municipalities and counties.

American Community Survey (2009-2013).

FBI UCR data.

We build three models to address (a) perceptions of safety, (b) participation of seniors, youth, and families with children in community planning processes, and (c) provision of services for youth. We analyze the impact of both community policing and collective efficacy variables on each of these outcomes. In addition, we control for standard socioeconomic variables that are considered to be drivers of crime.

Dependent Variables

Perception of safety

We measure perception of safety from the survey items “Crime rates are low in my community” and “Residents feel safe and secure in streets and parks.” Each of these items is measured on a Likert-type scale of 1 (strongly disagree) to 5 (strongly agree). We combine these two into an index (values 2-10). As shown in Table 1, overall most municipalities feel relatively safe with an average score of 7.89. Table 2 shows that suburbs have a significantly higher perception of safety than metro core or rural communities, but we see no difference between high and low crime communities. This is not a surprise, as literature often indicates that crime rates are not correlated with safety perceptions (Duffy et al., 2008; Roberts & Stalan, 1997). As one of the major goals of both community policing and collective efficacy is to make communities feel safer, we expect a positive correlation with safety perceptions.

Senior, youth, and family participation in community planning

This index was based on the sum of answers to three survey questions: “How engaged are the following groups (seniors, families with children, youth) in planning for their needs?” Answers were coded as follows: 0 = not at all engaged, 1 = somewhat engaged, 2 = very engaged for a total possible score of 6. We find that engagement is average across the sample with a mean value of 2.56 (Table 1). Participation is highest in metro core communities, but there is no difference in high and low crime communities (Table 2). We hypothesize that communities that engage in more community policing will have more youth and family participation in community planning. In addition, because collective efficacy entails citizens coming together to participate, we expect communities with higher collective efficacy to have higher participation levels.

Services for youth

This index measures the existence (yes/no) of seven services for youth. These include five services provided in the community: youth center, family literacy/parenting programs, after-school programs, summer camps, and youth employment programs, and two services provided by informal networks of residents: mentoring children and recreation programs. With a mean value of 3.40, most communities provide approximately half the measured services, but service levels are highest in metro core and lowest in rural communities. We hypothesize that communities with higher rates of community policing will provide more youth services. In regard to collective efficacy, the more a community comes together, the more likely residents will advocate for services, which should result in a higher youth services levels.

Independent Variables

Our primary independent variables of interest are community policing and collective efficacy. Our models test if elements of community policing and collective efficacy differ in their relationship to perceptions of safety, participation, and provision of youth services in models disaggregated by metro status and by high and low crime rates.

Community policing

Based on the community policing literature, we hypothesize that the community policing variables should be positively correlated with our outcome variables: perceptions of safety, participation, and provision of youth services, in all models. Three variables were used to measure different aspects of community policing: trust in police, active policing, and citizen action. For police trust, we combine into a single variable, responses to two dichotomous questions asking if “police are a trusted institution for information and services” for either families with children or seniors. Active policing is measured by a dichotomous response to a “police” option for two questions: “Are any of the following engaged in cross-agency partnerships to serve children or seniors?” and “Does your local government work with the following institutions to deliver information or services?” Eyes on the street, citizen action, is measured by combining dichotomous responses to two questions asking if local governments support or facilitate informal neighborhood programs for “checking in on neighbors” or “neighborhood watch.” We note that police trust is highest in suburbs and lowest in rural communities. Active policing is similar in metro core and suburbs but lower in rural communities. Eyes on the street, citizen action, is highest in the urban core, lower in suburbs and rural communities. These results reflect the fact that formal community policing and neighborhood watch programs are most common in metropolitan areas. We do not see a difference in these community policing variables by high/low crime communities, except for eyes on the street, which is lowest in high crime communities.

Collective efficacy

Collective efficacy combines two concepts: adequate services and social cohesion (Nix et al., 2015; Sampson & Bartusch, 1998; Sampson & Raudenbush, 1999). Our survey provides data on city managers’ perceptions of community attitudes regarding adequacy of services for families and youth, and attitudes regarding diversity, and families as a resource. Items were ranked from 1 (strongly disagree) to 5 (strongly agree). Adequacy of services combined responses to two statements into a single index (values 2-10): “Public schools are of high quality in my community” and “Families with children can find the range of services they need within my community.” Social cohesion is measured by three statements: “My community is not divided by race, class, or old timer newcomer divisions”; “Children are a resource for the community”; and “Community has a responsibility to care for children and youth.” We find relatively positive attitudes toward service quality (mean 7.52 of 10) and relatively high social cohesion (11.77 of 15). These measures of social cohesion do not differ by metro status or high and low crime communities, except for service adequacy which is lower in rural communities. We expect these two measures of collective efficacy to be positively related to perceptions of safety, youth, and family participation, and provision of youth services.

Crime rates

Crime rates are drawn from the 2012 FBI Uniform Crime Report which consists of all criminal activity (both violent and property crimes) brought to the attention of law enforcement. These data are drawn from police agency reports and aggregated to the county level. The FBI does not provide crime data at the municipal level because policing agencies often cross jurisdictions. For each municipality, we use the crime rate of the county in which it is a part, and divide the number of crimes by the population of the county. Table 2 shows that crime is highest in rural communities on a per capita basis, and lowest in suburbs and metro cores. Crime can vary substantially across a county, thus we include county as a control, as 29% of our sample is counties.

We expect lower participation in higher crime communities. While literature argues that participation is higher in suburban and rural communities due to smaller size and greater social cohesion (Connell et al., 2008; Frank & Leiderbach, 2003), our data do not support that. Despite similar rates of social cohesion, formal community participation rates are lower in suburbs and rural areas. Based on research showing the positive impacts of youth services on education and employment, both of which decrease the likelihood of criminal activity (Sampson & Laub, 1993; Schweinhart et al., 2005), we expect crime rates to be negatively related to youth services.

Socioeconomic variables

The urban literature shows clearly that crime is higher in communities with declining populations, more youth, more poverty, more inequality, more unemployment, and in communities with more minorities (Venkatesh, 2006). We would expect these communities to have lower safety perceptions, community participation, and youth services (Sampson, Raudenbush, & Earls, 1997). But what do we find in a sample that is more representative of suburban and rural communities in the United States?

Data for our socioeconomic control variables come from the American Community Survey (rolling averages 2009-2013) and the 2000 and 2010 U.S. Census. We control for population growth, and find that it is almost twice as high in the suburbs (26.7%) as in the metro core (13.2%). Rural areas show miniscule growth rates (3.8%). In fact, 333 places have negative population growth from 2000 to 2010, and most of these are rural, 167, a third of our rural subsample. We also find population decline among suburbs (135) and metro core (31) but these are a smaller percentage of those subsamples. We expect higher crime in places with declining populations. Unemployment is higher in the metro core, but poverty and inequality are higher in both metro core and rural communities. Income is highest in suburbs, lower in metro cores, and lowest in rural communities.

We include population age 15 to 24 as this group is most likely to commit crime (Hirschi & Gottfredson, 1983). We note that rural and suburban communities have lower percentages of young adults (age 15-24) than metro cores. At first, this was a surprise, but we disaggregated this group by youth (15-17) and young adults (18-24), and found that suburbs and rural places retain a lower percentage of young adults than the metro core, and this explains the lower rates for this group as a whole.¹ High crime communities have also lower percentages of young adults. One explanation is that crime rates are highest in rural communities, and some of the highest crime rural places are recreation destinations with small populations such as Denali, Alaska. Other rural places may face higher crime because they are declining, and young adults have moved away to seek opportunities elsewhere, as indicated by rural growth rates which are miniscule compared suburban and metro cores.

Suburbs traditionally rely on spillover benefits from neighboring cities, and thus typically have a more limited array of services and less robust planning processes (Warner et al., 2016). However, suburban communities are diversifying, and poverty has increased more in suburbs than in the metro core in the last decade (Frey, 2015). We wanted to investigate the impact of suburban poverty on our dependent variables as poor suburban communities, such as Ferguson, Missouri, tend to have more social challenges, less professional police departments, and face more fiscal stress (Dolan & Carr, 2015; U.S. Department of Justice Civil Rights Division, 2015). As suburbs develop big city problems, due to rising poverty, they need to develop service delivery and finance strategies to meet public needs (Kim & Warner, 2016). We created an interaction term of Suburb × Poverty Rate and expect higher poverty suburban communities will rank lower on all of our dependent variables.

Results

Table 3 presents our models for perception of safety, community participation, and youth services differentiated by metro status.² What stands out in our results is that community policing variables are generally not significant on perceptions of safety or community participation in any models. Where we see a consistent positive impact of community policing across metro status is with youth services, where both active policing and citizen action (eyes on the street) are positively associated with youth services. For rural communities, police trust also leads to more services for youth. These results suggest that the most important community policing impact may be in building coalitions among police, other local government agencies and citizens to support increased services for youth. The results on trust show an important urban and rural difference. Trust leads to more community participation in the metro core. Indeed, this has been a focus of urban community policing programs (Russell & Gascón, 2014). In the rural model, where police trust is higher, perceptions of safety are lower, a finding that deserves additional research.

Table 3.

Regression Models: Perception of Safety, Community Participation, Services for Youth (Standardized Betas).

	Metro core			Suburban			Rural
	Perception of safety	Community participation	Services for youth	Perception of safety	Community participation	Services for youth	Perception of safety	Community participation	Services for youth
Community policing
Police trust^a	0.01	0.29**	0.01	−0.05	0.06	0.02	−0.11*	−0.03	0.10*
Active policing^a	0.11	−0.10	0.24**	0.07	0.08	0.10**	0.02	0.06	0.11*
Eyes on the street^a	−0.00	0.02	0.22**	0.05	0.04	0.12**	−0.06	−0.04	0.17**
Collective efficacy
Service adequacy^a	0.39**	0.08	0.12	0.28**	0.16**	0.16**	0.28**	0.10	0.08
Social cohesion^a	0.12*	0.06	0.06	0.10**	0.14**	0.02	0.12**	0.09	0.01
Socioeconomic variables
Population^b	−0.05	0.04	0.26**	−0.11**	0.07	0.37**	−0.24**	−0.04	0.45**
Population growth^b	0.03	−0.09	−0.04	0.07*	0.05	−0.00	0.13**	−0.04	−0.08
Unemployment rate (%)^c	−0.04	−0.04	0.13	−0.09*	−0.03	−0.01	0.03	0.10	−0.05
Poverty (%)^c	−0.31**	−0.19	0.08	−0.31**	0.00	0.13**	−0.09	−0.15*	−0.04
White population (%)^c	0.12	−0.13	0.07	0.01	−0.01	−0.08*	0.14*	0.09	−0.07
Gini^c	0.11	0.10	−0.03	0.08*	0.01	0.05	0.04	−0.00	0.06
Population between 15 and 24 (%)^c	0.15	0.16	−0.01	0.05	−0.01	−0.04	0.03	0.03	0.05
Per capita county crime^d	0.10	−0.04	0.01	−0.03	0.03	0.04	0.00	−0.01	−0.02
County	−0.16*	−0.06	0.04	0.03	−0.03	−0.03	−0.00	0.13*	−0.14**
Perception of safety^a	—	0.00	−0.04	—	0.02	0.02	—	−0.06	0.11*
Community participation^a	0.00	—	0.21**	0.01	—	0.12**	−0.06	—	0.21**
Services for youth^a	−0.03	0.23**	—	0.02	0.16**	—	0.14*	0.28**	—
n	206	206	206	680	680	680	429	429	429
R ²	.43	.22	.28	.32	.18	.37	.23	.15	.37

Note. FBI = Federal Bureau of Investigation; UCR = Universal Crime Reports.

Planning Across Generations Survey (2013).

U.S. Census (2000, 2010).

American Community Survey (2009-2013).

FBI UCR data.

p < .05. **p < .01.

By contrast, our measures of collective efficacy have a positive relation to safety perceptions in all three models, but collective efficacy is only positively related to community participation and youth services in the suburb models. Recall suburbs rank lower on both services and participation, so collective efficacy can make a difference in promoting more participation and services in suburban communities. Rural areas also rank lower on services and participation, but our models find no impact of collective efficacy on participation and service levels in rural communities. This challenges traditional conceptions of rural areas as being more cohesive based on informal, personal social ties.

Regarding our socioeconomic variables, we find that safety perceptions are lower in suburbs and rural communities with larger populations, but higher when those communities are growing. This makes sense, as decline is associated with higher crime. Perception of safety is lower in metro core and suburbs with higher poverty, as expected. Poverty had no effect in rural areas, but rural areas with more White population had a higher perception of safety. Suburbs with higher income inequality showed higher safety perceptions. Per capita county crime had no effect on safety perceptions, as expected. Counties had lower safety perceptions in the metro core. Recall the crime variable is at the county level, so it is reasonable that counties have a lower perception of safety, all else equal.

None of our socioeconomic controls are significant in the community participation models, except for the rural model where poverty was associated with less participation and county was associated with more. This makes sense as counties may have more capacity than smaller municipalities in nonmetropolitan areas to promote community participation in planning.

Across metro status, we find that places with larger population have more youth services, but only in suburbs do places with more poverty offer more youth services. In no model does the proportion of youth in the population affect youth service levels. Thus, service levels are primarily a function of population size, community policing, and social cohesion.

Our second set of models, presented in Table 4, provide results for high crime and low crime communities.³ None of the community policing variables are significant in the high crime models, except for the positive impact of eyes on the street on the level of youth services. In the low crime models, police trust leads to more community participation, and active policing and eyes on the street lead to more services for youth. Thus, community policing appears to work best in low crime communities.

Table 4.

Regression Models: Perception of Safety, Community Participation, Services for Youth (Standardized Coefficients).

	High crime rate			Low crime rate
	Perception of safety	Community participation	Services for youth	Perception of safety	Community participation	Services for youth
Community policing
Police trust^a	−0.06	−0.02	0.08	−0.05	0.08*	0.04
Active Policing^a	0.11	0.11	0.03	0.05	0.05	0.11**
Eyes on the street^a	−0.02	−0.01	0.17**	0.03	0.02	0.13**
Metro status
Rural	—	—	—	0.05	0.08	−0.06
Suburban	−0.00	0.18	−0.13	0.19**	−0.10	−0.23**
Suburban × Poverty	−0.14	−0.16	0.15	−0.15**	0.12	0.11*
Collective efficacy
Service adequacy^a	0.26**	0.14	0.04	0.31**	0.12**	0.15**
Social cohesion^a	0.16*	0.10	0.06	0.09**	0.12**	0.01
Socioeconomic variables
Population^b	−0.17*	−0.06	0.47**	−0.15**	0.06	0.42**
Population growth^b	0.13*	−0.03	−0.01	0.05	0.05	−0.00
Unemployment rate (%)^c	0.05	0.07	−0.06	−0.06	−0.02	0.01
Poverty (%)^c	−0.08	−0.07	−0.07	−0.08	−0.05	0.04
White population (%)^c	0.07	0.02	−0.10	0.03	−0.01	−0.01
Gini^c	0.00	−0.00	0.05	0.02	−0.03	0.03
Per capita income^c	0.07	−0.00	−0.01	0.15*	0.12	0.01
Population between 15 and 24 (%)^c	0.01	−0.01	0.09	0.07*	0.05	−0.01
Per capita county crime (%)^d	−0.04	−0.02	−0.05	0.00	0.06*	−0.02
County	0.01	0.12	−0.07	−0.02	0.02	−0.06*
Perception of safety^a	—	−0.02	0.05	—	−0.01	0.03
Community participation^a	−0.02	—	0.16**	−0.01	—	0.14**
Services for youth^a	0.06	0.24**	—	0.04	0.21**	—
Constant	—	—	—	—	—	—
n	243	243	243	1,072	1,072	1,072
R ²	.24	.15	.43	.33	.17	.47

Note. FBI = Federal Bureau of Investigation; UCR = Universal Crime Reports.

Planning Across Generations Survey (2013).

U.S. Census (2000, 2010).

American Community Survey (2009-2013).

FBI UCR data.

p < .05. **p < .01.

Collective efficacy is positively correlated with safety perceptions in both high and low crime models, but only in the low crime models does it have a positive effect on participation and youth services. Our controls show no differences in high crime communities by metro status, while in low crime communities, suburbs have a higher safety perception but lower services for youth. Our interaction term shows that suburbs with high poverty have a slightly lower perception of safety but slightly higher service levels. Regarding our control variables, in high crime communities, perception of safety is lower when population is larger in both high and low crime models. But population growth has a positive effect on safety perceptions in high crime communities only. Almost half of high crime communities experienced actual population decline from 2000 to 2010.⁴ Communities experiencing decline often face higher rates of crime (Cullen & Levitt, 1999). This notion is also supported in the low crime models where communities that retain their young adults have higher safety perceptions.

No variables affect community participation in the high crime model except for services for youth, which is correlated with higher participation. In the low crime model, collective efficacy and per capita county crime are both positively correlated with community participation. In low crime communities, we see a positive effect of social cohesion, and collective efficacy is not seen in the high crime communities.

Youth services models show that eyes on the street, community participation, and population lead to more services for youth in both the high and low crime models. Thus, citizen action primarily leads to more youth services. Only in the low crime model do we find active policing and collective efficacy associated with more services for youth. Counties have lower services for youth, in part because many of these services are neighborhood based and not provided by county government.

Discussion

Together, these models offer some disturbing results. Community policing is never associated with safety perceptions. Police trust is only linked to community participation in the metro core and low crime communities. One of the main goals of community policing is to provide communities with needed services. Youth services are particularly important because of their ability to positively influence outcomes later in life (Sampson & Laub, 1993; Schweinhart et al., 2005). In low crime communities, two of the three community policing variables are correlated with increased services. The dynamic in high crime neighborhoods is different, where only eyes on the street is associated with increased services. It appears that community policing is working best in terms of youth services, where it is needed least, low crime communities, and is not working where it is needed most.

Similar to our results of community policing, collective efficacy appears to have the most impact where it is needed least—in low crime communities. In these communities, collective efficacy is correlated with increased safety perception, community participation, and youth services. By contrast, in high crime communities, collective efficacy only has a positive impact on safety perception. This could be an endogeneity problem, as communities could be in the low crime category because of collective efficacy, an effect our model cannot disentangle.

Collective efficacy may play a greater role than community policing, as these variables are more likely to be significant and the standardized betas are larger. From a policy perspective, these results suggest that community policing could be more effective if its focus were on building collective efficacy. Literature shows that cultural issues in traditional police management can cause problems in trying to build social cohesion (Chappell & Lanza Kaduce, 2010). For community policing to have more impact, departments may need to give officers more training in cultural sensitivity and community building.

Our study is unique, in that it draws from a cross section of communities which are representative of the nation. Hence, it provides a better look at dynamics more representative of a typical police department. This is important because most police agencies operate in suburban or rural environments (Reaves, 2015). Our research demonstrates that generalizing the findings from metro core departments may not be appropriate to understand suburbs and rural communities. Suburbs, which make up the largest number of municipalities in the United States, show a divergence. While suburbs generally have a higher perception of safety, higher poverty suburbs have a lower safety perception. While youth services are lower in suburbs in general, they are slightly higher in higher poverty suburbs.

Limitations

As is the case with all studies, our study has limitations. First, our measures come from city managers, not police or citizens. This is both a benefit and a risk. The benefit is that city managers see the community in its totality and are in a good position to assess attitudes and services. The risk is that they may not be as close to policing concerns. Second, our measures are primarily attitudes, and these are necessarily subjective. Third, our crime rate variable, being an objective measure, was only available at the county level. We cannot determine the actual crime levels in individual municipalities because crime rates can vary throughout the county. The cross jurisdictional nature of policing makes this unavoidable as the unit of analysis for Uniform Crime Reports is the policing entity. Finally, implementation of community policing practices and departmental acceptance can play a major role in the success of these efforts (Chappell & Lanza Kaduce, 2010). We are unable to measure these aspects from our survey data, but future research should do so.

Our study is a point in time snapshot, so causality cannot be determined. More research is needed on the relationship between collective efficacy and community policing and outcomes of safety, participation, and services. We need to understand why both variables show stronger effects in low crime communities. Do communities experience lower crime rates because they have a successful community policing strategy, or is it easier to implement community policing in low crime communities? Answering this question requires longitudinal data which assess specific community policing strategies over time. It would also be helpful to obtain measures of police department cultures to better determine what approaches might break down resistance to community policing practices by rank-and-file officers.

Conclusion

Most studies of community policing focus on large cities, but our data enable us to look across metro status at rural and suburban communities. We find that community policing may be working best where it is needed least. We find no association between community policing and perceptions of safety, and we find a positive association of community policing with community participation only in the metro core and low crime models. This suggests that community policing needs to develop practices that work better in suburban and rural communities, especially those facing high crime. Active policing and eyes on the street programs are associated with higher levels of youth services in low crime communities, but active policing has no effect on youth services in high crime communities. Designing community policing strategies that work in high crime communities, especially in suburban and rural communities, should be an important goal going forward. Our study provides a first step in unraveling the nature of these dynamics. We find that collective efficacy is positively related to perceptions of safety across community types but only associated with more participation and youth services in suburbs and low crime models. Future studies should explore the relation between collective efficacy and community policing. Effective community policing needs active community participation bolstered by social cohesion. Exploring the interactions between these will be key to identifying community policing strategies that work for suburban and rural communities.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported in part by US Department of Agriculture, National Institute for Food and Agriculture Grant no. 2011-68006-30793.

Notes

Author Biographies

Joseph Rukus is an assistant professor of criminology at Arkansas State University. His research focuses on community-based crime fighting strategies, social justice, prisons, and jails.

Mildred E. Warner is a professor of city and regional planning at Cornell University. Her research focuses on local government service delivery, community development policy, and economic development.

Xue Zhang is a PhD student in regional science at Cornell University. Research interests are in quantitative modeling and community development.

References

Buffel

Philllipson

Scharf

(2013). Experiences of neighborhood exclusion and inclusion among older people living in deprived inner-city areas in Belgium and England. Ageing & Society, 33, 89-109.

Chappell

Lanza Kaduce

(2010). Police academy socialization: Understanding the lessons learned in a paramilitary-bureaucratic organization. Journal of Contemporary Ethnography, 39, 187-214.

Community Oriented Policing Services. (2014). Community policing defined. Washington, DC: U.S. Department of Justice.

Connell

Miggans

McLoin

(2008). Can a community policing initiative reduce serious crime? A local evaluation. Police Quarterly, 11, 127-150.

Cordner

(2015). Community policing: Elements and effects. In Dunham

Alpert

(Eds.), Critical issues in policing: Contemporary issues (7th ed., pp. 481-498). Long Grove, IL: Wavelength Press.

Cullen

Levitt

(1999). Crime, urban flight, and consequences for cities. The Review of Economics and Statistics, 81, 159-169.

Dolan

Carr

(2015). The poor get prison: The alarming spread of the criminalization of poverty. Washington, DC: Institute for Policy Studies.

Drakulich

(2015). Concerns for self or family? Sources of and responses to altruistic fear. Journal of Interpersonal Violence, 30, 1168-1207.

Duffy

Wake

Burrows

Bremmer

(2008). Closing the gap—Crime and public perceptions. International Review of Law, Technology & Computers, 22, 17-44.

10.

Falcone

Wells

Weisheit

(2002). The small-town police department. Policing: An International Journal of Policing Strategies & Management, 25, 371-384.

11.

Federal Bureau of Investigation. (2015). Uniform crime reports 2013. Retrieved from https://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2013/crime-in-the-u.s.-2013/cius-home

12.

Frank

Leiderbach

(2003). The work routines and citizens interactions of small-town and rural police officers. In Thurman

McGarrell

(Eds.), Community policing in a rural setting (2nd ed., pp. 49-60). New York, NY: Routledge.

13.

Frey

W. H.

(2015). Diversity explosion: How new racial demographics are remaking America. Washington, DC: Brookings Institution Press.

14.

Goldstein

(1979). Improving policing: A problem-oriented approach. Crime & Delinquency, 25, 236-258.

15.

Grinc

(1998). Angels in marble: Problems in stimulating community involvement in community policing. In Karp

(Ed.), Community Justice: An Emerging Field (pp. 167-192). New York, NY: Rowman and Littleman.

16.

Heller

(2014). Summer jobs reduce violence among disadvantaged youth. Science, 346, 1219-1223.

17.

Hipp

(2013). Assessing crime as a problem: The relationship between residents’ perception of crime and official crime rates over 25 years. Crime & Delinquency, 59, 616-648.

18.

Hirschi

Gottfredson

(1983). Age and the explanation of crime. American Journal of Sociology, 89, 552-584.

19.

Kang

(2015). Participation in the community social control, the neighborhood watch groups: Individual- and neighborhood-related factors. Crime & Delinquency, 61, 188-212.

20.

Kim

Warner

M. E.

(2016). Pragmatic municipalism: Local government service delivery after the great recession. Public Administration, 94, 789-805.

21.

Kelling

Wasserman

Williams

(1988). Police accountability and community policing (Perspectives on Policing No. 7). Washington, DC: National Institute of Justice.

22.

Lyons

(2002). Partnerships, information, and public safety: Community policing in a time of terror. Policing: An International Journal of Policing Strategies & Management, 25, 530-542.

23.

Maguire

Kuhns

Uchida

Cox

(1997). Patterns of community policing in nonurban America. Journal of Research in Crime & Delinquency, 34, 368-394.

24.

Moore

(1992). Problem-solving and community policing. Crime and Justice, 15, 99-158.

25.

Morgan

Mason

(2014). Crimes against the elderly, 2003-2013. Washington, DC: U.S. Department of Justice.

26.

Nix

Wolfe

Rojeck

Kaminski

(2015). Trust in the police: The influence of procedural justice and perceived collective efficacy. Crime & Delinquency, 61, 610-640.

27.

Parfrement-Hopkins

Green

(2010). Public perceptions. In Flatley

Kershaw

Smith

Chaplain

Moon

(Eds.), Crime in England and Wales 2009/2010 (pp. 109-148). London, England: British Home Office.

28.

Reaves

(2015). Local police departments, 2013: Personnel, policies, and practice. Washington, DC: Bureau of Justice Statics.

29.

Reisig

Parks

(2004). Can community policing help the truly disadvantaged? Crime & Delinquency, 50, 139-167.

30.

Ren

Cao

Lovrich

Gaffney

(2005). Linking confidence in the police with the performance of the police: Community policing can make a difference. Journal of Criminal Justice, 33, 55-66.

31.

Roberts

Stalan

(1997). Public opinion, crime, and criminal justice. Boulder, CO: Westview Press.

32.

Rosenbaum

Lurigio

(1994). An inside look at community policing reform: Definitions, organizational changes, and evaluation findings. Crime & Delinquency, 40, 299-314.

33.

Rukus

Warner

(2013). Crime rates and collective efficacy: The role of family friendly planning. Cities, 31, 37-46.

34.

Russell

Gascón

(2014). Defining “policeability”: Cooperation, control, and resistance in South Los Angeles community-police meetings. Social Problems, 61, 237-258.

35.

Sampson

Bartusch

(1998). Legal cynicism and (subcultural?) tolerance of deviance: The neighborhood context of racial differences. Law & Society Review, 32, 777-804.

36.

Sampson

Laub

(1993). Crime in the making: Pathways and turning points in life. Cambridge, MA: Harvard University Press.

37.

Sampson

Raudenbush

(1999). Systematic social observation of public spaces: A new look at disorder in urban neighborhoods. American Journal of Sociology, 105, 603-651.

38.

Sampson

Raudenbush

Earls

(1997). Neighborhoods and violent crime: A multi-level study of collective efficacy. Science, 27, 918-924.

39.

Schweinhart

Monite

Xiang

Bennett

Belfield

Nores

(2005). Lifetime effects: The high/scope Perry preschool study through age 40. Ypsilanti, MI: High/Scope Press.

40.

Shucksmith

Brown

D. L.

Shortall

Vergunst

Warner

M. E.

(2012). Rural transformations and rural policies in the US and UK. New York, NY: Routledge.

41.

Skogan

(2006). Police and community in Chicago: A tale of two cities. New York, NY: Oxford University Press.

42.

Skogan

Frydll

(2004). Fairness and effectiveness in policing: The evidence. Washington, DC: National Academies Press.

43.

Sousa

Kelling

(2006). Of “broken windows,” criminology, and criminal justice. In Weisburd

Braga

A. A.

(Eds.), Police innovation: Contrasting perspectives (pp. 77-97). Cambridge, MA: Cambridge University Press.

44.

Sullivan

Jolliffe

(2012). Peer influence, mentoring, and the prevention of crime. In Welsh

Farrington

(Eds.), Oxford handbook of crime prevention (pp. 207-255). New York, NY: Oxford University Press.

45.

Tönnies

(2001). Community and civil society ( Harris

, Ed. & Trans.). Cambridge, UK: Cambridge University Press. (Original work published 1887)

46.

Tulloch

(2004). Parental fear of crime: A discursive analysis. Journal of Sociology, 40, 362-377.

47.

U.S. Census Bureau. (2013a). Core based statistical areas and combined statistical areas. Office of Management and Budget. Retrieved from census.gov/population/metro/files/lists/2013/List1.xls

48.

U.S. Census Bureau. (2013b). Principal cities of metropolitan and micropolitan statistical areas. Office of Management and Budget. Retrieved from census.gov/population/metro/files/lists/2013/List2.xls

49.

U.S. Department of Justice, Civil Rights Division. (2015). Investigation of Ferguson police department. Washington, DC: U.S. Department of Justice.

50.

Venkatesh

(2006). Off the books: The underground economy of the urban poor. Cambridge, MA: Harvard University Press.

51.

Warner

M. E.

Homsy

G. H.

Morken

L. M.

(2017). Planning for aging in place: Stimulating a market and government response. Journal of Planning Education and Research, 37, 29-42.

52.

Warner

M. E.

Morken

L. J.

(2016). What explains differences in availability of community health-related service provision for seniors in the United States? Journal of Aging and Health. doi:10.1177/0898264316654675

53.

Warner

Rukus

(2013). Planners’ role in creating family-friendly communities: Action, participation and resistance. Journal of Urban Affairs, 35, 627-644.

54.

Wells

Weisheit

(2004). Patterns of urban and rural crime: A county level comparison. Criminal Justice Review, 29, 1-19.