The Influence of Neighborhood Risk Factors on Parolee Recidivism in Philadelphia,Pennsylvania

Introduction

The mass incarceration experience of the past 40 years has led to the release of more than 700,000 offenders annually (West, Sabol, & Greenman, 2010). For many, successful reentry may prove difficult due to limited education and vocational skills (Harlow, 2003), untreated disorders (James & Glaze, 2006), and unstable living arrangements (Mallik-Kane & Visher, 2008). These barriers to reentry are evidenced by the approximately 68% of offenders rearrested within 3 years of release (Durose, Cooper, & Snyder, 2014). Volume and complexity of problems experienced by ex-offenders, along with high rates of reoffending, raise concern for public safety (Petersilia, 2003), bringing prisoner reentry and recidivism to a prominent position in scholarly literature and public policy.

Recidivism research has largely focused on individual-level attributes, demonstrating age and gender to be robust predictors of recidivism, with younger releasees at greater odds to recidivate (Hipp, Petersilia, & Turner, 2010; Miller, Caplan, & Ostermann, 2016b) and males rearrested at significantly higher rates than females (Miller et al., 2016b). Prior involvement in the criminal justice system (Chamberlain and Wallace, 2016), illicit drug use (Stahler et al., 2013), and unemployment post incarceration (Makarios, Steiner, & Travis, 2010) are also considered risk factors of recidivism. Offenders with stakes in conformity (Hirschi, 1969) including strong family ties and those married post release have been shown to be more successful in the reintegration process, as are those with higher educational attainment (Huebner & Pleggenkuhle, 2013).

Although studies at the individual level have made an important contribution, there is a growing body of research examining the degree to which land use and neighborhood context influence recidivism. The influence of the neighborhood context on criminal behavior has long been recognized in criminological literature (Shaw & McKay, 1942). Understanding the social ecology of neighborhoods on recidivism is critical, particularly when we consider the geographic concentration of returning offenders, typically in disadvantaged high-crime urban communities (Lynch & Sabol, 2001). Indeed, Visher and Farrell (2005) found that more than half (54%) of male prisoners returned to only seven of Chicago’s 77 community areas. High rates of resident turnover through incarceration and reentry may serve to increase crime by disrupting social networks needed to maintain informal social control (Rose & Clear, 1998). The geographic clustering of returning offenders to core communities has been shown to have a “spatial contagion” effect—living in close proximity to others who are offending promotes reoffending behavior (Mennis & Harris, 2011). Indeed, Chamberlain and Wallace (2016) found that the clustering of parolees in concentrated neighborhoods differentially affects individual recidivism, with parolees living in areas with high concentrations of returning offenders 67% more likely to reoffend.

The current study addresses the paucity of this line of research by examining the predictive value of neighborhood contextual factors on the reincarceration of parolees released to the City of Philadelphia. We control for both neighborhood risk and protective factors net of individual-level variables associated with recidivism. We also examine whether the neighborhood context is associated with time to reincarceration.

Neighborhood Context and Crime

Macrosociological explanations of crime are not new to criminological literature (Shaw & McKay, 1942). Shaw and McKay’s (1942) social disorganization theory argued that crime and delinquency are best explained by the traits of the neighborhood, not the traits of the individual. They proposed that dynamic urban change through rapid population turnover and population heterogeneity, often the result of economic deprivation, break down the necessary social networks of informal social control, allowing high rates of crime to occur (Bursik, 1988). Indeed, the independent effect of neighborhood contextual factors has been shown to influence crime (Boggess & Hipp, 2016), as well as other similar outcomes, including victimization (Foshee, Chang, McNaughton, Chen, & Ennett, 2015), mental illness (Rudolph, Stuart, Glass, & Merikanga, 2014), and substance use (Mennis et al., 2016). The breakdown of social institutions within neighborhoods often limits the political power necessary to acquire resources, including government and social service programs, while being unable to fight off undesirable land uses (Bursik & Grasmick, 1993), further contributing to high-crime rates and hindering reintegration.

Although the nexus between the social ecology of neighborhoods and spatial variations in crime has long been recognized, empirical research examining the structure and place-based conditions of the neighborhood on recidivism is limited: “A major premise of these studies is that individual rates of offending are determined, to some extent, by social forces in their wider environment” (Kubrin & Stewart, 2006, p. 169). In a natural experiment following Hurricane Katrina, Kirk (2009) found that ex-offenders displaced from their residence to another parish were less likely to be reincarcerated, including those with extensive criminal histories. His findings suggest a beneficial effect on recidivism in removing offenders from their criminogenic environment.

Ecological studies of recidivism have generally operationalized community-level constructs based on social disorganization theory, yielding mixed results. In their seminal work examining neighborhood economic deprivation on recidivism, Kubrin and Stewart (2006) found that offenders returning to neighborhoods with higher levels of concentrated disadvantage were more likely to be rearrested within the first year of release, controlling for individual-level factors. Their findings suggest an independent contextual effect on recidivism. Hipp et al. (2010) further suggest that the impact of concentrated disadvantage extends beyond the parolee’s neighborhood to the surrounding environs. Furthermore, other studies have found no relationship between neighborhood economic deprivation and reincarceration (Stahler et al., 2013), parole failure (Miller et al., 2016b), or reconviction (Tillyer & Vose, 2011).

More recently, research has begun to explore the interactive effect of environmental and individual-level factors in explaining recidivism. For example, Wang, Hay, Todak, and Bales (2014) examined the effects of criminal propensity and social context on recidivism. Their findings showed limited support for contextual factors that promote crime and individual criminal propensity on recidivism. Chung and Steinberg (2006), however, found that the effect of economic disadvantage on persistent delinquency was mediated by parenting behavior and peer deviance. Similarly, Howard (2016) explored the effects of individual- and neighborhood-level characteristics on drug court graduations. At the individual level, race was found to be a significant predictor of graduation. However, when neighborhood-level variables were introduced, race was no longer significant, suggesting that neighborhood context mediates the individual-level effect of race.

Crime Attractors

Nonresidential land uses that draw outside users, particularly motivated offenders, for their criminal opportunity may also contribute to crime (Brantingham & Brantingham, 1995; McCord, Ratcliffe, Garcia, & Taylor, 2007) and recidivism. Crime attractor locations (e.g., drug markets, alcohol outlets) are, by their very nature, locations that attract offenders for their criminal opportunities. Drawing from crime pattern theory (Brantingham & Brantingham, 1995), we examine the criminogenic influence of drug markets and retail alcohol outlets in a parolee’s neighborhood.

Drug Markets

Crime has been shown to be disproportionately concentrated in drug hot spot areas independent of social disorganization indicators (Martinez, Rosenfeld, & Mares, 2008). Indeed, Wooditch, Lawton, and Taxman (2013) found that probationers were more likely to test positive for illicit drugs with an increased availability of drugs in their immediate environs. However, in a study by Miller et al. (2016b), no relationship was found between the presence of drug markets in a parolee’s neighborhood, and their rearrest or parole revocation. In a follow-up study, Miller, Caplan, and Ostermann (2016a) reoperationalized crime opportunities by examining local violent, drug, and property crime hot spots with only modest support for drug and property crime hotspots and recidivism.

Beer Bars and Alcohol Outlets

The geographical association of alcohol outlets with social disorder is well documented, independent of neighborhood context (Conrow, Aldstadt, & Mendoz, 2015). This is not surprising, given that alcohol often reduces inhibitions and increases risk-taking behaviors (Centers for Disease Control and Prevention [CDC], 2014). To our knowledge, the study by Miller et al. (2016b) of New Jersey parolees is the only study to have examined the presence of alcohol outlets on recidivism. Contrary to their expectations, alcohol outlets were not predictive of a new arrest or parole revocation.

Protective Factors

Much of the recidivism literature focuses on risk factors with limited attention given to the role of protective factors. Protective factors within a neighborhood can serve to insulate and offset exposure to negative influences (Jessor, Turbin, & Costa, 1998). It is reasonable to posit that resource-rich neighborhoods that provide needed services will influence recidivism. Wallace’s (2015) study of neighborhood organization availability on recidivism found that current levels of availability were not related to recidivism; however, the loss of two or more educational organizations was found to increase recidivism. Interestingly, concentrated disadvantage was found to moderate the effect of organizational gains within more affluent neighborhoods. Specifically, organizational gains were found to increase recidivism, suggesting a criminogenic effect of nonresidential land uses where offenders gather (McCord et al., 2007). Wallace and Papachristos (2014) reported similar findings when they contextualized the losses and gains of healthcare organizations with neighborhood concentrated disadvantage. Hipp et al. (2010) suggest that the impact of social services is often diminished by the overburdened and understaffed nature of the providers. The current study examines whether the presence of houses of worship and drug treatment centers in a parolee’s neighborhood reduces recidivism by providing needed social and clinical support.

Religious Institutions, Neighborhoods, and Recidivism

Faith-based institutions may provide returning offenders with spiritual support, prosocial networks, and ancillary services, including shelters and food pantries. Studies suggest that spirituality influences individual behavior (Pullen et al., 2015) with involvement in faith-based communities improving social networks (Lim & Putnam, 2010). Strong social support systems have been well documented to have a positive effect on recidivism outcomes (Duwe, 2011) and to prevent the formation of a criminal identity within the offender (Rocque, Bierie, & MacKenzie, 2011).

Substance Use, Treatment Centers, and Recidivism

Limited numbers of offenders receive prison-based treatment (Mumola & Karberg, 2006) often returning home with lingering substance use disorders (Solomon, Visher, La Vigne, & Osborne, 2006). Moreover, more than half (53%) of substance-using offenders report three or more prior probations or incarcerations (Mumola & Karberg, 2006). Although not fully supported in the literature, community drug treatment centers can provide returning offenders with ease of accessibility to meet their clinical needs and compliance with conditions of release. Several studies have demonstrated that participation in community-based substance use programs reduces drug use and recidivism (Visher & Courtney, 2007), as well as improves outcomes among probationers (Krebs, Strom, Koetse, & Lattimore, 2009).

Current Study

The current study examines reincarceration and time to reincarceration over a 3-year period with a sample of 4,851 parolees. We examine neighborhood contextual factors associated with social disorganization and crime pattern theories. In addition, we consider whether the presence of neighborhood protective factors is associated with reducing recidivism.

Method

Analyses

We estimated a series of logistic regression models for a dichotomous measure of reincarceration. Model 1 controlled for individual-level predictors. Model 2 added neighborhood-level socioeconomic predictors (concentrated disadvantage, race heterogeneity, resident stability). Model 3 included the predictors from Models 1 and 2, and the neighborhood protective and risk factors (drug treatment centers, churches, drug sales arrests, liquor stores, beer outlets, restaurant liquor outlets). Following our logistic regression models, we estimated a series of linear regression models (all three models conducted as above) to measure the time to return to prison by the number of days for parolee’s reincarcerated for a new offense or technical violation within 3 years. Multilevel modeling was not conducted because results from a Durbin–Watson test of independence indicated that they were not necessary with values very close to 2.0.

Findings

Table 1 provides the distribution of the dependent and independent variables and covariates. The majority of parolees were male (95%) and non-White (90%). The descriptives also show that 40% of the parolees’ latest convictions were for violent offenses, 39% were for drug offenses, 11% for property offenses, and 10% fell into the “other” offense-type category. Seventy-seven percent of parolees reported having a drug problem at some point in their lives; 48% reported problems with alcohol in the course of their lifetime.

OPEN IN VIEWER

Table 1.

Descriptive Statistics for Sample.

Variable	M	SD	Range
Reincarcerated in 3yrs	0.49	0.50	0-1
Time to reincarceration	420.91	290.75	2-1,094
Reason for return	0.08	0.28	0-1
Male	0.94	0.21	0-1
Age at release	34.89	9.52	18-76
White non-Hispanic	0.10	0.29	0-1
Married	0.15	0.50	0-1
Unemployed	0.068	0.25	0-1
Problem alcohol	0.48	0.50	0-1
Problem drugs	0.77	0.42	0-1
Arrested prior 16	0.43	0.50	0-1
Two or more prior convictions	0.68	0.47	0-1
Institutional misconduct	0.54	0.50	0-1
Property offense	0.11	0.31	0-1
Drug offense	0.39	0.49	0-1
Other offense	0.10	0.31	0-1
Concentrated disadvantage	0.62	0.75	−1.37-2.52
Racial heterogeneity	0.35	0.21	0.00-0.77
% moved in the last year	13.73	7.12	0-70.2
Drug treatment centers	0.31	0.65	0-5
Churches	6.47	4.42	0-28
Drug sales arrests	38.22	62.16	0-654
Liquor stores	0.14	0.39	0-2
Beer bars/outlets	0.41	0.65	0-4
Restaurants	3.51	3.36	0-76

Note. Drug treatment centers, churches, drug sales arrests, liquor stores, beer bars/outlets, and restaurants are measured by a count within a quarter mile of the subject’s residence.

Table 2 displays the results of the series of logistic regression models estimating return to prison. The results of our first model estimating individual-level predictors of recidivism were consistent with the findings of previous studies with males more likely to be reincarcerated (odds ratio [OR] = 2.46) than females. Age was also significant with each year older at release associated with a 3% decrease in the likelihood of returning to prison. Consistent with prior research, parolees who reported being unemployed prior to their incarceration were at greater odds of being reincarcerated (OR = 1.3). Race was not found to be significantly related to reincarceration; nor was being married at the time of incarceration.

OPEN IN VIEWER

Table 2.

Results of Binary Logistic Regression to Predict Reincarceration Within 3 Years.

Variable	Individual odds ratio	Individual + neighborhood odds ratio	Full model odds ratio
Male	2.46***	2.40***	2.45***
Age at release	0.97***	0.97***	0.97***
White	1.08	1.08	1.08
Married	1.03	1.04	1.04
Unemployed	1.38*	1.35*	1.36*
Problem alcohol	1.10	1.12	1.12
Problem drugs	1.31**	1.30**	1.32**
Arrested prior 16	1.17*	1.15*	1.16*
Two or more prior convictions	1.69***	1.71***	1.70***
Institutional misconduct	1.89***	1.91***	1.91***
Property vs. violent	1.25	1.25	1.26
Drug offense	0.73***	0.73***	0.73***
Other offense	0.90	0.90	0.91
Concentrated disadvantage	—	0.98	0.98
Racial heterogeneity	—	0.72	0.76
%moved in the last year	—	1.01	1.01
Drug treatment centers	—	—	0.97
Churches	—	—	1.00
Drug sales arrests	—	—	1.00
Liquor stores	—	—	1.13
Beer bars/outlets	—	—	1.21***
Restaurants	—	—	0.99
Constant	.32***	.34***	.32***
Nagelkerke R2	.107	.111	.115

Note. See Table 1 for description of variables and how they were coded.

*

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

Parolees who reported problems with drugs in their lifetime were 1.3 times more likely to return to prison; parolees who reported alcohol-related problems were neither more nor less likely to be reincarcerated. However, parolees convicted of drug-related offenses were 27% less likely to return to prison compared with violent offenders; property and “other” offense types were not significant in predicting reincarceration. Criminal history was significantly related to reincarceration with parolees reporting arrests prior to the age of 16 (OR = 1.1) and a history of two more convictions (OR = 1.6) at greater odds of returning to prison. Offenders charged with prison misconduct were found to be 1.8 times more likely to return to prison.

Our next model estimated for both neighborhood- and individual-level factors. Comparison between the models revealed a small increase in the variance explained in Model 2 (Nagelkerke R² = .107 and .111, respectively). All of the individual-level predictors that were significant in the first model remained significant in Model 2 with the direction of the beta value staying the same. All of the neighborhood-level variables including concentrated disadvantage, racial heterogeneity, and percentage moved in the last year were not significantly related to reincarceration.

The final model included neighborhood-level protective and risk factors within 1 quarter mile of the parolee’s residence. Again, we find that the predictors that were significant in Models 1 and 2 remained significant, and the direction of the beta values remained the same. Among neighborhood-level risk factors, the presence of beer bars and liquor outlets increased the likelihood of a parolee being reincarcerated (OR = 1.2). However, we did not find drug sales arrests to be significantly related to a parolee returning to prison. Furthermore, neighborhood protective factors including houses of worship and drug treatment centers were not found to have any effect on the likelihood of a parolee being reincarcerated.

To further explore the result of the logistic regression analyses, an additional series of linear regression models were estimated (Table 3) to examine the time to reincarceration among parolees returned to prison for new offenses or technical violations. Similar to the logistic regression analyses, three models were conducted: The first model included individual-level risk factors of reincarceration. The results of the first model found that males were returned to prison approximately 90 days earlier than female parolees. Again, we find employment to be a significant predictor, with parolees who were unemployed prior to incarceration returning to prison an average of 51 days earlier. Race was significant with White non-Hispanics returned to prison approximately 46 days sooner than non-White parolees.

OPEN IN VIEWER

Table 3.

Results of Multiple Linear Regression to Predict Time to Recidivism.

Variable	IndividualB	Individual + neighborhoodB	Full modelB
Reason for return	342.35***	343.48***	347.97***
Male	−90.17*	−85.62*	−88.39*
Age at release	−0.53	−0.29	−0.16
White	−46.08*	−56.29*	−52.57*
Married	0.43	0.43	−1.40
Unemployed	−51.48	−49.63	−53.15
Problem alcohol	3.52	2.90	2.40
Problem drugs	−47.11**	−49.79**	−49.86**
Arrested prior 16	−35.70**	−35.56**	−35.66**
Two or more prior convictions	18.18	13.14	11.38
Institutional misconduct	−34.31*	−32.10*	−32.43*
Property vs. violent	−99.68***	−101.90***	−100.81***
Drug offense	−47.87**	−46.98**	−45.53**
Other offense	−74.78**	−74.28**	−75.66**
Concentrated disadvantage	—	−5.08	1.50
Racial heterogeneity	—	89.72**	72.87*
%moved last year	—	−0.90	−0.55
Drug treatment centers	—	—	5.15
Churches	—	—	−0.96
Drug sales arrests	—	—	−0.06
Liquor stores	—	—	34.30
Beer bars/outlets	—	—	−20.72*
Restaurants	—	—	−2.59
Constant	678.12***	654.63***	674.96***
R 2	.047	.050	.056

Note. See Table 1 for description of variables and how they were coded.

*

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

Additional findings showed that offenders with drug problems in their lifetime returned to prison on average 41 days earlier than those without reported drug problems. Consistent with prior studies showing contact with the criminal justice system at a young age to be a robust predictor of future criminal behavior, parolees with histories of arrests prior to the age of 16 were returned to prison approximately 35 days earlier. Parolees who were charged with prison misconduct were also found to be reincarcerated on average 34 days earlier compared with offenders with no prison infractions.

Results of our study also found a relationship between the type of offense for which a parolee was convicted and his or her time to reincarceration. Specifically, we found that property offenders were returned to prison nearly 100 days sooner than violent offenders; drug offenders were approximately 47 days quicker to return to prison, and “other” offense type was 74 days sooner compared with the referent group, violent offenders. In addition, we found that parolees remained in the community for approximately 342 days longer if their return to prison was for a new offense compared with a technical violation.

The second model included the predictors from Model 1 with the addition of neighborhood-level factors. Comparison between the models revealed a small increase in the variance explained in Model 2 (Nagelkerke R² = .047 and .050, respectively). Findings of the individual-level predictors in Model 1 were similar to those observed in Model 2 (see full results in Table 3). Examining the influence of neighborhood factors, we found that the racial heterogeneity of the parolee’s neighborhood was a significant predictor in explaining time to reincarceration as we expected. However, our finding was somewhat unexpected, in that parolees residing in neighborhoods characterized as racially heterogeneous remained in the community for longer periods of time (approximately 89 days) prior to being returned to prison. Similar to the findings of our logistic regression, concentrated disadvantage and resident mobility did not predict time to reincarceration.

In the final model, we included all of the predictors from Models 1 and 2, and added neighborhood risk and protective factors. There is a small increase in the variance explained compared with Model 2 (Nagelkerke R² = .050 and .056, respectively). We found similar results for the individual- and neighborhood-level factors in the full model to those in the first two models. However in the final model, we found that the presence of beer bars and liquor outlets in the parolees’ neighborhood reduced their time in the community by an average of 20 days. Drug sales arrests, the presence of houses of worship, or drug treatment centers in the parolee’s neighborhood were not found to have a significant relationship with time to reincarceration.

Discussion and Conclusion

Although criminological research has long sought to understand the influence of the social ecology of neighborhoods on crime, it is only recently that the neighborhood context has been examined in relation to offender reintegration and recidivism. Considering the vast numbers of offenders returning to core communities each year, this is a notable omission in the recidivism literature. The current study contributes to our understanding of the predictive value of the neighborhood context on recidivism by examining neighborhood contextual factors put forth by social disorganization and crime pattern theories using a sample of parolees released to the City of Philadelphia over a period of 2 years.

Overall, our findings suggest a limited role of the neighborhood context in reincarceration for a new offense or technical violation. Our findings do suggest, however, that for reincarcerated parolees, living in more racially heterogeneous neighborhoods is predictive of longer stays in the community prior to reincarceration (89 days for Model 2 and 72 days for Model 3, respectively). Although contrary to what would be expected from a theoretical perspective, this finding is in keeping with current literature that suggests more homogeneous neighborhoods, particularly Black, Hispanic, and foreign born, have higher crime rates and perceived crime and disorder (Johnson, 2016; Lindblad, Manturuk, & Quercia, 2013).

We examined the presence of drug treatment centers and houses of worship within walking distance of parolees’ home for their perceived benefit to returning offenders. Contrary to our expectations, they were found to neither increase nor decrease the likelihood of reincarceration. Although continuum of care in the community setting has been broadly accepted for its beneficial effects (Knight, Simpson, & Hiller, 1999), parolees may not be self-motivated to pursue treatment or access social services due, in part, to limited finances, lack of insurance (Mallik-Kane, 2005; Nelson, Deess, & Allen, 1999), and difficultly navigating services and programs (Rukus & Lane, 2014). Thus, the sheer ease of accessibility may not be sufficient to serve as a protective factor and may benefit longer term after parolees have managed their more immediate transition needs.

Also contrary to our expectations, the percentage of drug sales arrests in parolees’ neighborhood was not significantly related to their reincarceration or time spent in the community. Our findings may reflect an effort by many states, including Pennsylvania, to reduce prison populations through investment in evidence-based practices of risk assessment, community-based treatment, and prison diversion programs for high-risk and substance-using offenders, including parole violators.

One of the most notable findings in our study was the presence of beer bars and liquor outlets in predicting a parolee’s return to prison (OR = 1.2). Moreover, their presence in the neighborhood reduced the time a parolee remained in the community by an average of 20 days. This is not surprising, given the criminogenic nature of alcohol outlets (Conrow, Aldstadt, & Mendoz, 2015). Block and Block (1995) suggest that, by their very nature, bars and taverns are places where strangers congregate, often with limited surveillance and standards of behavior. Alcohol consumption is also known to reduce inhibitions and increase risk-taking behavior (CDC, 2014). For many parolees, the close proximity of alcohol outlets provides ease of accessibility to a population that is disproportionately affected by alcohol use disorders (James, 2004). Caution should be exercised in comparing and generalizing findings of the predictive value of alcohol outlets on recidivism as state licensing laws vary, which may affect availability and presence within defined geographic boundaries.

At the individual level, our results are consistent with prior research showing males, younger offenders, and unemployed parolees at greater risk of reoffending. Prior criminal history and a history of arrest prior to the age of 16 were also associated with an increased likelihood of reincarceration. Although parolees reporting a history of drug use were more likely to be reincarcerated, parolees whose most recent incarceration was for a drug-related offense were 27% less likely to return to prison. Once again, this finding may reflect Pennsylvania’s investment in community-based programming for high-risk offenders and parole violators. We also found that prison misconduct was associated with reincarceration and an earlier return to prison, suggesting that institutional adjustment may serve as an important indicator of postrelease success.

Limitations

Although our findings contribute to the recidivism literature and the role of the neighborhood context in reincarceration, some limitations must be addressed: First, we used the most conservative measure of recidivism—reincarceration. Reincarceration may not reflect the full scope of an individual’s involvement with the criminal justice system. Individuals who were rearrested or reconvicted without being sentenced to state prison will not be reflected in our results. Second, our measures of substance use and criminal history are self-reported as part of the LSI-R, and therefore caution must be exercised in the offender’s recollection and accuracy of responses. However, the LSI-R is a standardized instrument that has been empirically validated on diverse samples of offenders (Andrews & Bonta, 1995). Third, this was a cross-sectional sample of parolees limited to one geographic location, Philadelphia, Pennsylvania. Therefore, caution should be exercised in generalizing these findings to be reflective of all parolees across the nation.

Conclusion

To our knowledge, this is the first study to examine both risk and protective factors of a neighborhood for their predictive value on reincarceration. Although we found limited support for the neighborhood context on recidivism, our finding of alcohol outlets for their ability to predict reincarceration suggests the importance of examining nonresidential land uses in recidivism research. Replicating and expanding these findings may provide for more informed decisions regarding postrelease case plans for individuals returning to high-risk neighborhoods. Future research should further consider the importance of examining the interaction effect of specific land uses with individual-level traits, including criminogenic risks and needs.