Property Crime Clearance in Small Jurisdictions

Abstract

Research on the factors that influence crime clearance rates has primarily studied violent crimes in large cities. However, property crimes are among the most commonly occurring and least frequently cleared offenses, and the majority of police departments in the United States serve small jurisdictions. Thus, this study undertook an examination of the predictors of clearance rates for burglary, larceny, and vehicle theft in a sample of agencies serving populations of 50,000 people or fewer. Independent variables included both policing factors (e.g., workload, funding, broken windows arrests) and social disorganization indicators (e.g., residential instability, poverty). Negative binomial regression analyses revealed variation in the significance of the predictors across the three crimes. Additionally, many predictors found to be influential in prior work were insignificant in this study, which suggests differences in the nature of crime clearance between large cities and smaller jurisdictions and a need for further research in this area.

Keywords

clearance rates burglary larceny vehicle theft

In the United States, a crime is considered cleared when an offender is arrested, charged, and delivered to the courts for prosecution or when there is sufficient evidence to arrest a person but unusual circumstances (e.g., death of the suspect) prevent that from occurring (U.S. Department of Justice, 2016). The percentage of crimes that are cleared—the clearance rate—is commonly used as an indicator of the effectiveness of law enforcement agencies by policy makers, the media, and criminologists (Pare & Ouimet, 2004; Worrall, 2016). Low clearance rates can have a variety of adverse effects, including decreased trust in law enforcement, greater fear of crime, poor morale among officers, and reduced deterrence (Lockwood, 2014; Roberts, 2008). Thus, researchers have done much to assess the incident details, policing factors, and community traits that affect clearance rates (Cordner, 1989; Doerner & Doerner, 2012; Jang, Hoover, & Lawton, 2008; Litwin, 2004; Litwin & Xu, 2007; Lockwood, 2014; Ousey & Lee, 2010; Puckett & Lundman, 2003; Roberts & Roberts, 2016; Worrall, 2016). However, the body of research on crime clearance is imbalanced in two ways. One, these studies have typically focused on violent crime, especially homicide (Doerner & Doerner, 2012; Taylor, Holleran, & Topalli, 2009). This is problematic, given the differences between violent and property offenses. Violent crimes are more likely to have witnesses because they require an encounter between victim and offender (Pare, Felson, & Ouimet, 2007). Additionally, they tend attract publicity and so may receive greater investigative attention (Pare et al., 2007; Puckett & Lundman, 2003). Unsurprisingly, then, violent crimes have higher clearance rates than property offenses. Over half of homicides and aggravated assaults are cleared in the United States, compared to less than one quarter of burglaries, larcenies, and vehicle thefts (U.S. Department of Justice, 2016). A second shortcoming of the clearance literature is that it has typically focused on agencies in large cities. Small and rural law departments differ in important ways from large urban ones, such as the availability of resources and the relationship with the community (Falcone, Wells, & Weisheit, 2002; Weisheit, Wells, & Falcone, 1995). Small departments also clear property crime at different rates. In 2015, departments serving populations smaller than 50,000 people cleared 15.3% of burglaries, 27.6% of larcenies, and 18.3% of vehicle thefts; those figures are 9.7%, 16.2%, and 8.8% for jurisdictions with populations over 250,000 (U.S. Department of Justice 2016). In light of these two weaknesses in the literature, the present study examined the predictors of property crime clearance rates using a sample of law enforcement agencies serving small populations.

Literature Review

Research has examined the predictors of crime clearance at the incident, neighborhood, and city level (e.g., Cihan, Zhang, & Hoover, 2012; Lockwood, 2014; Roberts & Roberts, 2016; Wong, 2010), and each approach offers unique advantages that are considered in detail later in this article. The reader should note that, given the different limitations of each type of research, caution is necessary when comparing findings across different levels of aggregation. Since this literature review precedes a city-level study, it focuses on other macro-level work and notes whether the studies of property crime examined cities or neighborhoods.

Community Factors

Scholars have known for decades that the social environment of a community can influence crime clearance (Bennett, 1982; Cordner, 1989), and social disorganization theory is one perspective that has been applied to that relationship (Mustaine, Tewksbury, Corzine, & Huff-Corzine, 2013; Regoeczi & Jarvis, 2013). This theory grew out of work at the University of Chicago in the early 20th century (Park, Burgess, & McKenzie, 1925/1984; Shaw & McKay, 1942). It contends that community traits (e.g., residential instability, racial diversity, etc.) permit crime to increase because they obstruct the development of the relationships among residents who provide a foundation for informal social control. The body of work testing these claims has since provided “strong empirical support” for this perspective (Pratt & Cullen, 2005, p. 405). The original formulation of the theory focused on crime in communities as has the bulk of research testing its claims. However, since some cities and counties experience more social disorganization than others, scholars have also tested the theory using larger aggregate units (Barnett & Mencken, 2002; Miethe, Hughes, & McDowall, 1991; Mosher, 2001). The possible macro-level effects of social disorganization variables have also been considered by clearance researchers, particularly at the city level where clearance rates are commonly reported (Borg & Parker, 2001; Jang et al., 2008; Ousey & Lee, 2010; Roberts, 2008; Roth, 2017). There are several reasons that social disorganization may be relevant to crime clearance. One, social disorganization is expected to increase crime and so may be related to police workload and the number of crimes to be cleared. Two, it may be more difficult to clear crimes in cities with much social disorganization if residents of disorganized areas are less cooperative with investigators due to fear of retaliation from other citizens or a general distrust of law enforcement (Roberts, 2008). Overall, there is mixed evidence about the impact of social disorganization on crime clearance and thus a need for additional research in this area (Lockwood, 2014).

Racial or ethnic diversity is one social disorganization variable that may affect clearance rates, particularly given that the poor relationship between police and minorities in the United States may limit reporting of crime or resident assistance with investigations (see Peck, 2015). However, this effect may vary by offense. Variables such as the portion of minorities or immigrants in the population are associated with lower homicide clearance in some studies (Keel, Jarvis, & Muirhead, 2009; Ousey & Lee, 2010; Puckett & Lundman, 2003), although they are insignificant in other work (Litwin & Xu, 2007). Regarding property crime, larger immigrant populations were associated with reductions in overall property crime clearance in Canadian municipalities, but the presence of aboriginal populations produced no effect (Wong, 2010). In a study of large American cities, larger non-White populations were negatively associated with the clearance of burglary, larceny, and vehicle theft (Roth, 2017). Conversely, in the 35 largest cities in Texas, larger Black populations increased clearance of larceny and vehicle theft but had no effect on burglary (Jang et al., 2008). Racial heterogeneity was also insignificant in a recent comparison of burglary clearance across Philadelphia neighborhoods (Lockwood, 2014).

Theoretically, residential stability should affect clearance rates by facilitating the development of relationships among residents that increase their ability to recognize crime and willingness to report it. Borg and Parker (2001) did find increased homicide clearance in large cities with greater residential stability. Similarly, Litwin (2004) found greater homicide clearance in Chicago neighborhoods with more home ownership. However, in a longitudinal analysis of Chicago data by Ousey and Lee (2010), residential instability did not affect homicide clearance. Roberts (2008) also found that residential stability did not affect robbery or assault clearance and had a negative relationship with rape clearance in a sample of mid-sized U.S. cities. Regarding property crime, Wong (2010) found an unexpected positive relationship between residential instability (percent of the population who had moved in the last year) and a summed measure of property crime clearance in Canadian cities. In the United States, Lockwood (2014) found that neighborhood-level residential stability did not affect burglary clearance in Philadelphia, while Roth (2017) found no effect on burglary, larceny, or vehicle theft clearance across large cities.

Economic status is another social disorganization variable that is often included in studies of clearance rates. However, variables such as unemployment and income frequently exhibit no association with homicide clearance across cities or neighborhoods (Davies, 2007; Litwin, 2004; Puckett & Lundman, 2003). Even when economic variables are significant, the direction of their effect varies. In Chicago, Litwin and Xu (2007) found a negative relationship between economic disadvantage and homicide clearance in one of the three time periods they studied, which Roberts (2008) also found for robbery and assault across cities. Conversely, Pare and Ouimet (2004) noted lower overall clearance rates in jurisdictions with higher incomes, and Ousey and Lee (2010) found a positive relationship between economic deprivation and homicide clearance in large cities. More relevant to the present study is the limited evidence regarding the relationship between economic conditions and property crime clearance. Lockwood’s (2014) work at the neighborhood level found no effect of economic disadvantage on burglary clearance. In Texas, Jang, Hoover, and Lawton (2008) found that unemployment reduced the clearance of robbery and assault but had no effect on clearance rates for burglary, larceny, and vehicle theft. Roth (2017) reached a similar conclusion regarding the effects of unemployment and poverty on those three offenses across large U.S. cities. Income and unemployment were also insignificant predictors of overall property crime clearance rates across Canadian municipalities (Wong, 2010), although a multilevel (incident and community) analysis of Quebecois communities did find a positive relationship between property crime clearance and the portion of low-income families (Pare et al., 2007).

Finally, some literature suggests a need to control for demographic factors. Larger total populations, for example, are often associated with reduced clearance rates for homicide, rape, robbery, assault, and overall property crime (Davies, 2007; Litwin & Xu, 2007; Pare & Ouimet, 2004; Pare et al., 2007; Wong, 2010). The effect of young people in a population is less clear. Pare and Ouimet (2004) suggested that departments serving populations containing more young people will have higher clearance rates in general. However, this effect may vary by crime type. Jang et al. (2008) found that the portion of young males in a population reduced robbery clearance, increased larceny clearance, and was insignificant for burglary, vehicle theft, and assault. In Roth (2017) the portion of juveniles in a population was associated with lower clearance rates for burglary and vehicle theft but had no effect on larceny clearance. Thus, additional research is needed to clarify the relationship between crime clearance and the presence of young people.

Policing Factors

It is reasonable to suspect that a police department’s workload and available resources affect its ability to clear crime, and prior work has tested the effects of several such variables. Bennett’s (1982) cross-national analysis, for example, found that the number of police personnel was negatively associated with overall clearance rates. However, Cordner (1989) did not find this effect among Maryland police departments. More recent analyses suggest a negative association between staffing levels and property crime clearance. Jang et al. (2008) found that the number of sworn personnel in 35 large Texas cities was negatively associated with the clearance of burglary, larceny, and vehicle theft. Similarly, in Doerner and Doerner’s (2012) analysis of several hundred Florida police departments, each standard deviation increase (about 120 officers) in sworn personnel was associated with a 2.6% decrease in the combined clearance rate for property crimes.

However, police staffing variables may be proxies for jurisdiction size, since departments serving large populations are likely to have more officers. Additionally, such variables do not account for possible differences in workload or resources among similarly sized departments. Thus, some studies have incorporated other measures such as the number of crimes per officer, the number of investigators per 1,000 crimes, or the number of officers per resident. However, these variables are frequently found to have no influence on overall clearance rates or those for violent crimes (Cordner, 1989; Davies, 2007; Ousey & Lee, 2010; Puckett & Lundman, 2003; Roberts, 2008). There is limited evidence regarding the relationship between police workload and property crime clearance. Pare, Felson, and Ouimet (2007) found no association between workload and property crime clearance in 116 departments in Quebec. Conversely, Jang et al.’s (2008) work in Texas concluded that the number of index crimes per sworn officer significantly reduced the clearance of burglary, larceny, and vehicle theft. Roth (2017) found a negative relationship between a similar workload measure and the clearance of burglary and larceny but not vehicle theft. Finally, a few studies have also included measures of law enforcement spending and have found it to be negatively associated with homicide and property crime clearance (Davies, 2007; Doerner & Doerner, 2012).

Several recent studies have also considered whether the use of broken windows policing can affect clearance rates. The concept of “broken windows” originated with an analogy in Wilson and Kelling’s (1982) article in Atlantic Monthly. In that article, they contend that just as a house where a broken window remains unrepaired invites vandalism, communities where disorderly behavior is tolerated will attract serious crime. Broken windows policing thus seeks to prevent or reduce crime through strict enforcement of disorder offenses such as loitering, public drunkenness, and vagrancy. There is mixed evidence about the effect of this strategy on crime (e.g., Corman & Mocan, 2005; Hinkle & Weisburd, 2008; Worrall, 2006). A recent meta-analysis found that it can produce a moderate decrease in offending (Braga, Welsh, & Schnell, 2015) although that effect may have partly resulted from reduced physical disorder. Jang et al. (2008) suggested three ways that broken windows policing could affect clearance rates. One possibility is that it will decrease clearance rates by reducing community cooperation and leaving fewer resources available for investigating serious offenses (Jang et al., 2008). Alternatively, it may increase crime clearance by giving officers opportunities to search offenders and obtain information that clears other crimes. For example, a person stopped for loitering might possess items stolen in a recent theft. Finally, broken windows policing may equally affect both the number of serious crimes and the arrest count, producing zero net effect on clearance rates (Jang et al., 2008).

Worrall (2006) and Jang et al. (2008) discuss the merits of several options for measuring broken windows policing at the city level. Ultimately, they operationalized this concept as the portion of misdemeanors or Part II arrests comprised by public order crimes such as vagrancy, disorderly conduct, and public drunkenness. This measure thus indicates an agency’s emphasis on enforcing disorder offenses relative to other minor crimes. However, it has produced inconsistent effects in the literature. Jang et al. (2008) found that across Texas cities, this variable was associated with increased burglary clearance and decreased larceny clearance but was insignificant for vehicle theft, robbery, and assault. Lockwood (2014) also found that it was associated with increased burglary clearance across Philadelphia neighborhoods. Conversely, broken windows policing was not associated with the clearance of burglary, larceny, or vehicle theft across large U.S. cities (Roth, 2017).

Policing in Small Towns and Rural Areas

An important feature of the clearance literature is that, like policing research in general, it has primarily focused on large urban departments (Falcone et al., 2002; Liederbach & Frank, 2003; Payne, Berg, & Sun, 2005; Rhodes & Johnson, 2008). However, the nationwide clearance rates mentioned in the introduction reveal lower clearance rates in large cities (U.S. Department of Justice 2016) and Cordner (1989) found differences in the clearance rates of urban and nonurban police departments in Maryland. Although Liederbach and Frank (2003) concluded that small town and urban police officers spend their time similarly when not interacting with citizens, studies do suggest several notable differences between policing in small or rural areas when compared to urban ones. (The Methods section discusses the difficulty of defining these communities.) It is thus unsafe to assume that studies of crime clearance in urban areas are generalizable to small or rural jurisdictions (Sampson, 1986).

Perhaps the greatest difference between urban and small towns is the strength of community relationships. Small town police officers tend to be more familiar with their residents (Liederbach & Frank, 2003; Payne et al., 2005), and citizens are more familiar with one another (Weisheit et al., 1995). Weisheit, Wells, and Falcone (1995, p. 26) described this as a “density of acquaintanceship” that can increase residents’ ability to notice crime and identify particular suspects, as well as their willingness to notify law enforcement. Similarly, Falcone, Wells, and Weisheit (2002, p. 377) noted that this familiarity between police and the community produces investigative leads that are “unheard of” in urban departments. These relationships may also contribute to the tendency of small and rural departments to rely on informal resolutions rather than criminal or legal ones (Falcone et al., 2002; Payne et al., 2005; Weisheit et al., 1995).

In addition to their community relationships, rural and small town police departments differ from their urban counterparts in several other ways. They tend to have fewer financial resources, perhaps due to due smaller tax bases (Falcone et al., 2002; Weisheit et al., 1995). Small town officers also receive fewer calls per shift, although that figure may underestimate their workload by omitting informal requests for aid that are more common in these settings (Payne et al., 2005; Rhodes & Johnson, 2008). Police activity in these departments also varies more widely, with officers providing a range of social services and acting as generalists rather than specializing in a particular type of activity (Falcone et al., 2002; Payne et al., 2005; Weisheit et al., 1995).

Summary

Some readers may be struck by the divergence in the findings of prior studies. However, these variations may be explained in part by methodological differences. For example, some studies use clearance rates for individual crimes as dependent variables (Jang et al., 2008; Lockwood, 2014; Roberts, 2008), while others examine categories of crime (e.g., the combined property crime clearance rate used by Doerner & Doerner, 2012; Pare et al., 2007; Wong, 2010). However, specific property offenses are cleared at different rates (e.g., theft is cleared at higher rates than burglary), and so the results of studies using combined outcome variables may reflect the mixture of crimes within a jurisdiction in addition to police success in clearing them (see Pare & Ouimet’s, 2004 discussion of expected clearance rates).

The choice of geographic units may also account for differences across studies, as each level of analysis presents unique advantages and drawbacks that can affect the results. Incident-level studies can examine the effects of offense traits, such as police response time, that are difficult for macro-level studies to address (e.g., Coupe & Girling, 2001). However, in some jurisdictions, the apparent effects of incident traits may actually reflect the influence of contextual variables (e.g., if victim race is a partial proxy for the residential stability of the community where the crime occurred). Some research resolves this issue by modeling neighborhood-level variables in addition to incident traits (Litwin & Xu, 2007; Lockwood, 2014) but may face questions about whether the findings can be generalized to other jurisdictions. That is, when a single city is studied, its endemic cultural and geographic features may affect the results. Multijurisdictional studies face a similar generalizability problem when they examine departments from a single state or province (e.g., Cordner, 1989 in Maryland; Doerner & Doerner, 2012 in Florida; Jang et al., 2008 in Texas; Pare et al., 2007 in Quebec), and their lack of lower level (incident and neighborhood) data may conceal the effects of incident traits or neighborhood context.

Ultimately, the crime clearance literature is uneven in at least two ways. One, it has disproportionately focused on violent crime, particularly homicide (Doerner & Doerner, 2012; Taylor et al., 2009). Furthermore, some research on property crime clearance has combined these offenses into overall measures that partly reflect differences in the distribution of crimes across offense type and cannot examine intercrime differences in clearance predictors. Two, scholars have devoted insufficient attention to crime clearance in small jurisdictions, despite evidence regarding the unique challenges and advantages of policing in those settings. The present research strengthens the literature in both of these areas by examining the predictors of clearance rates for burglary, larceny, and vehicle theft in a sample of law enforcement agencies serving small populations.

Methods

Data and Sample

This study sought to sample law enforcement agencies serving small populations in the United States. However, there is no consensus about the standard for distinguishing large and small police departments. Some studies have selected departments based on the number of officers (Gilbin, Burruss, & Schafer, 2014; Pare et al., 2007), while others have used the annual number of incidents (Pare & Ouimet, 2004) or the total population as criteria (Ousey & Lee, 2010). Roberts and Roberts (2016) focused on jurisdictions with populations larger than 50,000; the present study used the same standard but examined jurisdictions smaller than that figure. The sample consisted of 362 randomly selected law enforcement agencies serving populations of 50,000 people or fewer¹ who responded to the 2013 Law Enforcement Management and Statistics (LEMAS) survey conducted by the Bureau of Justice Statistics (U.S. Department of Justice, 2014a).² Four tribal agencies were removed due to the unique nature of law enforcement on reservations, where a “complicated web” of jurisdictional rules based on offense seriousness, crime location, and offender and victim ethnicity complicate law enforcement and contribute to confusion about who residents should contact to report crime (Goldberg & Champagne, 2005; Reaves, 2011; Wilkins & Lomawaima, 2001, p. 5). Thus, the final sample contained 358 agencies. The 2013 Uniform Crime Report (UCR) supplied information regarding clearance rates (U.S. Department of Justice, 2014b), while the 2013 American Community Survey 5-year estimates provided data for the structural variables (U.S. Census Bureau, 2014).

Dependent Variables

The dependent variables in this study were the clearance rates (given as a percentage) for burglary, larceny, and vehicle theft as indicated by the 2013 UCR (U.S. Department of Justice, 2014b). These crimes were chosen because they are among the most commonly occurring but least successfully cleared Part I crimes (U.S. Department of Justice, 2016) and because they have received insufficient attention in the clearance literature. The imperfections of clearance rates are noted in the limitations section, but they remain a standard measure of law enforcement performance (Doerner & Doener, 2012; Jang et al., 2008).

Independent Variables

Existing studies suggest several policing variables that may affect clearance rates. Police spending was operationalized by dividing the total agency budget by the total population to compute law enforcement spending per capita (as done by Doerner & Doerner, 2012). Police workload was measured as the number of Part I crimes per full-time officer (Jang et al., 2008). Additionally, the analysis used two variables to account for the presence of county sheriffs (n = 67) and municipal (i.e., noncounty) police agencies (n = 291) in the sample. Since county sheriffs often have larger territories than municipal ones, police density was included as the number of full time officers per square mile of the jurisdiction’s land area. To account for other unique traits of county sheriffs (Liederbach & Frank, 2006), a dummy variable was also included with noncounty agencies as the reference category. Finally, as noted in the literature review, the use of particular enforcement strategies may affect property crime clearance (Jang et al., 2008; Lockwood, 2014). One such strategy is broken windows policing which was measured as the percentage of all Part II arrests comprised by broken windows offenses (disorderly conduct, public drunkenness, vagrancy, loitering, and curfew violations) as indicated by the 2013 UCR (Jang et al., 2008; Worrall, 2006 used similar measures). This variable thus reflects agencies’ prioritization of broken windows offenses relative to other minor crimes.

The analyses also controlled for several structural variables. The presence of youth may affect crime clearance (Pare & Ouimet, 2004; Roth, 2017) and was included as the percentage of the total population comprising juveniles. In the United States, racial minorities often have less trust in police (see Peck, 2015), which may affect crime reporting and cooperation with investigations. Thus, the percentage of the population that was non-White was included as a predictor. Similarly, residential instability may decrease social connections among residents and hinder their ability or willingness to recognize and report crime; it was measured as the percentage of the city’s residents who had moved in the last year (Lockwood, 2014). Although unemployment may influence property crime by increasing the number of financially motivated offenders, unemployed people may also have more home-centric routines that permit them to notice and report crime, thus decreasing police response time and increasing their chances of catching the offender in the act (Cihan et al., 2012; Coupe & Girling, 2001). Because poorer areas may receive greater police attention, the percentage of residents living in poverty was also included as a predictor (Cihan et al., 2012; Pare et al., 2007). The analysis also controlled for differences in population size by including the total population (in thousands) as a variable. Finally, rurality was measured using the U.S. Department of Agriculture’s (United States Department of Agriculture [USDA], 2013) rural–urban continuum. This measure not only distinguishes between metropolitan and nonmetropolitan counties but also assigns a value to every county in the United States based on the size of its urban population and proximity to a large metropolitan area; possible scores range from 1 (highly urban) to 9 (highly rural).

Analysis

Table 1 contains descriptive statistics regarding the variables. Initial examination of the data revealed a need to address missing information. Specifically, there were 104 instances of missing data for the broken windows variable, 38 for policing spending, 10 for police workload, 4 for police density, and 2 for the remaining predictors with the exception of the county agency dummy and rurality variables, which were complete. Researchers seeking to address missing data have a variety of options (for an extensive review, see Graham, 2009). Recent work suggests that modern approaches such as multiple imputation are superior to older methods such as analyzing only cases with complete data or replacing missing values with the mean (Graham, 2009; Johnson & Young, 2011; Schlomer, Bauman, & Card, 2010). Thus, the analysis imputed missing values using the multiple imputation procedure in the Statistical Package for the Social Sciences (SPSS), version 24. The number of imputations was set to 40, which Graham, Olchowski, and Gilreath (2007) found to be suitable (loss of power <1%) for instances of up to 50% missing information. In the present study, the variable with the most missing data (broken windows enforcement) had 29% of cases missing.

Table 1.
Descriptive Statistics.

Measure Mean SD Min. Max.

Total burglaries 75.56 126.97 0.00 1,467.00

Burglary clearance % 13.83 15.78 0.00 100.00

Total larcenies 284.47 410.88 0.00 3,287.00

Larceny clearance % 21.98 19.73 0.00 100.00

Total vehicle thefts 19.49 39.78 0.00 291.00

Vehicle theft clearance % 23.47 28.48 0.00 100.00

Police spending Police spending per capita 247.15 220.06 19.88 2,402.99

Police workload Part I crimes per officer 11.97 9.02 0.00 52.44

Broken windows Portion of part II arrests for broken windows offenses 9.63 10.69 0.00 66.70

Police density Officers per square mile 4.09 10.43 0.00 120.00

Total population Total population of jurisdiction 14,687.81 13,235.73 281.00 56,965.00

Juveniles Portion of population <18 23.52 4.97 6.25 45.12

Non-White Portion of population non-White 19.34 19.30 0.00 99.56

Unemployment Unemployed portion of workforce 9.75 4.80 0.70 30.10

Poverty Portion of population in poverty 16.00 9.56 1.00 49.40

Residential instability Portion of population moved in last year 14.94 6.42 2.00 43.00

Rurality USDA urban–rural continuum score 3.56 2.57 1.00 9.00

Note. USDA = United States Department of Agriculture.

Subsequently, the clearance rates were analyzed using negative binomial regression, which was chosen for several reasons. One, although least squares regression methods are often used to analyze rates, they become inappropriate when those figures are computed based on a small number of crimes. Osgood (2000) provides a thorough review of the problems associated with applying least squares techniques to crime rates and recommends the use of Poisson or negative binomial techniques when analyzing units with populations less than several hundred thousand. The present study is one such case; the sampled agencies served small populations and some had low numbers of offenses and clearances. Two, negative binomial was chosen over Poisson regression because the data were overdispersed; that is, the variance was larger than the means for the dependent variables (burglary M = 13.83 and variance = 249.01; larceny M = 21.98 and variance = 389.18; vehicle theft M = 23.47 and variance = 811.15). Poisson regression assumes the equality of the mean and variance, and so negative binomial regression was more appropriate for the present analysis (Osgood, 2000). Three, zero-inflated negative binomial regression is necessary if there are excess zeros in a data set produced by the inability of some cases to experience the dependent variable. For example, it would be impossible for an agency that has no reported larcenies to have any cleared larcenies. Lacking information about which agencies had zero clearances for that reason, a zero-inflated approach might have been necessary. However, in the present study, the regression analyses for each offense excluded departments that had not experienced any incidents of that crime (as did Worrall, 2016), thus precluding the need for zero-inflated negative binomial models.

Table 2 contains the results of the negative binomial regressions. In each model, the dependent variable was the number of cleared offenses for that particular crime (which is log-transformed in negative binomial regression). The log of the number of reported instances of that crime (in hundreds) was included an offset (i.e., exposure) variable, which accounted for the number of clearances out of the number of reported crimes and effectively made the dependent variable a rate: clearances per hundred crimes (Osgood, 2000). The dispersion parameters for each of the three models were significantly greater than zero (burglary = .727, larceny = .964, vehicle theft = .695, p < .01), indicating overdispersion and supporting the use of negative binomial rather than Poisson models. Overall, the likelihood ratio χ²values indicate that the models for larceny and vehicle theft were a significant improvement over null models.

Table 2.
Negative Binomial Regressions of Property Crime Clearance Rates on Police and Structural Predictors.

Burglary (n = 318) Larceny (n = 327) Vehicle Theft (n = 291)

Coefficient SE IRR Coefficient SE IRR Coefficient SE IRR

County agency −0.123 .191 0.884 −0.348 .198 0.706 −0.022 .2288 0.978

Police spending 0.000 .000 1.000 0.000 .000 1.000 0.000 .0003 1.00

Police workload 0.001 .007 1.001 0.015* .008 1.015 −0.005 .0093 0.995

Broken windows −0.002 .007 0.998 −0.001 .007 0.999 0.003 .0081 1.003

Police density −0.003 .007 0.997 0.002 .006 1.002 −0.017* .0075 0.983

Total population −0.009 .005 0.991 −0.007 .005 0.993 −0.013* .0059 0.987

Juveniles −0.029* .012 0.971 −0.009 .012 0.991 −0.017 .0149 0.984

Racial diversity −0.003 .004 0.997 0.003 .004 1.003 −0.007 .0044 0.993

Unemployment 0.017 .016 1.017 0.012 .016 1.012 −0.001 .0201 0.999

Poverty 0.005 .010 1.005 −0.001 .010 0.999 0.021 .0118 1.021

Residential instability −0.009 .011 0.991 0.006 .011 1.007 −0.008 .0124 0.992

Rurality −0.023 .032 0.977 −0.011 .033 0.989 0.001 .0382 1.001

Constant 3.551 .389 34.911 3.062 .381 21.393 3.772** .4927 43.534

Likelihood ratio χ² 14.962 23.025* 24.217**

Note. IRR = incident rate ratio; SE = standard error.

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

For burglary, the presence of juveniles was the only significant variable. The incidence rate ratio (IRR) indicates that each unit (i.e., percentage point) increase in the portion of juveniles was associated with a 2.9% decrease in the burglary clearance rate. However, this variable was not significant for either of the other crimes. For larceny, a one-unit increase in workload (Part I crimes per officer) was associated with a 1.5% increase in the larceny clearance rate. Two variables had significant negative effects on vehicle theft clearance. Each unit increase in police density (officers per square mile) was associated with a 1.7% reduction in that clearance rate, while each thousand-person increase in total population reduced it by 1.3%.

The effects of the rurality variable were insignificant across all three crimes. In addition to the models presented in Table 2, which used the standard nine-point scale for that measure, other analyses explored two other versions of the rurality measure. One analysis converted the 9-point scale into a dummy variable for metropolitan counties with populations over one million, with all other counties (smaller metropolitan counties and nonmetropolitan counties) combined as the reference category. When this dummy variable replaced the 9-point scale in the analysis, the effect remained insignificant across all three crimes. Another analysis recoded the rurality measure into a dummy variable comparing metropolitan and nonmetropolitan counties. When the analysis used this variable instead of the original 9-point scale, the results were again insignificant for all crimes.

Discussion

In the case of burglary, none of the policing variables significantly predicted clearance rates. This finding suggests that burglary clearance in smaller jurisdictions is not the product of investigative attention or resources but of situational factors such as whether the burglar is noticed and reported in sufficient time to be caught in the act (Coupe & Girling, 2001). Such a conclusion is consistent with early work indicating that nonpolicing factors such as the availability of evidence and witnesses were the strongest influences on property crime clearance (Eck, 1983; Greenwood, Chaiken, Petersilia, & Prusoff, 1975; Greenwood & Petersilia, 1975). However, the insignificance of the policing variables contrasts prior city-level research that found negative associations between burglary clearance, total officers, and police workload (Jang et al., 2008; Roth, 2017). The only variable that influenced burglary clearance in this study was the presence of juveniles, which had a negative effect. As noted in the literature review, there is mixed evidence on this point in earlier studies. Roth, 2017 found a negative effect from juvenile presence in large U.S. cities, although Jang et al. (2008) found no effect from the presence of young males. One possibility is that compared to adults, juvenile commit burglary in ways that make clearance more difficult. Although the few studies of juvenile burglars indicate that they choose targets based on similar considerations as adults (Logie, Wright, & Decker, 1992; Wright & Logie, 1988), their offenses may differ in other ways, such as the amount of time spent in the home or methods of disposing of stolen items.

In contrast to burglary, one policing variable was significantly associated with larceny clearance. Police workload (Part I crimes per full-time officer) was positively associated with larceny clearance. This finding runs counter to the expectation that a greater workload leaves less time and resources for solving any one offense (Bennett, 1982). Moreover, the positive effect of workload contrasts the negative relationship between workload and larceny clearance found by Jang et al. (2008) and Roth, 2017. Future research would benefit from further exploration of the connection between workload and clearance rates in small jurisdictions.

For vehicle theft, two variables were significant that were insignificant for the other crimes. Police density (officers per square mile) was negatively associated with vehicle theft clearance. Police density is affected by two factors: the number of officers (the numerator) and the agency’s land area (the denominator). Thus, vehicle theft was cleared at lower rates in geographically smaller jurisdictions with larger police departments. Total population had a similar negative relationship with vehicle theft clearance. Taken together, these effects suggest lower clearance rates among the more urban departments, which is consistent with the overall pattern of UCR data described in the introduction. These results also align with prior studies regarding the connection between other indicators of urbanity (e.g., number of sworn officers and police spending) and auto theft clearance (Jang et al., 2008; Roth, 2017). However, as noted in the analysis, the USDA rural–urban continuum had no effect despite the exploration of several forms of this variable.

Social disorganization theory was originally proposed as an explanation for crime rates, and prior work suggested the relevance of its structural variables to crime clearance. The insignificance of most of these predictors may cause some to question theory’s relevance crime in small towns and rural areas. Some research does indicate the applicability of the theory to rural places (Li, 2011; Osgood & Chambers, 2000), although other work is much less favorable (Kaylen & Pridemore, 2011, 2013). The findings of the present study could be viewed as evidence of the irrelevance of social disorganization to small town crime clearance. However, there are several other interpretations of the insignificant results, which may also explain the differences between the findings of prior research and the present study. One is that social disorganization has similar effects in urban centers, small towns, and rural areas but is associated with different structural variables in smaller communities, such as a decline in the presence of local small businesses (Kaylen & Pridemore, 2012). Alternatively, the conditions that produce social disorganization in urban areas may be mitigated by the unique aspects of small and rural communities (e.g., “density of acquaintanceship”) that were highlighted in the literature review (Chilenski, Syvertsen, & Greenberg, 2015; Holmes, Painter & Smith, 2017; Weisheit et al., 1995, p. 26; Wells & Weisheit, 2012). Another option suggested by Li (2011) is that in smaller places, exogenous factors such as regional economic conditions carry greater weight and overwhelm the effects of local structural variables. A final explanation for the insignificance of the social disorganization variables in this study is heterogeneity among small towns and rural communities (Li, 2011), which the present study sought to model with the police density, rurality, and county dummy variables. Small towns may be more cohesive than geographically dispersed rural communities and places where residents work and socialize primarily within their own town may be more cohesive than those, where many residents travel to surrounding areas for those activities (Kaylen & Pridemore, 2011). Overall, the lack of research on clearance rates in rural areas and small towns makes it difficult to assess the likelihood of these explanations; clearly, more work is needed to clarify the effects of structural variables on crime clearance outside of urban areas.

Several of the possibilities described above (e.g., greater influence of exogenous factors, greater social cohesion, different structural correlates of social disorganization) may also explain the disparate findings of the present study and prior work. That divergence could also result from other important differences between this study and prior research on property crime clearance, such as examining specific property crimes (in contrast to Doerner & Doerner, 2012; Pare et al., 2007; Wong, 2010) at the city level (in contrast to Lockwood, 2014) for a nationwide sample (in contrast to Doerner & Doerner, 2012; Jang et al., 2008) consisting only of departments serving small populations (in contrast to Roth, 2017).

Limitations

One limitation of this work concerns the weaknesses of clearance rates as an outcome measure. Patterns in citizens’ crime reporting can affect them, and agencies may manipulate the figures to their benefit (Cordner, 1989; Jang et al., 2008; Pare & Ouimet, 2004; Tilley & Burrows, 2005). In some cases, arrested offenders may be responsible for additional uncleared crimes for which they are never charged, resulting in artificially low clearance rates. Additionally, Jang et al. (2008) warned that focusing on crime clearance ignores the other diverse activities required for effective law enforcement. More recently, deterrence theorists have noted that clearance rates are a poor measure of offenders’ probability of apprehension and do not account for crimes that are prevented because of effective policing (Nagin, 2016; Nagin, Solow, & Lum, 2015; Pickett & Roche, 2016a, 2016b; Pogarsky & Loughran, 2016; Raphael, 2016). Although these flaws are well known among law enforcement and criminologists, clearance rates receive attention from policy makers, influence funding decisions, and are held up for public judgement in the media (Buduson, 2017; Jorgensen, 2017; Kaste, 2015; Moore, 2016; Palomino & Veklerov, 2017; Worrall, 2016). Thus, they “remain a key indicator police performance” (Worrall, 2016, p. 2). Since agencies continue to be held accountable for them, clearance rates (and other clearance-based metrics such as time to clearance) remain a common outcome measure in recent policing research (e.g., Carter & Carter, 2016; Doerner & Doerner, 2012; Lockwood, 2014; McCluskey, Cancino, Tillyer, & Tillyer, 2014; Pastia, Davies, & Wu, 2017; Roth, 2017; Stacey, Martin, & Brick, 2017; Worrall, 2016).

A second limitation concerns the UCR’s hierarchy rule and its potential effect on the broken windows variable. The hierarchy rule stipulates that when multiple offenses occur at the same time and place, only the most serious crime is recorded in the UCR (U.S. Department of Justice, 2004). For example, if a person is arrested for public drunkenness and then assaults the arresting officer, only the assault will appear in the UCR. Thus, as noted in other work using this measure (Roth, 2017), UCR data likely reflect incomplete counts of the minor crimes used to calculate the portion of Part II arrests comprised by broken windows offenses. It is difficult to estimate the severity of this limitation, since the range of values on the broken windows variable does suggest interagency differences in the enforcement of these offenses. It is possible that the hierarchy rule affects the actual counts of Part II offenses but not the ranking of agencies relative to one another. For example, if the hierarchy rule results in the exclusion of 10% of Part II offenses from the UCR, then the agencies that prioritized broken windows enforcement would still have higher values on the broken windows variable.

A final limitation was the difficulty of controlling for rurality. Although all of the sampled agencies served small populations, rural agencies face different challenges than urban ones, such as larger jurisdictions or fewer resources (Falcone et al., 2002; Weisheit et al., 1995). This study attempted to account for these differences with variables such as law enforcement spending, police density, and the USDA urban–rural continuum. However, even these measures are imperfect indicators that cannot completely account for rurality across small jurisdictions; for example, a town may be located in a rural portion of a county near a metropolitan area that scores near the urban end of the continuum.

Conclusion

There are ample routes for future research to expand upon the findings of this work. Subsequent studies might examine factors related to the clearance of violent crime in smaller jurisdictions. They might also explore other criteria for defining small departments or measuring rurality, as there is no consensus regarding ideal measures of these concepts (Falcone et al., 2002; Weisheit et al., 1995). For example, Gilbin, Burruss, and Schafer (2014) assessed geographic proximity to (and interaction with) large agencies as predictors of the homeland security preparedness of small town police. Similar spatial variables could be used in the analyses of crime clearance. Finally, since police–community collaboration is particularly important to small town law enforcement (Falcone et al., 2002; Weisheit et al., 1995), future research could attempt to measure whether the strength of that relationship produces any effect on clearance rates for these departments.

Ultimately, despite the significance of some predictors, the notable finding of this study was the collective absence of effects from most variables. This was true even of factors such as racial diversity and residential instability that are frequently related to clearance rates in prior research. In addition to the theoretical implications discussed above, these findings provide additional evidence against the assumption that policing functions similarly in large cities and smaller jurisdictions (Falcone et al., 2002; Liederbach & Frank, 2003; Payne et al., 2005; Rhodes & Johnson, 2008). The consequences of this assumption are not purely academic, as it can lead flawed attempts to replicate urban policing strategies in smaller jurisdictions (Falcone et al., 2002; Sampson, 1986). For instance, strict criminal enforcement of broken windows offenses may contravene a community’s preference that these minor crimes be resolved informally (Falcone et al., 2002; Liederbach & Frank, 2006; Weisheit et al., 1995). When misapplied, such tactics may harm the community relationships that are vital to effective small town law enforcement. Overall, this research bolsters the clearance literature by focusing on small jurisdictions and on specific property crimes, but additional work is needed on both of those fronts to refine the understanding of intercrime and community differences in clearance rates.

	Measure	Mean	SD	Min.	Max.
Total burglaries		75.56	126.97	0.00	1,467.00
Burglary clearance %		13.83	15.78	0.00	100.00
Total larcenies		284.47	410.88	0.00	3,287.00
Larceny clearance %		21.98	19.73	0.00	100.00
Total vehicle thefts		19.49	39.78	0.00	291.00
Vehicle theft clearance %		23.47	28.48	0.00	100.00
Police spending	Police spending per capita	247.15	220.06	19.88	2,402.99
Police workload	Part I crimes per officer	11.97	9.02	0.00	52.44
Broken windows	Portion of part II arrests for broken windows offenses	9.63	10.69	0.00	66.70
Police density	Officers per square mile	4.09	10.43	0.00	120.00
Total population	Total population of jurisdiction	14,687.81	13,235.73	281.00	56,965.00
Juveniles	Portion of population <18	23.52	4.97	6.25	45.12
Non-White	Portion of population non-White	19.34	19.30	0.00	99.56
Unemployment	Unemployed portion of workforce	9.75	4.80	0.70	30.10
Poverty	Portion of population in poverty	16.00	9.56	1.00	49.40
Residential instability	Portion of population moved in last year	14.94	6.42	2.00	43.00
Rurality	USDA urban–rural continuum score	3.56	2.57	1.00	9.00

	Burglary (n = 318)	Larceny (n = 327)	Vehicle Theft (n = 291)
County agency	−0.123	.191	0.884	−0.348	.198	0.706	−0.022	.2288	0.978
Police spending	0.000	.000	1.000	0.000	.000	1.000	0.000	.0003	1.00
Police workload	0.001	.007	1.001	0.015*	.008	1.015	−0.005	.0093	0.995
Broken windows	−0.002	.007	0.998	−0.001	.007	0.999	0.003	.0081	1.003
Police density	−0.003	.007	0.997	0.002	.006	1.002	−0.017*	.0075	0.983
Total population	−0.009	.005	0.991	−0.007	.005	0.993	−0.013*	.0059	0.987
Juveniles	−0.029*	.012	0.971	−0.009	.012	0.991	−0.017	.0149	0.984
Racial diversity	−0.003	.004	0.997	0.003	.004	1.003	−0.007	.0044	0.993
Unemployment	0.017	.016	1.017	0.012	.016	1.012	−0.001	.0201	0.999
Poverty	0.005	.010	1.005	−0.001	.010	0.999	0.021	.0118	1.021
Residential instability	−0.009	.011	0.991	0.006	.011	1.007	−0.008	.0124	0.992
Rurality	−0.023	.032	0.977	−0.011	.033	0.989	0.001	.0382	1.001
Constant	3.551**	.389	34.911	3.062**	.381	21.393	3.772**	.4927	43.534
Likelihood ratio χ²	14.962			23.025*			24.217**

Footnotes

Declaration of Conflicting Interests

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

Funding

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

Notes

References

Barnett

Mencken

F. C.

(2002). Social disorganization and the contextual nature of crime in nonmetropolitan counties. Rural Sociology, 67, 372–393.

Bennett

(1982). The effect of police personnel levels on crime clearance rates: A cross-national analysis. International Journal of Comparative and Applied Criminal Justice, 6, 177–193.

Borg

M. J.

Parker

K. F.

(2001). Mobilizing law in urban areas: The social structure of homicide clearance rates. Law and Society Review, 35, 435–466.

Braga

A. A.

Welsh

B. C.

Schnell

(2015). Can policing disorder reduce crime? A systematic review and meta-analysis. Journal of Research in Crime and Delinquency, 52, 567–588.

Buduson

(2017, 5 19). Waiting for justice: Cleveland police struggle to solve murder cases. News 5 Cleveland. Retrieved from http://www.news5cleveland.com/longform/cleveland-police-struggle-to-solve-murder-cases

Carter

D. L.

Carter

J. G.

(2016). Effective police homicide investigations: Evidence from seven cities with high clearance rates. Homicide Studies, 20, 150–176.

Chilenski

S. M.

Syvertsen

A. K.

Greenberg

M. T.

(2015). Understanding the link between social organization and crime in rural communities. Journal of Rural and Community Development, 10, 109–127.

Cihan

Zhang

Hoover

(2012). Police response time to in-progress burglary: A multilevel analysis. Police Quarterly, 15, 308–327.

Cordner

(1989). Police agency size and investigative effectiveness. Journal of Criminal Justice, 17, 145–155.

10.

Corman

Mocan

(2005). Carrots, sticks, and broken windows. The Journal of Law and Economics, 48, 235–266.

11.

Coupe

R. T.

Girling

A. J.

(2001). Modelling police success in catching burglars in the act. Omega: The International Journal of Management Science, 29, 19–27.

12.

Davies

H. J.

(2007). Understanding variations in murder clearance rates: The influence of the political environment. Homicide Studies, 11, 133–150.

13.

Doerner

W. M.

Doerner

W. G.

(2012). Police accreditation and clearance rates. Policing: An International Journal of Police Strategies and Management, 35, 6–24.

14.

Eck

J. E.

(1983). Solving crimes: The investigation of burglary and robbery. Washington, DC: Police Executive Research Forum.

15.

Falcone

D. N.

Wells

L. E.

Weisheit

R. A.

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

16.

Gilbin

M. J.

Burruss

G. W.

Schafer

J. A.

(2014). A stone’s throw from the metropolis: Re-examining small-agency homeland security practices. Justice Quarterly, 31, 368–393.

17.

Goldberg

Champagne

(2005). Is Public Law 280 fit for the twenty-first century? Some data at last. Connecticut Law Review, 38, 697–730.

18.

Graham

J. W.

(2009). Missing data analysis: Making it work in the real world. Annual Review of Psychology, 60, 549–576.

19.

Graham

J. W.

Olchowski

A. E.

Gilreath

T. D.

(2007). How many imputations are really needed? Some practical clarifications of multiple imputation theory. Prevention Science, 8, 206–213.

20.

Greenwood

P. W.

Chaiken

J. M.

Petersilia

Prusoff

(1975). The criminal investigation process volume III: Observations and analysis. Santa Monica, CA: Rand.

21.

Greenwood

P. W.

Petersilia

(1975). The criminal investigation process volume I: Summary and policy implications. Santa Monica, CA: Rand.

22.

Hinkle

J. C.

Weisburd

(2008). The irony of broken windows policing: A micro-place study of the relationship between disorder, focused police crackdowns and fear of crime. Journal of Criminal Justice, 36, 503–512.

23.

Holmes

M. D.

Painter

M. A.

Smith

B. W.

(2017). Citizens’ perceptions of police in rural US communities: A multilevel analysis of contextual, organizational and individual predictors. Policing and Society, 27, 136–156.

24.

Jang

Hoover

L. T.

Lawton

B. A.

(2008). Effect of broken windows enforcement on clearance rates. Journal of Criminal Justice, 36, 526–538.

25.

Johnson

D. R.

Young

(2011). Toward best practices in analyzing datasets with missing data: Comparisons and recommendations. Journal of Marriage and Family, 73, 926–945.

26.

Jorgensen

(2017, 5 9). Pol to propose bill mandating NYPD to report quarterly on precinct-by-precinct clearance rates for major crimes. New York Daily News. Retrieved from http://www.nydailynews.com/new-york/nyc-crime/bill-require-nypd-report-clearance-rates-major-crimes-pol-article-1.3151176

27.

Kaste

(2015, 3 30). How many crimes do your police ‘clear’? Now you can find out. National Public Radio. Retrieved from http://www.npr.org/2015/03/30/395799413/how-many-crimes-do-your-police-clear-now-you-can-find-out

28.

Kaylen

M. T.

Pridemore

W. A.

(2011). A reassessment of the association between social disorganization and youth violence in rural areas. Social Science Quarterly, 92, 978–1001.

29.

Kaylen

M. T.

Pridemore

W. A.

(2012). Systematically addressing inconsistencies in the rural social disorganization and crime literature. International Journal of Rural Criminology, 1, 134–152.

30.

Kaylen

M. T.

Pridemore

W. A.

(2013). Social disorganization and crime in rural communities. British Journal of Criminology, 53, 905–923.

31.

Keel

T. G.

Jarvis

J. P.

Muirhead

Y. E.

(2009). An exploratory analysis of factors affecting homicide investigations: Examining the dynamics of murder clearance rates. Homicide Studies, 13, 50–68.

32.

Y. Y.

(2011). Social structure and informal social control in rural communities. International Journal of Rural Criminology, 1, 63–88.

33.

Liederbach

Frank

(2003). Policing Mayberry: The work routines of small-town and rural officers. American Journal of Criminal Justice, 28, 53–72.

34.

Liederbach

Frank

(2006). Policing the big beat: An observational study of county level patrol and comparisons to local small town and rural officers. Journal of Crime and Justice, 29, 21–44.

35.

Litwin

K. J.

(2004). A multilevel multivariate analysis of factors affecting homicide clearances. Journal of Research in Crime and Delinquency, 41, 327–351.

36.

Litwin

K. J.

(2007). The dynamic nature of homicide clearances: A multilevel model comparison of three time periods. Homicide Studies, 11, 94–114.

37.

Lockwood

(2014). What clears burglary offenses? Estimating influences of multiple perspectives of burglary clearance in Philadelphia. Policing: An International Journal of Police Strategies and Management, 37, 746–761.

38.

Logie

Wright

Decker

(1992). Recognition memory performance and residential burglary. Applied Cognitive Psychology, 6, 109–123.

39.

McCluskey

J. D.

Cancino

J. M.

Tillyer

M. S.

Tillyer

(2014). Does organizational structure matter? Investigation centralization, case clearances, and robberies. Police Quarterly, 17, 250–275.

40.

Miethe

T. D.

Hughes

McDowall

(1991). Social change and crime rates: An evaluation of alternative theoretical approaches. Social Forces, 70, 165–185.

41.

Moore

(2016, 12 11). City’s performance portal reveals clearance rates for Topeka Police Department. The Topeka Capital-Journal. Retrieved from http://cjonline.com/news/2016-12-11/city-s-performance-portal-reveals-clearance-rates-topeka-police-department

42.

Mosher

(2001). Predicting drug arrest rates: Conflict and social disorganization perspectives. Crime and Delinquency, 47, 84–104.

43.

Mustaine

E. E.

Tewksbury

Corzine

Huff-Corzine

(2013). Social disorganization and unfounded sexual assault case clearances. Violence and Victims, 28, 90–102.

44.

Nagin

D. S.

(2016). What we’ve got here is failure to communicate. Criminology & Public Policy, 15, 753–765.

45.

Nagin

D. S.

Solow

R. M.

Lum

(2015). Deterrence, criminal opportunities, and police. Criminology, 52, 74–100.

46.

Osgood

D. W.

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

47.

Osgood

D. W.

Chambers

J. M.

(2000). Social disorganization outside the metropolis: An analysis of rural youth violence. Criminology, 38, 81–116.

48.

Ousey

G. C.

Lee

M. R.

(2010). To know the unknown: The decline in homicide clearance rates, 1980-2000. Criminal Justice Review, 35, 141–158.

49.

Palomino

Veklerov

(2017, 4 7). In Richmond, high number of homicides go unsolved. San Francisco Examiner. Retrieved from http://www.sfchronicle.com/crime/article/In-Richmond-many-murders-go-unsolved-11055724.php

50.

Pare

Felson

R. B.

Ouimet

(2007). Community variation in crime clearance: A multilevel analysis with comments on assessing police performance. Journal of Quantitative Criminology, 23, 243–258.

51.

Pare

Ouimet

(2004). A measure of police performance: Analyzing police clearance and crime statistics. The Canadian Review of Policing Research, 1.

52.

Park

R. E.

Burgess

E. W.

McKenzie

R. D

. (1925/1984). The city. Chicago, IL: The University of Chicago Press.

53.

Pastia

Davies

. (2017). Factors influencing the probably of clearance and time to clearance of Canadian homicide cases, 1991-2011. Homicide Studies. Advanced online publication. doi:10.1177/1088767916686148

54.

Payne

B. K.

Berg

B. L.

Sun

I. Y.

(2005). Policing in small town America: Dogs, drunks, disorder, and dysfunction. Journal of Criminal Justice, 33, 31–41.

55.

Peck

J. H.

(2015). Minority perceptions of the police: A state-of-the-art review. Policing: An International Journal of Police Strategies and Management, 38, 173–203.

56.

Pickett

J. T.

Roche

S. P.

(2016a). Arrested development: Misguided directions in deterrence theory and policy. Criminology & Public Policy, 15, 727–751.

57.

Pickett

J. T.

Roche

S. P.

(2016b). A few clarifying comments on Pickett and Roche (2016). Criminology & Public Policy, 15, 831–836.

58.

Pogarsky

Loughran

T. A.

(2016). The policy-to-perceptions link in deterrence: Time to retire the clearance rate. Criminology & Public Policy, 15, 777–790.

59.

Pratt

Cullen

(2005). Assessing macro-level predictors and theories of crime: A meta-analysis. Crime and Justice, 32, 373–450.

60.

Puckett

J. L.

Lundman

R. J.

(2003). Factors affecting homicide clearances: Multivariate analysis of a more complete conceptual framework. Journal of Research in Crime and Delinquency, 40, 171–193.

61.

Raphael

(2016). Optimal policing, crime, and clearance rates. Criminology & Public Policy, 15, 791–798.

62.

Reaves

(2011). Tribal law enforcement, 2008. Washington, DC: Bureau of Justice Statistics.

63.

Regoeczi

W. C.

Jarvis

J. P.

(2013). Beyond the social production of homicide rates: Extending social disorganization theory to explain homicide case outcomes. Justice Quarterly, 30, 983–1014.

64.

Rhodes

T. R.

Johnson

R. R.

(2008). Patrol officer call for service workloads: An urban and small town comparison. The Police Journal, 81, 184–195.

65.

Roberts

(2008). The influences of incident and contextual characteristics on crime clearance of non-lethal violence: A multi-level event history analysis. Journal of Criminal Justice, 36, 61–71.

66.

Roberts

J. M.

(2016). Crime clearance and temporal variation in police investigative workload: Evidence from National Incident-Based Reporting System (NIBRS) Data. Journal of Quantitative Criminology, 32, 651–674.

67.

Roth

(2017). A city-level analysis of property crime clearance rates. Criminal Justice Studies, 30, 45–62.

68.

Sampson

R. J.

(1986). The effects of urbanization and neighborhood characteristics on criminal victimization. In Figlio

R. M.

Hakim

Rengert

G. F.

(Eds.), Metropolitan crime patterns (pp. 3–25). Monsey, NY: Willow Tree Press.

69.

Schlomer

G. L.

Bauman

Card

N. A.

(2010). Best practices for missing data management in counseling psychology. Journal of Counseling Psychology, 57, 1–10.

70.

Shaw

C. R.

McKay

H. D.

(1942). Juvenile delinquency and urban areas. Chicago, IL: University of Chicago Press.

71.

Stacey

Martin

K. H.

Brick

B. T.

(2017). Victim and suspect race and the police clearance of sexual assault. Race and Justice, 7, 226–255.

72.

Taylor

T. J.

Holleran

Topalli

(2009). Racial bias in case processing: Does victim race affect police clearance of violent crime incidents? Justice Quarterly, 26, 562–591.

73.

Tilley

Burrows

. (2005). An overview of attrition patterns. London, England: Home Office Research Development and Statistics Directorate. Retrieved from http://webarchive.nationalarchives.gov.uk/20080305180705/http://www.homeoffice.gov.uk/rds/pdfs05/rdsolr4505.pdf

74.

U.S. Census Bureau. (2014). 2013 American Community Survey 5-Year Estimates. Washington, DC. Retrieved from http://www.census.gov/acs/www/data/data-tables-and-tools/index.php

75.

United States Department of Agriculture. (2013). 2013 rural-urban continuum codes. Retrieved from http://www.ers.usda.gov/data-products/rural-urban-continuum-codes.aspx

76.

U.S. Department of Justice, Federal Bureau of Investigation. (2004). Uniform Crime Reporting (UCR) Handbook. Washington, DC: Author. Retrieved from https://www.fbi.gov/about-us/cjis/ucr/additional-ucr-publications/ucr_handbook.pdf/view

77.

U.S. Department of Justice, Federal Bureau of Investigation. (2014a). Law Enforcement Management and Administrative Statistics (LEMAS), 2013 (ICPSR36164). Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor].

78.

U.S. Department of Justice, Federal Bureau of Investigation. (2014b). Uniform Crime Reporting Program Data: Arrests by Age, Sex, and Race, Summarized Yearly, 2013 (ICPSR36116-v1). Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor].

79.

U.S. Department of Justice, Federal Bureau of Investigation. (2016). Crime in the United States, 2015. Retrieved from https://ucr.fbi.gov/crime-in-the-u.s/2015/crime-in-the-u.s.-2015

80.

Weisheit

R. A.

Wells

L. E.

Falcone

D. N.

(1995). Crime and policing in rural and small-town America: An overview of the issues. Washington, DC: National Institute of Justice.

81.

Wells

L. E.

Weisheit

R. A.

(2012). Explaining crime in metropolitan and non-metropolitan counties. International Journal of Rural Criminology, 1, 153–183.

82.

Wilkins

D. E.

Lomawaima

K. T.

(2001). Uneven ground: American Indian sovereignty and federal law. Norman, OK: University of Oklahoma Press.

83.

Wilson

J. Q.

Kelling

G. L.

(1982, 3). Broken windows: The police and neighborhood safety. The Atlantic Monthly, 127, 29–38.

84.

Wong

S. K.

(2010). Crime clearance rates in Canadian municipalities: A test of Donald Black’s theory of law. International Journal of Law, Crime and Justice, 38, 17–36.

85.

Worrall

J. L.

(2006). Does targeting minor offenses reduce serious crime? A provisional, affirmative answer based on an analysis of county-level data. Police Quarterly, 9, 47–72.

86.

Worrall

J. L

. (2016). Investigative resources and crime clearances: A group-based trajectory approach. Criminal Justice Policy Review, online before print. doi:10.1177/0887403416650251

87.

Wright

Logie

R. H.

(1988). How young house burglars choose targets. The Howard Journal of Criminal Justice, 27, 92–104.

	Burglary (n = 318)			Larceny (n = 327)			Vehicle Theft (n = 291)
	Coefficient	SE	IRR	Coefficient	SE	IRR	Coefficient	SE	IRR
County agency	−0.123	.191	0.884	−0.348	.198	0.706	−0.022	.2288	0.978
Police spending	0.000	.000	1.000	0.000	.000	1.000	0.000	.0003	1.00
Police workload	0.001	.007	1.001	0.015*	.008	1.015	−0.005	.0093	0.995
Broken windows	−0.002	.007	0.998	−0.001	.007	0.999	0.003	.0081	1.003
Police density	−0.003	.007	0.997	0.002	.006	1.002	−0.017*	.0075	0.983
Total population	−0.009	.005	0.991	−0.007	.005	0.993	−0.013*	.0059	0.987
Juveniles	−0.029*	.012	0.971	−0.009	.012	0.991	−0.017	.0149	0.984
Racial diversity	−0.003	.004	0.997	0.003	.004	1.003	−0.007	.0044	0.993
Unemployment	0.017	.016	1.017	0.012	.016	1.012	−0.001	.0201	0.999
Poverty	0.005	.010	1.005	−0.001	.010	0.999	0.021	.0118	1.021
Residential instability	−0.009	.011	0.991	0.006	.011	1.007	−0.008	.0124	0.992
Rurality	−0.023	.032	0.977	−0.011	.033	0.989	0.001	.0382	1.001
Constant	3.551**	.389	34.911	3.062**	.381	21.393	3.772**	.4927	43.534
Likelihood ratio χ²	14.962			23.025*			24.217**