Neighborhood Disadvantage and Reliance on the Police

Abstract

Contemporary theories suggest that, due to limited access and generalized distrust, residents of disadvantaged neighborhoods are relatively unlikely to report matters to police. Although existing studies reveal few ecological differences in crime reporting, findings may be limited to victim/offense subsets represented in aggregated victimization data. Using calls-for-service (CFS) data from a Pacific Northwest city, this study assesses the degree to which neighborhood block groups (N = 164) vary in incidents reported to police overall and subsequent to the elimination of a major nonemergency-reporting mechanism. Two hypotheses are assessed: First, CFS rates will vary inversely with neighborhood disadvantage, net of the effect of objective levels of crime and other control variables; second, CFS originating in affluent neighborhoods will exhibit greater year-to-year decreases relative to disadvantaged neighborhoods following reduction of local reporting services in 2004. Findings from spatial analyses indicate that residents of disadvantaged neighborhoods tend to rely on police for assistance as much as, if not more than, people elsewhere.

Keywords

crime reporting neighborhood disadvantage police calls for service

Police and other law enforcement officials are called on to perform a variety of important functions. They are expected to protect and serve members of the general public and to do so equitably, without regard for race, class, gender, or any other extralegal variable. There is evidence, however, that law enforcement officers do not always exercise their authority uniformly across differing population groups and that the actions of police are influenced not just by what people do but also by their social status. Indeed, studies show that minorities and members of economically disadvantaged populations often receive inferior police protection and services and experience disproportionately high rates of police scrutiny, arrests, and use of force (e.g., Beckett, Nyrop, & Pfingst, 2006; Browning, Cullen, Cao, Kopache, & Stevenson, 1994; Fagan & Davies, 2000; Kane, 2002, 2005; Mastrofski, Reisig, & McCluskey, 2002; B. W. Smith & Holmes, 2003; D. A. Smith & Klein, 1984; D. A. Smith & Visher, 1981; Terrill, Paoline, & Manning, 2003; Terrill & Reisig, 2003; Worden, 1996; see Weitzer & Tuch, 2006). Especially among African Americans and the poor, these differences have exacerbated, if not engendered, distrust of the police and dissatisfaction with police job performance (see, e.g., Brunson & Miller, 2006; Carr, Napolitano, & Keating, 2007; Jones-Brown, 2000; Weitzer & Tuch, 1999, 2006).

Although recent research suggests that such experiences reduce the likelihood that individuals who belong to affected groups will call the police for help (see, e.g., Carr et al., 2007; see also Avakame, Fyfe, & McCoy, 1999; cf. Carbone-Lopéz, 2007), little is known about ecological variations in reliance on the police. To date, research has detected few crime-reporting differences across neighborhoods (Baumer, 2002; Bennett & Wiegand, 1994; cf. Goudriaan, Wittebrood, & Nieuwbeerta, 2006), contradicting theoretical claims concerning the link between disadvantage and voluntary involvement of police in the affairs of residents (see Anderson, 1999; Black, 1976; Bursik & Grasmick, 1993; Conklin, 1975; Laub, 1981; Manning & Van Mannen, 1978). However, because most relevant studies are based on aggregated victimization data, they reflect only a limited subset of neighborhood residents and criminal offenses and thus do not permit examination of broader neighborhood differences in reliance on the police. Using calls-for-service data from a midsize city in the Pacific Northwest, we apply spatial modeling techniques to examine the relationship between neighborhood characteristics and reliance on police for assistance in dealing with serious and minor crimes, as well as physical and social disorder.

Theories of Neighborhood Reliance on Police

Contrary to current trends, traditional perspectives of neighborhood reliance on the police tended to associate socioeconomic disadvantage with increased dependence on legal bodies of the state for assistance, intervention, and protection. Contributing scholars viewed agents of the law, such as the police, as substituting for informal social controls in communities lacking adequate institutional strength and support (Conklin, 1975; Laub, 1981; see also Manning & Van Mannen, 1978). In The Behavior of Law, Donald Black (1976) formalized this relationship as one of five basic principles regarding the mobilization and behavior of law and its official representatives: “Law is stronger where other social control is weaker” (p. 107).

For reasons not entirely clear, recent research has focused more closely on Black’s (1976) stratification hypothesis and later work pertaining to crime as a means by which people with few legal options address their grievances (see, e.g., Baumer, 2002). In addition to formulating the principle of social control, Black drew on a wealth of historical and contemporary examples to suggest that the quantity of law varies directly with social stratification such that people of low social status—namely, the poor and minorities—are unable to access and utilize legal resources to the same extent as those in more privileged positions. Neglected by law, he said, these people would be more likely than those enjoying greater legal protections to take matters into their own hands as a form of self-help or social control (Black, 1983, 1989).

Although Elijah Anderson’s (1999) work predicts a similar neighborhood-level outcome, the theorized mechanism clearly differs. On the basis of extensive ethnographic observation and interviews, Anderson’s conclusion was that the profound lack of faith in the police and the judicial system among residents of disadvantaged African American communities has contributed to an oppositional culture, or “code of the street,” emphasizing self-reliance, respect, and violence:

The police . . . are most often viewed as representing the dominant white society and as not caring to protect inner-city residents. When called, they may not respond, which is one reason many residents feel they must be prepared to take extraordinary measures to defend themselves and their loved ones against those who are inclined to aggression. Lack of police accountability has in fact been incorporated into the local status system: the person who is believed capable of “taking care of himself” is accorded a certain deference and regard, which translates into a sense of physical and psychological control. The code of the street thus emerges where the influence of the police ends and where personal responsibility for one’s safety is felt to begin. (p. 34)

Thus, Anderson argues that residents of disadvantaged communities are unlikely to call the police for several reasons: first, because of a general lack of confidence in the ability of police to adequately resolve issues; second, because of the perceived risk of poor or substandard treatment; and third, out of fear for the potential loss of respect on the street. As Baumer (2002) notes, Anderson also implies that the willingness of inner-city residents to contact the police may be further diminished by disproportionate involvement in a variety of street crimes (see also Skogan, 1984).

Unlike Black’s (1976) theory, which explains neighborhood variation in reliance on the police as a compositional rather than contextual effect (see Baumer, 2002), Anderson’s (1999) work focuses on a cultural process in which residents of the inner cities learn from their own and others’ experiences to handle their affairs informally, often through the administration of street justice or retaliatory violence (see also Anderson, 1992, 1994; Kubrin & Weitzer, 2003). This point of view complements structural perspectives suggesting that lack of informal social control mechanisms among socially disadvantaged and disorganized neighborhoods is exacerbated by difficulties that residents face in accessing public services (including the police) and in countering policies that worsen neighborhood conditions (Bursik, 1989; Bursik & Grasmick, 1993). This view is also compatible with contemporary policing theories and perspectives, such as broken windows (Skogan, 1989; Wilson & Kelling, 1982) and problem-oriented policing (Goldstein, 1990), which suggest that the failure to attend to minor issues risks a spiral of decay in which crime and other social problems become progressively worse.

Empirical Background

Despite recent interest in the relationship between the police and community characteristics, relatively few studies have examined the extent to which neighborhood contexts affect reliance on the police. Much of the existing research focuses instead on neighborhood variations in police– community cooperation and coproduction (Reisig & Parks, 2004; Scott, 2002; Triplett, Sun, & Gainey, 2005), perceptions of police misconduct and biases (Kane, 2002; Weitzer, 1999, 2000), attitudes toward and satisfaction with the police (Jones-Brown, 2000; Reisig & Parks, 2000; Sampson & Bartusch, 1998; Schuck & Rosenbaum, 2005; Vélez, 2001; Webb & Marshall, 1995), police practices (Fagan & Davies, 2000; Jackson & Boyd, 2005; Klinger, 1997; Mastrofski et al., 2002; Vélez, 2001), and police recording of citizen complaints of crime (Warner, 1997, 2007).

Evidence is limited regarding ecological patterns in reliance on the police. In analyses of data from the 1974–1976 National Crime Survey, for example, Gottfredson and Hindelang (1979) found that poverty levels of neighborhood subgroups did not influence the extent to which victims of violence reported their experiences to the police. Warner’s (1992) analyses of these data also revealed limited effects of neighborhood income inequality on reporting of robbery and assault. Regarding assault, however, Warner noted a significant interaction between neighborhood racial composition (measured as percentage non-White) and race of victim. Whereas assault reporting varied directly with neighborhood racial composition among White victims, a negative effect was observed among Black victims. Using data from the U.S. Census and the National Crime Victimization Survey, Baumer (2002) found that with the exception of the curvilinear relationship detected for reporting of simple assault, there were no significant differences in the extent to which residents of various neighborhoods reported their victimization experiences to the police. Similarly, analyses of survey data gathered from residents of an urban city in the Midwest showed virtually no effect of neighborhood disadvantage or collective efficacy on respondents’ decision to contact police in response to what they perceived to be the most significant problem facing their neighborhoods (Wells, Shafer, Varano, & Bynum, 2006).¹

Cross-national studies have revealed similar findings. Bennett and Wiegand’s (1994) analysis of data from a household survey conducted in Belize showed victim reporting to the police to be unrelated to environment-specific correlates, including neighborhood cohesion, wealth, police effectiveness, and prevalence of drugs and crime. Whereas results from Fishman’s (1979) study in Haifa, Israel, showed that neighborhood disadvantage influences the extent to which residents report their victimizations, the observed effect was small and indirect, mediated by crime severity and attitudes toward police. Research based on victimization data collected in the Netherlands (Goudriaan et al., 2006) also revealed a weak inverse relationship between neighborhood disadvantage and victim reporting for a variety of person and property crimes. Similar to Baumer’s (2002) findings, the observed effect was most pronounced at extremely high levels of neighborhood disadvantage.

Overall, findings from existing studies suggest that reliance on the police is, at best, only marginally related to neighborhood contexts. However, given that most such studies are based exclusively on victim reports, it is unclear whether findings can be generalized beyond individuals reporting a personal connection to crime or if the strength of the relationship between neighborhood disadvantage and reliance on the police depends on the nature of the incident and the amount of reporting discretion involved (see also Laub, 1981; but see Wells et al., 2006).² Using calls-for-service data aggregated to the block-group level, we assess the extent to which neighborhood characteristics affect requests for police assistance in matters spanning a broader range of incidents than considered in past studies. Specifically, we focus on the relationship between neighborhood disadvantage and calls made to police for assistance with violent and property crimes, as well as physical and social disorder.

Data and Methods

Data for this study are from computer-aided-dispatch system records of police-initiated activity and calls for service in a metropolitan city (population size: approximately 200,000) located in the Pacific Northwest. These data consist of more than 200,000 annual police-involved incidents, ranging from reports of suspicious persons to homicides and including a variety of information pertaining to the substantive categorization of the call, response priority, final disposition, date and time, and geographic coordinates for the originating call site.

Computer-aided-dispatch (calls for service) data are unique in that they present an opportunity to unveil a portion of the dark figure of crime and to circumvent some of the problems associated with victimization surveys. These data represent those incidents that citizens believe are worthy of police attention but that may not be reported in future interviews or recorded and counted as crimes in official statistics or victimization statistics. As such, they provide a broad picture of crime and disorder, relatively free of well-known police-response and survey biases (other than the potential screening that occurs at the call-taking level; Warner & Pierce, 1993; cf. Klinger & Bridges, 1997).

Another unique feature of these data is the capability to compare levels of reported incidents both before and after the elimination of an important crime-reporting service within the city under examination (2004 versus 2005). Under fiscal pressures to streamline budgets, local law enforcement officials decided to reduce costs by eliminating an existing 24-hr nonemergency-reporting mechanism. In its place, the jurisdiction expanded the availability of Internet reporting for selected crimes and provided an alternative nonemergency call-in reporting service with limited hours of coverage. Only 911 services were maintained at their previous operating and funding levels.³ The data used in this study thus provide an opportunity for quasi-experimental assessment of the relationship between neighborhood disadvantage and reliance on police before and after the reduction of call services.

Hypotheses

Guided by the contemporary theoretical perspectives described above, we hypothesize that the number of calls for service will vary inversely with neighborhood disadvantage, net of the effects of objective levels of crime and other control variables. If there is relatively less reliance on the police in disadvantaged neighborhoods, owing to distrust or limited access, then we expect to see fewer calls for official assistance originating in these areas than in more affluent neighborhoods. This may be especially true for high-discretion calls, such as disorder and property-related incidents.

Hypothesis 1: The number of calls for service will vary inversely with neighborhood disadvantage, net of the effect of objective levels of crime and other control variables.

The reduction of local 24-hr nonemergency-reporting services on January 1, 2005, reduced the capacity of the community to report minor incidents of interest to police. If contemporary perspectives are correct in suggesting that reporting levels are comparatively low in disadvantaged neighborhoods, then changes in reporting services are unlikely to affect the number of calls in these areas as greatly as in places where residents are theorized to be more reliant.⁴ Thus, we hypothesize that calls for service in advantaged neighborhoods will exhibit greater year-to-year decreases than will calls made in disadvantaged neighborhoods.⁵

Hypothesis 2: Following the reduction of local reporting services in 2005, calls for service originating in more affluent neighborhoods will exhibit greater decreases than will those made in disadvantaged neighborhoods.

Independent Variables

Drawing on prior research, we measure disadvantage for each block group in the city with the following variables derived from the 2000 U.S. Census: median household income, percentage of the population below the official poverty line, percentage of female-headed households with children under 18, percentage of families receiving public assistance, percentage of persons aged 5 and over who have changed residences in the last 5 years, and percent non-White (Baumer, Horney, Felson, & Lauritsen, 2003; Kubrin & Weitzer, 2003; Sampson & Groves, 1989).⁶ Because these variables are highly correlated at the bivariate level, principal components analysis with varimax rotation (Kaiser normalized) was employed. This analysis yielded one factor with an eigenvalue greater than the conventional cutoff of 1.00 and factor loadings higher than .75 (standardized alpha = .90).

Control Variables

For each block group, we include the log of residential population as an offset variable in count models, permitting interpretation of effects in terms of rates of calls (calls per logged population unit) rather than raw counts. When modeling change in the number of calls between 2004 and 2005, we treat prior year’s calls in each block group as an exposure variable, which is automatically log transformed by statistical software in negative binomial regression models. To rule out the possibility that population exerts independent effects on change, we include this variable as a control. In addition, we control for the crime-prone proportion of the population by including a measure of the percentage of males between the ages of 15 and 25. To control for the possibility that reporting practices are influenced by variations in the presence, availability, and vigor of police officers in each area, we also include a measure of the number of officer-initiated incidents within each block group in 2005.

Because tests for spatial autocorrelation are all significant at the p < .001 level (Moran’s I ranges from .302 to .464 across our dependent variables, discussed below), we analyze the data with a spatial lag model (adjacent boundary rook weights, row standardized). We use a spatial lag model rather than a spatial error model to control for spatial dependence, because community characteristics and crime in one area are likely to affect crime and reporting practices in geographically proximate locales (see Land & Deane, 1992). Whereas the spatial error model is preferable when concern is with “omitted (spatially correlated) covariates that if left unattended would affect inference” (Baller, Anselin, Messner, Deane, & Hawkins, 2001, p. 566), the spatial lag model is more appropriate when predictors and events in one place are likely to affect those in another. Lagrange multiplier tests carried out with logged counts and rates confirm the superior fit of the latter model with our data (see Anselin, 2005).

Finally, although our data do not allow us to control for objective levels of each form of crime and disorder, we include a proxy measure based on the rate of 2005 Uniform Crime Report counts for aggravated assaults and robberies.⁷ Although this measure undoubtedly fails to account for all the neighborhood variance in crime and disorder, evidence suggests that serious violent Uniform Crime Report crimes (excluding rape) are the least susceptible to citizen underreporting and among those most likely to be detected by the police, regardless of citizen behavior. In addition, when reliance on the police is modeled as the change in calls from one year to the next, each neighborhood serves as its own control, and differences in levels of crime across neighborhoods are less relevant.⁸

Dependent Variables

Using ArcMap 9.2, we conducted spatial join operations in which theoretically relevant categories of calls were aggregated to the block-group level to provide a count of reported incidents. These categories were constructed to measure a range of events with differential likelihood of being reported: total calls, disorder, threats and harassment, property incidents, and violent incidents. The operational definitions for these categories were then presented to and endorsed by the analytical staff within the planning and research unit of the municipal policing agency. Details on the specific call types within each category are included in the appendix.

To assess our second hypothesis, we constructed a measure of change in calls following elimination of reporting services. Consistent with other measures of crime, data on change in calls was overdispersed, necessitating the use of negative binomial regression. To accommodate the assumptions of this method, several transformations were made to the raw data. For each of the calls-for-service categories, the number of calls in 2005 was first subtracted from the number of calls in 2004. This provided a count of the volume of change in calls subsequent to elimination of services (expressed as a negative number). Because negative binomial regression does not allow negative numbers, it was necessary to add a small constant equal to the maximum increase from 2004 to 2005 for each category. The net result was a variable with a minimum value of 0 for each variable. The mathematical operations are as follows:

Y_{Transformed} = (Y_{2004} - Y_{2005}) + Y_{2005 - 2004 Max Increase}

These transformations and their interpretation are best illustrated through an example. Consider three block groups representing highest (Block Group A), lowest (Block Group C), and no change (Block Group B) in calls with the following characteristics:

Block Group A—Calls for service: 2004 = 1,000; 2005 = 800

Block Group B—Calls for service: 2004 = 1,000; 2005 = 1,000

Block Group C—Calls for service: 2004 = 1,000; 2005 = 1,050

Because Block Group C (the only block group with an increase from 2004 to 2005) produces a disallowed negative value, we reverse this sign by adding a constant equal to the maximum positive increase. This transformation yields a relative measure of change with the greatest decrease from 2004 to 2005 equal to the maximum positive value and with the greatest increase from 2004 to 2005 equaling 0:

Block Group A: 200 fewer calls in 2005 with a constant of 50 added = 250

Block Group B: No change between 2004 and 2005 with a constant of 50 added = 50

Block Group C: 50-call increase in 2005 (–50), with a constant of 50 added = 0

The net result of this transformation is a measure of change in which a score of 0 indicates a slight increase or no change (with an observed maximum of 88 call increases from 2004 to 2005 across all categories), whereas the maximum score is equal to the greatest level of observed decrease in calls for each block group within the respective call category (with an observed maximum of 1,225 decreases in calls from 2004 to 2005 across all categories). Stated another way, a block group with a calculated score of 0 for disorder stayed relatively consistent in yearly levels of reporting despite the elimination of services, whereas a block group with a score of 250 exhibited a greater decrease in calls for police service between 2004 and 2005.

In sum, rather than represent raw volume of calls within a given year, this measure allows us to tap the covariation between neighborhood disadvantage and change in calls from one year to the next. Contrary to a traditional percentage-change measure, the transformed count measure allows us to address the overdispersion of the data through negative binomial regression modeling, as well as assess the extent of change after factoring in the offset for prior year’s calls and incorporating other important controls. Tables 1 and 2 presents descriptive statistics and bivariate correlations for all variables used in the analysis.⁹

Table 1.

Descriptive Statistics

	Variable	M	SD	Min	Max
1	UCR violent: Logged (robbery + aggravated assault + 1/population)	.005	.005	.000	.037
2	UCR vehicle theft (square root transform)	.009	.008	.000	.060
3	Percentage male 15–25 (logged)	7.417	4.427	.000	35.365
4	Officer initiated: (officer initiated / population) ^ .2	.455	.562	.007	4.237
5	Disadvantage	.000	1.000	−1.999	2.464
6	Population (logged)	1,130.506	432.615	484	2,851
7	Citizen CFS 2004	873.341	776.127	96	5,940
8	Disorder CFS 2004	262.787	229.984	16	1,528
9	Threat CFS 2004 (+1)	19.665	17.290	0	119
10	Property CFS 2004	150.890	133.794	14	910
11	Violent CFS 2004	71.976	62.897	2	410
12	Citizen CFS 2005	701.939	657.288	68	4,715
13	Disorder CFS 2005	197.902	181.310	12	1,174
14	Threat CFS 2005 (+1)	12.402	10.996	0	73
15	Property CFS 2005	109.799	100.326	6	621
16	Violent CFS 2005	69.384	61.119	1	333
17	Change [(Citizen CFS 2004 − 2005) + 66]	−171.402	155.277	−1,225	66
18	Change [(Disorder CFS 2004 − 2005) + 39]	−64.884	62.096	−409	39
19	Change [(Threat CFS 2004 − 2005) + 7]	−7.262	9.398	−46	7
20	Change [(Property CFS 2004 − 2005) + 28]	−41.091	47.845	−308	28
21	Change [(Violent CFS 2004 − 2005) + 88]	−2.591	18.266	−77	88

Note: N = 164. URC = Uniform Crime Reports; CFS = calls for service. All descriptive statistics provide information about variables before transformation.

Table 2.

Bivariate Correlations Between Transformed Variables

	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20
2	.706*
3	.236*	.288*
4	.748*	.764*	.258*
5	.632*	.694*	.404*	.701*
6	−.156*	−.009	.243*	.022	.035
7	.588*	.703*	.340*	.730*	.615*	.311*
8	.627*	.721*	.348*	.749*	.678*	.294*	.966*
9	.614*	.677*	.327*	.691*	.654*	.321*	.897*	.915*
10	.536*	.673*	.383*	.688*	.558*	.320*	.935*	.868*	.808*
11	.674*	.699*	.347*	.739*	.694*	.310*	.901*	.927*	.921*	.804*
12	.593*	.714*	.308*	.734*	.611*	.281*	.990*	.955*	.878*	.913*	.875*
13	.635*	.728*	.330*	.751*	.668*	.270*	.956*	.982*	.903*	.860*	.913*	.963*
14	.627*	.648*	.332*	.705*	.681*	.324*	.862*	.887*	.872*	.784*	.897*	.855*	.897*
15	.569*	.731*	.372*	.724*	.603*	.322*	.944*	.904*	.836*	.957*	.830*	.945*	.907*	.823*
16	.710*	.730*	.313*	.744*	.699*	.294*	.891*	.923*	.906*	.791*	.957*	.891*	.932*	.902*	.842*
17	.430*	.490*	.397*	.544*	.488*	.366*	.807*	.785*	.763*	.809*	.802*	.716*	.702*	.688*	.720*	.684*
18	.467*	.543*	.325*	.581*	.559*	.299*	.785*	.836*	.754*	.706*	.767*	.724*	.718*	.665*	.702*	.699*	.859*
19	.396*	.486*	.212*	.447*	.407*	.212*	.641*	.646*	.820*	.568*	.645*	.616*	.611*	.434*	.574*	.612*	.600*	.608*
20	.307*	.349*	.292*	.406*	.295*	.222*	.635*	.532*	.506*	.791*	.509*	.573*	.502*	.467*	.578*	.448*	.752*	.503*	.385*
21	−.055	−.036	.145	.055	.050	.081	.121	.104	.139	.122	.241*	.031	.028	.070	.040	−.051	.473*	.302*	.175*	.256*

Note: See Table 1 for description of each variable.

p < .05.

Analytic Procedure

Because the distribution of calls for each category is overdispersed (with the variance substantially greater than the mean), negative binomial regression techniques are used to analyze the data (see Hilbe, 2007; Osgood, 2000).¹⁰ Following prior research with a spatial component (e.g., Kubrin & Weitzer, 2003), we also employ the Anselin Alternative 2SLS estimator, in which the spatial lag of the number of calls for service is instrumented using the spatial lag of the predicted values. Specifically, predicted values from the first stage of negative binomial regression are multiplied by the spatial weights matrix to create the variable used to control for spatial autocorrelation in the final model. Because correlated errors are a common problem in the use of instrumented variables, we apply the Huber-White/sandwich correction of standard errors in the second stage of estimation (see Hardin, 2002).

Results

Table 3 presents estimates from negative binomial regression models of the calls in each category (using population as the offset variable to approximate a rate). Contrary to expectations of an inverse relationship between reliance on the police and neighborhood disadvantage (Hypothesis 1), these data show comparatively high levels of calls for police service among disadvantaged neighborhoods, even after controlling for official crime. All else equal, a unit increase in disadvantage increases the expected rate of calls anywhere from 15.1% (property incidents: 100 × [exp(.141) – 1]) to 32.3% for calls related to threats/harassment (100 × [exp(.280) – 1]). Although this finding may be at least partially attributed to unmeasured variations in objective levels of crime, it illustrates a relatively strong tendency for residents of disadvantaged neighborhoods to contact the police for assistance in dealing with a variety of issues. As shown, however, the strength of the relationship between neighborhood disadvantage and calls for service is less pronounced for disorder and property incidents, suggesting fewer neighborhood differences for these high-discretion call categories than for those related to threats/harassment and physical violence.¹¹

Table 3.

Negative Binomial Regression of Counts for Calls-for-Service Categories With Population as Offset

		Calls for Service: 2005
Variable		Citizen	Disorder	Threat	Property	Violent
UCR violent	b	.222***	.230***	.310**	.184***	.422***
Logged (robbery + aggravated assault + 1/population)	t (SE)	5.918 (.038)	4.738 (.049)	3.140 (.099)	4.006 (.046)	6.255 (.067)
	p	.000	.000	.002	.000	.000
Percentage male 15 to 25 (logged)	b	−.039	−.032	−.038	.086	−.041
	t (SE)	−.694 (.056)	−.650 (.049)	−.707 (.054)	1.685 (.051)	−.723 (.057)
	p	.488	.516	.480	.092	.470
Officer initiated (officer initiated / population) ^ .2	b	2.396***	2.256***	1.482**	2.598***	1.501***
	t (SE)	9.486 (.253)	8.774 (.257)	3.269 (.454)	9.060 (.287)	4.852 (.309)
	p	.000	.000	.001	.000	.000
Disadvantage	b	.152***	.184***	.280***	.141**	.232***
	t (SE)	3.931 (.039)	4.108 (.045)	5.814 (.048)	3.185 (.044)	4.602 (.050)
	p	.000	.000	.000	.001	.000
Spatial effects (two-stage least squares regression)	b	.000	.000	−.004	−.001	.001
	t (SE)	−.622 (.000)	.490 (.000)	−.660 (.006)	−1.614 (.000)	.970 (.001)
	p	.534	.624	.509	.107	.332
Constant	b	−1.206**	−2.415***	−4.039***	−3.608***	−1.881**
	t (SE)	−3.285 (.367)	−5.455 (.443)	−4.466 (.904)	−8.766 (.412)	−3.084 (.610)
	p	.001	.000	.000	.000	.002
Ln alpha	b	−2.651***	−2.467***	−2.873***	−2.321***	−2.214***
	t (SE)	−20.621 (.129)	−22.549 (.109)	−6.101 (.471)	−16.566 (.140)	−15.456 (.143)
	p	.000	.000	.000	.000	.000
Log-likelihood		−1,044.429	−854.188	−457.030	−774.665	−707.157
Wald C²		1,026.070	978.307	505.274	699.517	674.750
Pseudo R²		.140	.156	.193	.149	.166
Model probability		.000	.000	.000	.000	.000
Alpha		.071	.085	.057	.098	.109
Deviance		165.071	167.578	212.090	167.150	178.870

Note: N = 164. UCR = Uniform Crime Reports.

p < .05. **p < .01. ***p < .001.

Official crime and officer-initiated incidents exhibit consistent positive relationships with all categories of calls, indicating that high-crime areas and areas with high police activity report more incidents of all types. No significant effects regarding percentage of males 15 to 25 emerge. The lack of significance associated with the measure of spatial dependence suggests that the distribution of calls across space may be attributed to the spatial distribution of the other variables included in the models.

Table 4 presents estimates from negative binomial regression models for change in calls between 2004 and 2005. As noted earlier, change is measured such that higher levels represent greater decreases whereas lower levels represent smaller decreases or slight increases subsequent to the reduction of reporting services. These data provide mixed support for Hypothesis 2, which predicted that disadvantaged neighborhoods would exhibit smaller decreases in reporting following program changes. For total incidents and disorder-related incidents, there are no neighborhood differences in the change in calls made to the police subsequent to the reduction of reporting services. Thus, either the neighborhoods were similarly affected by the reduction in program services, or such services were not being widely used in the first place. However, for threat/harassment, property, and violent incidents, neighborhood disadvantage is negatively related to change in calls, indicating that disadvantaged neighborhoods maintained contact with the police at levels equal to or higher than those of the previous year. Conversely, in 2005 more affluent neighborhoods tended to exhibit significantly greater decreases in calls related to threat/harassment, property, and violent incidents.

Table 4.

Negative Binomial Regression of Change in Calls-for-Service Counts for Categories With 2004 Calls for Service as Offset

		Calls for Service: 2005 – 2004
Variable		Citizen	Disorder	Threat	Property	Violent
UCR violent	b	−.092*	−.094*	−.133*	−.077	−.347***
Logged (robbery + aggravated assault + 1/population)	t (SE)	−2.138 (.043)	−2.155 (.043)	−2.385 (.056)	−1.527 (.050)	−4.407 (.079)
	p	.033	.031	.017	.127	.000
Percentage male 15 to 25 (logged)	b	.146	.069	.054	.071	.055
	t (SE)	1.659 (.088)	1.081 (.064)	.850 (.064)	1.556 (.046)	.791 (.070)
	p	.097	.280	.395	.120	.429
Officer initiated (Officer initiated / population) ^ .2	b	−1.265***	−1.285***	−.556	−1.250***	−1.653***
	t (SE)	−4.479 (.282)	−4.304 (.299)	−1.703 (.326)	−3.921 (.319)	−4.675 (.354)
	p	.000	.000	.089	.000	.000
Disadvantage	b	−.038	−.053	−0.130**	−.132***	−.367***
	t (SE)	−.692 (.054)	−.830 (.063)	−2.624 (.050)	−3.489 (.038)	−6.558 (.056)
	p	.489	.407	.009	.000	.000
Population (logged)	b	−.301***	−.368***	−.468***	−.476***	−1.121***
	t (SE)	−3.946 (.076)	−4.297 (.086)	−6.334 (.074)	−6.292 (.076)	−8.393 (.134)
	p	.000	.000	.000	.000	.000
Spatial effects (two-stage least squares regression)	b	−.058	−.098	−.151*	−.025	−.091
	t (SE)	−1.021 (.057)	−1.709 (.057)	−2.490 (.061)	−.423 (.060)	−1.386 (.066)
	p	.307	.087	.013	.672	.166
Constant	b	1.177*	2.205***	2.682***	3.136***	7.761***
	t (SE)	2.295 (.513)	4.424 (.498)	4.243 (.632)	6.043 (.519)	7.132 (1.088)
	p	.022	.000	.000	.000	.000
Ln alpha	b	−1.850***	−2.204***	−2.872***	−2.029***	−1.647***
	t (SE)	−6.236 (.297)	−12.261 (.180)	−6.398 (.449)	−8.124 (.250)	−7.396 (.223)
	p	.000	.000	.000	.000	.000
Log-likelihood		−958.463	−805.225	−494.056	−751.810	−845.869
Wald C²		198.120	276.583	135.454	203.270	529.227
Pseudo R²		.044	.079	.119	.077	.146
Model probability		.000	.000	.000	.000	.000
Alpha		.157	.110	.057	.131	.193
Deviance		180.548	182.023	197.615	181.288	180.386

Note: N = 164.

p < .05. **p < .01. ***p < .001.

To illustrate the spatial distribution of change assessed in Table 4, we present several maps (refer to Figures 1–3 showing the contiguity of disadvantage and high-change areas for selected call types: threat/harassment, property, and violent incidents. Each map is based on a Moran scatterplot, in which neighborhood disadvantage is located on the x-axis and w × Y (the spatial lag term) is located on the y-axis (see Anselin, 2005). Consistent with the procedures of Morenoff, Sampson, and Raudenbush (2001), our average levels of disadvantage and change in calls were the cut point for high and low levels of the measured variable. For neighborhood disadvantage, there are four possible categories: low disadvantage, low contiguity to disadvantage; low disadvantage, high contiguity to disadvantage; high disadvantage, low contiguity to disadvantage; high disadvantage, high contiguity to disadvantage. Although four categories were developed for change in calls, we present only those two providing the strongest contrast: low decrease, low contiguity to high decrease; high decrease, high contiguity to high decrease.

Figure 1.

Spatial Typology of Disadvantage and Change in Threat Calls

Figure 2.

Spatial Typology of Disadvantage and Change in Property Calls

Figure 3.

Spatial Typology of Disadvantage and Change in Violent Calls

Again, contrary to Hypothesis 2, the pattern shown in each map suggests a strong relationship between neighborhood disadvantage and reliance on the police, as indicated by continued tendency to initiate contact with the police following the elimination of an important reporting mechanism. Whereas neighborhoods with above-average disadvantage and high contiguity to such areas were relatively stable in their levels of reporting between 2004 and 2005, calls for police service were more likely to decrease in neighborhoods that were of below-average disadvantage and that were in close proximity to similarly situated areas.

Discussion and Conclusion

Contemporary theoretical perspectives suggest that residents of disadvantaged neighborhoods are unlikely to contact the police for assistance or report crimes because of limited access to the law (Black, 1976) or because of a generalized distrust of the criminal justice system (Anderson, 1999; see also Baumer, 2002; Kubrin & Weitzer, 2003). The purpose of this study was to examine ecological variations in citizen reliance on the police using calls-for-service data from a midsize city in the Pacific Northwest. Because most prior studies have been based on victimization data, it is unclear whether the persistent finding of minimal effects reflects general neighborhood patterns and extends beyond a limited subset of crime incidents. Contrary to these studies, our findings (based on calls-for-service data) reveal clear neighborhood variations, with disadvantage associated with significantly higher levels of reliance on the police. The largest neighborhood differences were observed for calls related to violence and threats/harassment, the two call categories expected to involve the least amount of caller discretion.

Klinger and Bridges (1997) argue that calls for service data

are more likely to undercount the total number of incidents that come to the attention of the police in neighborhoods where residents believe that officers respond more slowly to their calls, where residents are more fearful of crime, and where they experience more criminal victimizations. (p. 720)

However, our findings suggest that at least in the city examined, disadvantaged populations have not totally given up on the police and their services. A nuanced interpretation of Black’s Behavior of Law (1976) may be beneficial in explaining these findings (see Bernard, 2002). Black contends that there are four important forms of law that exist in society: conciliatory, therapeutic, compensatory, and penal. Contemporary perspectives have focused almost wholly on penal law, often presuming that mistrust and ineffectiveness of such law correspond to the emergence of informal law in the form of street justice. However, if disadvantaged neighborhoods experience a deficit in all other forms of law identified by Black, then the importance of and reliance on penal law may be enhanced in these areas (see also Conklin, 1975; Laub, 1981; Manning & Van Mannen, 1978). Put simply, whereas affluent areas have access to compensatory, therapeutic, and conciliatory resolutions when faced with reduced legal resources, disadvantaged areas are frequently left with only penal law.

This gives rise to an interesting paradox: On one hand, evidence supports contemporary perspectives and interpretations that residents of disadvantaged neighborhoods are leery of police. On the other, extensive utilization of police services by residents of disadvantaged neighborhoods suggests that the police may be their primary, if not only, recourse. As an officer once recounted to one of the authors of this article, “Where else can you get access to a 24-hr social service?” Thus, in situations of high immediacy, such as those involving suspicious persons or potential crimes in progress, members of disadvantaged populations may have few alternatives to “calling the cops.” In contrast, more affluent neighborhoods tend to be characterized by established social networks and a much broader range of informal and nonpenal mechanisms for resolving such issues (e.g., calling one’s neighbors, private security, etc.).

Carr and colleagues’ (2007) analysis of narrative data from young men and women in three high-crime neighborhoods—one predominantly African American, one predominantly Latino, and one predominantly White—provides support for this position. Despite holding negative attitudes toward the police and frequently expressing reluctance to call the police for assistance, these youth tended to point to tougher and more law enforcement as the best strategy for reducing crime in their neighborhoods. The authors concluded that “the inconsistency of the respondents between disavowing police and wanting to see more of them to solve crime” (p. 467) contradicts theoretical claims regarding the existence in disadvantaged neighborhoods of a subcultural response characterized by “complete opposition to the conventional order” (p. 468). Instead, the authors argue, distrust of police is transitory and context dependent, incapable of being generalized to all police officers in all situations or to all residents of disadvantaged neighborhoods.¹² Our findings are consistent with this contention, showing that disadvantaged populations may rely on police more for certain types of incidents than for others. Wilkinson’s (2007) study of two inner-city neighborhoods in New York suggest that willingness to call on the police to intervene in local problems depends on the age of the suspects and their relationship with neighborhood residents (insider versus outsider).

Although we have emphasized the need to employ alternate sources of data as a check on studies using aggregated reports made by respondents of victimization surveys, calls-for-service data are not without limitation. Klinger and Bridges (1997) note that such data are susceptible to a variety of errors, including duplicate calls, inaccuracy in callers’ descriptions of incidents, and clerical and interpretive mistakes on the part of the call takers. Fortunately, duplicate calls made to police in the city examined here were generally eliminated by call takers at dispatch; furthermore, other errors common to calls-for-service data are likely to be widely dispersed and randomly distributed across neighborhoods.

A potentially more serious limitation involves our use of official statistics to measure objective levels of crime. Although we attempt to control for actual rates of crime using Uniform Crime Report data for categories of crime that are less amenable to reporting biases and differences—including aggravated assault and robbery (and, alternately, vehicle theft)—the discrepancy between true crime rates and officially recorded crime may be large and exhibit systematic variation across differing neighborhood contexts. Consequently, it is possible that we are able to establish only whether neighborhoods differ in absolute levels of reliance on the police, not whether variation in such reliance is more or less proportionate to crime in these places.

In addition, our data do not permit causal inference. Even when approximating a quasi-experimental posttest-only design (Hypothesis 2), explanation is limited to variation in reporting based on population characteristics rather than on rigorously controlled experimental conditions. Given these limitations, subsequent research on reporting might seek to more systematically assess how changes in reporting services act when appropriate control groups are established over time. This may be done by coordinating research efforts with local law enforcement agencies. Qualitative data may also be of value in identifying the specific processes leading people to call the police or not.

Another significant limitation of the present data is the inability to generalize results to populations characterized by greater levels of heterogeneity and disadvantage. The city assessed in this study is predominantly White, and although pocketed disadvantage does exist, it is not of the sort that one would find in highly urbanized inner-city environments, such as the one described by Anderson (1999). Whereas the maximum level of disadvantage in the city we examined is 2.4 standard deviations above the mean, other studies report levels as high as 7.0 or 8.0 standard deviations above the mean. Given that past studies find that failure to report is especially pronounced at the highest levels of neighborhood disadvantage (Baumer, 2002; Goudriann et al., 2006), caution must be exercised in generalizing our findings.

Future research should consider such contextual differences as well as the possibility that reporting biases are limited to certain segments of the population within extremely disadvantaged neighborhoods. Future efforts should also examine relationships between neighborhood resources and the ways in which surpluses and deficits of law are manifest. As Anderson (1999) notes, most people who reside in areas of extreme disadvantage are fundamentally decent, and there is limited reason to believe that such individuals would rule out calling the police or would feel compelled to impose street justice. To the contrary, given a general deficit in other resources, calling the police is likely to be a primary means by which residents of disadvantaged neighborhoods deal with their problems.

Footnotes

Appendix

Appendix:

Call Type per Category

Social Disorder
ABAN	Abandoned vehicle	KEG	Kegger	SUSCIR	Suspicious circumstances
ANIMAL	Animal control issue	LANDIS	Landlord tenant dispute	SUSPER	Suspicious person
ARGUE	Argument	LEWD	Lewd conduct	SUSPW	Suspicious person with weapon
BARK	Barking dogs	LIQVIO	Liquor law violation	SUSVEH	Suspicious vehicle
CUSINT	Custodial interference	MALMS	Malicious mischief	TRBLU	Trouble unknown
DEMPRO	Demonstration/protest	MALMSP	Malicious mischief	TRESP	Trespassing
DETOX	Detox van	MENTAL	Mental problem	TRUANT	Truant
DISORD	Disorderly conduct	NEIGHB	Neighborhood dispute	UGUEST	Unwanted guest
DOMST	Domestic disturbance	NOISE	Noise disturbance	VEHTAG	Junk vehicle tagged
DRUGS	Drug activity	OD	Overdose	WELFAR	Check on welfare of individual
DUMP	Illegal dumping	PANHAN	Panhandling
EVICT	Eviction	PARDIS	Parental discipline
GAMB	Gambling violation	PARTY	Party disturbance
GANG	Gang activity	PERBOT	Person bothering
GRAF	Graffiti	PROST	Prostitution
HULK	Junk vehicle	REPO	Repossession
ILSHOT	Illegal shots	SHOOTV	Shooting violation
IMPND	Vehicle impound	SHOTU	Shots unknown
Threats/Harrassment
HARASS	Harrassment
THREAT	Threat
Violence
ABDUC	Abduction	DVR	DV report	ROBA	Robbery–Armed
ABDUCP	Abduction in progress	DVW	DV with weapon	ROBB	Robbery–Bank
ABUSE	Domestic abuse	EXTOR	Extortion	ROBC	Robbery–Commercial
ABUSEC	Domestic abuse of child	FIGHT	Fight	ROBP	Robbery in progress
ASLT	Assault	FIGHTW	Fight with weapon	ROBS	Robbery–Strongarm
ASLTJ	Assault involving juvenile	HOSTAG	Hostage	SEXCRI	Sex crime
ASLTP	Assault in progress	INTW	Intimidation–weapon	SHOOT	Shooting
ASLTR	Assault report	PERW	Person with weapon	SNIPER	Sniper
CARJAC	Carjacking	RAPE	Rape	STAB	Stabbing
DOAH	Homicide	RAPEP	Rape in progress	STALK	Stalking
DRIVBY	Drive-by shooting	RIOT	Riot	UNLIMP	Unlawful Imprisonment
DV	Domestic violence	ROB	Robbery
Property
ARSON	Arson	SHOPL	Shoplifting
BURG	Burglary	SHOPLR	Shoplifting report
BURGC	Burglary–commercial	THEFT	Theft
BURGG	Burglary–garage	THEFTP	Theft in progress
BURGJ	Burglary–juvenile	VEHPRO	Vehicle prowling
BURGP	Burglary in progress	VEHPRP	Vehicle prowling in progress
BURGR	Burglary–residential	VEHREC	Vehicle theft recovery
BURGRC	Burglary–commercial	VEHTFT	Vehicle theft
BURGRG	Burglary–garage	VEHTP	Vehicle theft in progress
BURGRR	Burglary–residential
FORGER	Forgery
FRAUD	Fraud

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

The authors received no financial support for the research and/or authorship of this article.

Notes

References

Anderson

(1992). The story of John Turner. Public Interest, 108, 3-34.

Anderson

(1994). The code of the streets. Atlantic Monthly, 273(5), 80-94.

Anderson

(1999). Code of the street: Decency, violence, and the moral life of the inner city. New York: Norton.

Anselin

(2005). Exploring spatial data with GeoDa: A workbook. Urbana, IL: Center for Spatially Integrated Social Science.

Avakame

E. F.

Fyfe

J. J.

McCoy

(1999). Did you call the police? What did they do? Justice Quarterly, 16(4), 765-792.

Baller

R. D.

Anselin

Messner

S. F.

Deane

Hawkins

D. F.

(2001). Structural covariates of U.S. county homicide rates: Incorporating special effects. Criminology, 39(3), 561-590.

Baumer

E. P.

(2002). Neighborhood disadvantage and police notification by victims of violence. Criminology, 40(3), 579-616.

Baumer

Horney

Felson

Lauritsen

J. L.

(2003). Neighborhood disadvantage and the nature of violence. Criminology, 41(1), 39-71.

Beckett

Nyrop

Pfingst

(2006). Race, drugs and policing: Understanding disparities in drug delivery arrests. Criminology, 44, 105-138.

10.

Bennett

R. R.

Wiegand

R. B.

(1994). Observations on crime reporting in a developing nation. Criminology, 32(1), 135-148.

11.

Bernard

T. J.

(2002). Donald Black’s influence in criminology. Contemporary Sociology, 31(6), 650-652.

12.

Black

D. J.

(1976). The behavior of law. New York: Academic Press.

13.

Black

D. J.

(1983). Crime as social control. American Sociological Review, 48(1), 34-45.

14.

Black

D. J.

(1989). Sociological justice. New York: Oxford University Press.

15.

Browning

S. L.

Cullen

F. T.

Cao

Kopache

Stevenson

T. J.

(1994). Race and getting hassled by the police: A research note. Police Studies, 17, 1-11.

16.

Brunson

R. K.

Miller

(2006). Young Black men and urban policing in the United States. British Journal of Criminology, 46(4), 613-640.

17.

Bursik

R. J.

Jr. (1989). Political decision making and ecological models of delinquency: Conflict and consensus. In Messner

S. F.

Krohn

M. D.

Liska

A. E.

(Eds.), Theoretical integration in the study of deviance and crime: Problems and prospects (pp. 105-117). Albany: State University of New York Press.

18.

Bursik

R. J.

Jr. Grasmick

H. G.

(1993). Neighborhoods and crime: The dimensions of effective community control. New York: Lexington Books.

19.

Carbone-Lopéz

K. C.

(2007). The “usual suspects”: How race affects decisions to report rape victimization. Journal of Ethnicity in Criminal Justice, 3(4), 29-47.

20.

Carr

P. J.

Napolitano

Keating

(2007). We never call the cops and here is why: A qualitative examination of legal cynicism in three Philadelphia neighborhoods. Criminology, 45(2), 445-480.

21.

Conklin

J. E.

(1975). The impact of crime. New York: Macmillan.

22.

Fagan

Davies

(2000). Street stops and broken windows: Terry, race and disorder in New York City. Fordham Urban Law Journal, 28, 457-504.

23.

Fishman

(1979). Patterns of victimisation and notification. British Journal of Criminology, 19, 146-157.

24.

Goldstein

(1990). Problem-oriented policing. Philadelphia: Temple University Press.

25.

Gottfredson

M. R.

Hindelang

M. J.

(1979). A study of the behavior of law. American Sociological Review, 44(1), 3-18.

26.

Goudriaan

Wittebrood

Nieuwbeerta

(2006). Neighbourhood characteristics and reporting crime. British Journal of Criminology, 46(4), 719-742.

27.

Hardin

J. W.

(2002). The robust variance estimator for two-stage models. The Stata Journal, 2, 253-268.

28.

Hilbe

J. M.

(2007). Negative binomial regression. Cambridge, UK: Cambridge University Press.

29.

Jackson

A. L.

Boyd

L. M.

(2005). Minority-threat hypothesis and the workload hypothesis: A community-level examination of lenient policing in high crime communities. Criminal Justice Studies, 18(1), 29-50.

30.

Jones-Brown

D. D.

(2000). Debunking the myth of officer friendly: How African American males experience community policing. Journal of Contemporary Criminal Justice, 16, 209-229.

31.

Kane

R. J.

(2002). The social ecology of police misconduct. Criminology, 40(4), 867-896.

32.

Kane

R. J.

(2005). Compromised police legitimacy as a predictor of violent crime in structurally disadvantaged communities. Criminology, 43(2), 469-498.

33.

Klinger

D. A.

(1997). Negotiating order in patrol work: An ecological theory of police response to deviance. Criminology, 35, 277-306.

34.

Klinger

D. A.

Bridges

G. S.

(1997). Measurement error in calls-for-service as an indicator of crime. Criminology, 35(4), 705-726.

35.

Kubrin

C. E.

Weitzer

(2003). Retaliatory homicide: Concentrated disadvantage and neighborhood culture. Social Problems, 50(2), 157-180.

36.

Land

Deane

(1992). On the large-sample estimation of regression models with spatial- or network-effects terms: A two-stage least squares approach. Sociological Methodology, 22, 221-248.

37.

Laub

J. H.

(1981). Ecological considerations in victim reporting to the police. Journal of Criminal Justice, 9, 419-430.

38.

Manning

P. K.

Van Mannen

(1978). Policing: A view from the street. Santa Monica, CA: Goodyear.

39.

Mastrofski

S. D.

Reisig

M. D.

McCluskey

J. D.

(2002). Police disrespect toward the public: An encounter-based analysis. Criminology, 40(3), 515-551.

40.

Morenoff

J. D.

Sampson

R. J.

Raudenbush

S. W.

(2001). Neighborhood inequality, collective efficacy, and the spatial dynamics of urban violence. Criminology, 39, 517-559.

41.

O’Brien

R. M.

(1985). Crime and victimization data. Beverly Hills, CA: Sage.

42.

Osgood

D. W.

(2000). Poisson-based regression analysis of aggregate crime rates. Journal of Quantitative Criminology, 16(1), 21-43.

43.

Reisig

M. D.

Parks

R. B.

(2000). Experience, quality of life, and neighborhood context: A hierarchical analysis of satisfaction with police. Justice Quarterly, 17(3), 607-630.

44.

Reisig

M. D.

Parks

R. B.

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

45.

Sampson

R. J.

Bartusch

D. J.

(1998). Legal cynicism and (subcultural?) tolerance of deviance: The neighborhood context of racial differences. Law & Society Review, 32(4), 777-804.

46.

Sampson

R. J.

Groves

W. B.

(1989). Community structure and crime: Testing social-disorganization theory. American Journal of Sociology, 94(4), 774-802.

47.

Schuck

A. M.

Rosenbaum

D. P.

(2005). Global and neighborhood attitudes toward the police: Differentiation by race, ethnicity and type of contact. Journal of Quantitative Criminology, 21(4), 391-418.

48.

Scott

J. D.

(2002). Assessing the relationship between police–community coproduction and neighborhood-level social capital. Journal of Contemporary Criminal Justice, 18(2), 147-166.

49.

Skogan

W. G.

(1984). Reporting crimes to the police: The status of world research. Journal of Research in Crime and Delinquency, 21, 113-137.

50.

Skogan

W. G.

(1989). Communities, crime, and neighborhood organization. Crime & Delinquency, 35(3), 437-457.

51.

Smith

B. W.

Holmes

M. D.

(2003). Community accountability, minority threat, and police brutality: An examination of civil rights criminal complaints. Criminology, 41, 1035-1064.

52.

Smith

D. A.

Klein

J. R.

(1984). Police control of interpersonal disputes. Social Problems, 31, 68-481.

53.

Smith

D. A.

Visher

C. A.

(1981). Street-level justice: Situational determinants of police arrest decisions. Social Problems, 29, 167-177.

54.

Terrill

Paoline

E. A.

Manning

P. K.

(2003). Police culture and coercion. Criminology, 41, 1003-1034.

55.

Terrill

Reisig

M. D.

(2003). Neighborhood context and police use of force. Journal of Research in Crime and Delinquency, 40, 291-321.

56.

Triplett

R. A.

Sun

I. Y.

Gainey

R. R.

(2005). Social disorganization and the ability and willingness to enact control: A preliminary test. Western Criminology Review, 6(1), 89-103.

57.

Vélez

M. B.

(2001). The role of public social control in urban neighborhoods: A multi-level analysis of victimization risk. Criminology, 39(4), 837-864.

58.

Warner

B. D.

(1992). The reporting of crime: A missing link in conflict theory. In Liska

A. E.

(Ed.), Social threat and social control (pp. 71-87). Albany: State University of New York Press.

59.

Warner

B. D.

(1997). Community characteristics and the recording of crime: Police recording of citizens’ complaints of burglary and assault. Justice Quarterly, 14(4), 631-650.

60.

Warner

B. D.

(2007). Directly intervene or call the authorities? A study of forms of neighborhood social control within a social disorganization framework. Criminology, 45(1), 99-129.

61.

Warner

B. D.

Pierce

G. L.

(1993). Re-examining social disorganization theory using calls to the police as a measure of crime. Criminology, 31(4), 493-517.

62.

Webb

V. J.

Marshall

(1995). The relative importance of race and ethnicity on citizen attitudes toward the police. American Journal of Police, 14(2), 45-66.

63.

Weitzer

(1999). Citizens’ perceptions of police misconduct: Race and neighborhood context. Justice Quarterly, 16(4), 819-846.

64.

Weitzer

(2000). Racialized policing: Residents’ perceptions in three neighborhoods. Law and Society Review, 34(1), 129-155.

65.

Weitzer

Tuch

S. A.

(1999). Race, class and perceptions of discrimination by the police. Crime & Delinquency, 45, 494-507.

66.

Weitzer

Tuch

S. A.

(2006). Race and policing in America: Conflict and reform. Cambridge, UK: Cambridge University Press.

67.

Wells

Shafer

J. A.

Varano

S. P.

Bynum

T. S.

(2006). Neighborhood residents’ production of order: The effects of collective efficacy on responses to neighborhood problems. Crime & Delinquency, 52, 523-550.

68.

Wilkinson

D. L.

(2007). Local social ties and willingness to intervene: Textured views among violent urban youth of neighborhood social control dynamics and situations. Justice Quarterly, 24, 185-220.

69.

Wilson

J. Q.

Kelling

G. L.

(1982). Broken windows. Atlantic, 249(3), 29.

70.

Worden

R. E.

(1996). The causes of police brutality: Theory and evidence on police use of force. In Geller

W. A.

Toch

(Eds.), Police violence: Understanding and controlling police abuse of force (pp. 23-51). New Haven, CT: Yale University Press.