The Impact of Neighborhood Context on Intragroup and Intergroup Robbery: The San Antonio Experience

Abstract

Guided by social disorganization theory, this article examines the influence of neighborhood characteristics on intragroup and intergroup robbery, net of spatial proximity in a predominantly native-born Latino/Mexican-origin city—San Antonio, Texas. From census tract and official police robbery data, the findings indicate that intragroup robbery is more common than intergroup robbery. Multivariate results show that variation in black intra-group robbery lies primarily in highly disadvantaged neighborhoods; whereas variation in Latino intergroup robbery is found in neighborhoods with more disadvantage, racial/ethnic heterogeneity, recent immigrants, and blacks. Residential instability persistently influences all robbery types. Disaggregating robberies by race and ethnicity reveals the importance of examining Latinos as offenders and victims. The case of San Antonio serves as a harbinger of conditions that may exist in the growing number of majority-Latino cities—and suggests the need to investigate crime experiences that move beyond studying racial dichotomies of violence.

Since the 1990s, researchers have documented the significance of examining race-specific homicide victimization rates (Martinez 2002). While most scholars have found differences in black and white violent victimization, they have directed less attention to the study of intraracial and interracial violence. The few existing studies focus on large units of analysis (e.g., cities) and often exclude ethnic groups such as Latinos (e.g., Wadsworth and Kubrin 2004). Indeed, we found no published neighborhood-level studies of intragroup or inter-group violence (see Sampson [1984] regarding the need to explore intergroup violence at the community level). This neglect is curious given the central role of communities in generating variation in crime including race/ethnic-specific violence (Wadsworth and Kubrin 2004). The absence of scholarly attention to Latino within-and between-group violence is also surprising in light of this population’s growth in American society over recent decades (Martinez 2002).

Drawing on social disorganization theory, we analyze data for neighborhoods in San Antonio, Texas, in an attempt to address this neglected area of crimino-logical research. The primary objective is to explore whether the social structure of neighborhoods is an important source of variation in intragroup/intergroup robbery among Latinos and blacks. The compositional characteristics of San Antonio make it an ideal research site. For more than thirty years, the Latino share of the city population has grown, and it is now a “majority-minority” city with Latinos composing 59 percent of the population. One consequence of the Latino increase is a notable reduction in the levels of social isolation (i.e., residential segregation) between Latinos and members of other racial/ethnic groups. Indeed, neighborhood exposure between blacks and Latinos in San Antonio has increased by nearly 40 percent over the past three decades, and blacks currently live in communities that are nearly 50 percent Latino.

Despite the changing racial/ethnic composition of San Antonio, Latinos and blacks tend to reside in structurally similar communities. Census data indicate that, on average, communities in which blacks and Latinos reside have comparable levels of poverty, unemployment, and vacant housing, factors commonly associated with social disruption and violence. We contend that contact between blacks and Latinos in primarily disadvantaged neighborhoods is likely to have implications for patterns of violence. For example, it is possible that greater between-group contact promotes intergroup conflict, owing to competition for limited resources and/or racial/ethnic tension. Social structural conditions may affect intragroup violence differently, with factors such as the presence of immigrants potentially providing a buffer against intragroup violence for Latinos, but not for blacks (Martinez 2002). Following the arguments advanced by Sampson (1984, 620), we recognize that “the same structural conditions that make cordial inter-group relations more likely also make interpersonal conflict between members of different groups more likely.” This analysis attempts to tease out the complexities associated with neighborhood effects on within- and between-group violence.

We expand on previous studies in three substantive ways. First, we examine the routine crime of robbery, a financially motivated crime that generally occurs between strangers in public settings and one that is linked to public perceptions of street violence (Wadsworth and Kubrin 2004). Second, we disaggregate robberies into combinations of offender and victim race/ethnicity, which permits us to move beyond the study of broad group-specific (i.e., black or Latino) victimization or offending rates. By extending robbery classifications, we cast new light on the presumed volatile relationship between racial/ethnic minority groups in the United States (Sampson 1984). Third, this study contributes to the violence literature by focusing on neighborhood rather than larger macro-level (i.e., nation, county, and city) effects. We posit that neighborhoods are more proximate contexts for intragroup/intergroup violence. Indeed, violent crime is not randomly distributed across areas within cities but is likely influenced by local levels of group interaction in addition to instability, disadvantage, and other factors. The present study advances knowledge about which types of neighborhood conditions, similarly or differentially, shape specific combinations of intragroup and intergroup violence. It also highlights how well social disorganization theory explains these various kinds of robbery, while transcending the black-white paradigm.

Conceptual Arguments and Hypotheses

Neighborhoods have long played a prominent role in explaining variation in crime (Shaw and McKay 1942). Socially disorganized communities characterized by high levels of poor economic conditions, population turnover, and racial heterogeneity are typified as places with high levels of violent crime, relative to more affluent, homogeneous neighborhoods with stable populations. Impoverished and unstable neighborhoods undermine citizens’ abilities to effectively exert informal mechanisms of social control, fostering a community context that is susceptible to crime. A central premise in Shaw and McKay’s (1942) theoretical writing is that high crime rates (and disorder) become long-standing features of some neighborhoods irrespective of which racial, ethnic, or native group(s) resides in those areas. A contemporary articulation of Shaw and McKay’s guiding principle is commonly referred to as the racial invariance hypothesis, which asserts that the “sources of violent crime appear to be remarkably invariant across race and rooted instead in the structural differences among communities” (Sampson and Wilson 1995, 41).

As American society grows increasingly diverse, however, there is continued discussion as to whether structural conditions indeed affect crime rates equally for all racial/ethnic groups (Martinez 2002). Existing research provides good reasons to expect that social structural influences may exert disparate influences on levels and types of offending across communities. For example, Latinos are more economically similar to blacks than they are to whites, yet Latinos consistently lack characteristics associated with the extremely disadvantaged and do far better on a number of social outcomes than blacks (Palloni and Morenoff 2001). The advantages experienced by Latinos across an array of factors (i.e., birth weight, infant mortality) have been characterized as the “Latino Paradox” and persist despite similarities with blacks in the social structural conditions they experience (Palloni and Morenoff 2001). Such Latino–black disparities may reflect that the ability of community social networks to fight against perilous outcomes, including crime, are not always undermined in the uniform manner characterized by the disorganization perspective (Moore and Pinderhughes 1993).

At the same time, it is important to consider that patterns of violence may also be shaped by cross-group interactions. Under some conditions, reduced residential isolation for both blacks and Latinos, coupled with less favorable economic conditions, may enhance existing social tensions between the groups. As a result, the increased exposure may present a barrier to the development of effective mechanisms of informal social control within communities. Partial support for this notion is found in the literature showing that community social networks tend to be established and accessed largely along racial/ethnic lines (McPherson, Smith-Lovin, and Cook 2001). The maintenance of primarily intraracial/ethnic networks within an otherwise diverse environment may promote feelings of conflict between groups, seemingly pitting members of one group against another in the competition for scarce resources (Camarillo 2004; Hernandez 2007). Considering the role of conflict, it may be that rather than impacting crime generally (as social disorganization holds), neighborhood composition has strong influences only on some types of crime, such as black on Latino, but not on others, such as Latino on black. Assessing this type of issue is timely because due to high fertility and immigration rates, the Latino population is growing, potentially raising conflict with African Americans and perhaps other groups.

Research hypotheses

As noted, despite the similar contextual conditions experienced by many Latinos and blacks in San Antonio, the effects of structural factors on violence may differ. For example, social disorganization theory identifies neighborhood disadvantage as one of the strongest predictors of community violence. Yet the evidence suggests that the link between disadvantage and violence may be less pernicious for Latinos than for blacks. Recently, Nielsen, Lee, and Martinez (2005) found that neighborhood disadvantage is a stronger predictor of black than Latino homicide victimization. Similar to the proponents of the “Latino Paradox,” criminologists conclude that Latinos may be better equipped to withstand crime-facilitating neighborhood conditions than other groups (Sampson 2008). Notably, residents of poor Latino neighborhoods have ties to strong employment and kinship networks that help to mitigate the impact of disadvantage on violence in these communities. Access to such protective mechanisms is typically more limited among black neighborhoods, and so disadvantage should have a strong impact on black victimization. In light of these findings, we hypothesize that disadvantage will not be a strong predictor of Latino victimization regardless of the race of the offender, but this factor will have a positive association with the victimization of blacks.

Residential instability is another community characteristic thought to be causally linked to crime. According to social disorganization theory, instability makes it difficult for residents to establish strong social ties, which undermines informal social control. Studies have generally shown that residential turnover is positively linked to levels of violence (Shaw and McKay 1942). The evidence also indicates that instability has a similar impact on victimization rates for blacks and Latinos. For example, Lee (2003, 75) found a consistent positive association between instability and homicide victimization across groups, which he argues is “entirely in keeping with the expectations” of social disorganization theory (cf. Martinez, Stowell, and Cancino 2008). For both Latinos and blacks, residential turnover may fracture established networks (due to out-migration) or require time and effort to establish new, and maintain old, networks (due to in-migration). Either situation may promote higher levels of crime. Thus, the weight of empirical evidence provides a solid basis for expecting a positive relationship between neighborhood residential instability and victimization risks, independent of race or ethnicity.

Neighborhood racial/ethnic heterogeneity is the third structural factor of primary interest in this study. Theory and past research suggest that such heterogeneity might or might not have uniform effects across types of interracial and intraracial robbery. According to social disorganization theory, all types of crime are likely to be higher in areas composed of racially/ethnically mixed populations because of diminished social control. Shaw and McKay (1942) argue that language and cultural differences associated with neighborhood heterogeneity make it difficult for residents to effectively communicate and solve common problems. One would not anticipate a differential impact of heterogeneity on levels of black versus Latino victimization irrespective of the race-ethnicity of the offender based upon this theory. In contrast, some research suggests that racial-ethnic heterogeneity may be a stronger predictor of interracial than intraracial violence. Such an effect might be expected because greater heterogeneity offers more opportunities for cross-group victimization (see Blau and Blau 1982). Increased exposure may also fuel existing antipathies between groups. For example, in a study of black-Latino transition in Compton, California, Camarillo (2004, 367) notes that the “persistence of stereotyping and other negative perceptions affect social relations between these groups.” Hernandez (2007) also argues that prejudices between blacks and Latinos may translate into selective targeting of out-group members for victimization. Accordingly, it seems plausible that diversity could promote intergroup conflict, subsequently leading to higher levels of out-group but not intragroup violence.

Finally, it is important for researchers to consider neighborhood social conditions that may diminish intragroup/intergroup violence. Immigration represents one such factor (Martinez 2002). Social disorganization theory proposes that immigration disrupts community organization (i.e., stability, homogeneity) such that communities with larger foreign-born populations should have higher levels of violent crime. Yet, literature on Latino crime has consistently demonstrated that immigrant areas have less crime than expected, for foreign-born residents and for native-born non-Latino whites and non-Latino blacks (Sampson 2008). Researchers have yet to explicate fully the mechanisms through which immigration is able to buffer communities against crime. A number of possibilities have been proposed, including the notion that immigrants “revitalize” structurally disadvantaged communities in which they settle (Lee 2003). Guided by the empirical evidence, we hypothesize that the size of the foreign-born population will be inversely associated with both intergroup and intragroup violence.

In sum, social disorganization theory and other contemporary community perspectives on immigration offer viable, but sometimes contradictory, expectations regarding the impact of neighborhood structural factors on within- and between-group violence. To assume a uniform effect of structural factors on violence is problematic because it overlooks the fact that structure shapes patterns of both in-group and out-group interaction, perhaps resulting in complex rather than uniform effects. The present study is designed as an initial attempt to assess the varying degrees to which neighborhood conditions promote or reduce specific types of robbery.

Research Setting

San Antonio has a unique demographic profile. Most of the population is composed of Latinos (58.7 percent), followed by non-Latino whites (31.1 percent), blacks (6.9 percent), and Asians (1.9 percent). The Latino population is primarily native-born, with the immigrant city proportion in line with national averages because of San Antonio’s close proximity to the Mexican border. Thus, not only is the city an attractive research setting for studying race/ethnicity and violent crime, it also provides an opportunity to assess these relationships in a place that absorbs many immigrants. As such, San Antonio provides a good context for exploring the complex roles of ethnicity, immigration, and race in producing violence.

Data and Methods

The study examines predictors of intragroup and intergroup robbery data for San Antonio’s 242 census tracts (as proxies for neighborhoods). Data come from two independent sources. Reported robbery information is from San Antonio Police Department (SAPD) official records for three calendar years (2002–2004). All other data are from the 2000 Census Summary File 3. Robbery data for three years are combined to minimize the impact of annual fluctuations and to increase the likelihood of having sufficient incidents to calculate reliable rates (Lee 2003). Robbery is selected over other violent crimes (e.g., assault and homicide) for substantive and methodological reasons. Substantively, robbery involves the threat or actual use of force to obtain an item of value, and it often occurs between strangers (Cohen, Felson, and Land 1980). Thus, robbery not only draws from violent crime but also has an element of property offending, thus distinguishing it from other types of violence. Methodologically, citizen reporting of robbery is consistent and high among violent crime victimizations when compared to assaults, which are subject to underreporting in part because the victim and offender are usually in some type of previous relationship (Smith, Frazee, and Davison 2000). Furthermore, robbery occurs more frequently compared to rare homicide events, which helps to yield a more generalizable pattern of estimates. The official robbery data include the address where the incident took place and offender/victim race and ethnicity. Robbery information is available for 89 percent of the offenders and 96 percent of the victims. Using geographic coordinates provided in the official SAPD data file, intragroup/intergroup robbery incidents were geocoded (with a successful match rate of more than 90 percent) and then aggregated to the census tract level.

Dependent variables

The dependent variables consist of three-year (2002–2004) census tract robbery counts for four combinations of outcomes pertaining to the racial and ethnic characteristics between offender and victim: (1) black offender–black victim, (2) Latino offender–Latino victim, (3) black offender–Latino victim, and (4) Latino offender–black victim. White robbery cases are not included because there are too few counts to examine in San Antonio. D’Alessio, Stolzenberg, and Elite (2002) reported that whites are more likely to offend against blacks in areas with high levels of economic competition. However, in San Antonio, non-Latino whites do not face such competition from Latinos or blacks; instead, for the latter groups, the local history is one of high levels of resource deprivation for both.

Independent variables

Consistent with previous neighborhood-level research, we can construct several indicators associated with social disorganization. The disadvantage index is operationalized as the average z score for percentage poverty, percentage female headed-households with children, percentage unemployed, and percentage no college education. Residential instability is measured as the average z score for percentage of the population that moved within the past five years and percentage vacant housing units. The crime-prone population is defined as the share of the population composed of males between the ages of eighteen and thirty-four. The racial/ethnic heterogeneity index assesses the proportion of a neighborhood’s population that is composed of the four largest racial/ethnic groups (i.e., non-Latino black, non-Latino white, Latino, non-Latino Asian). The measure used here, also known as the Herfindahl index, yields a value ranging from 0 to 1, with larger values indicating higher heterogeneity levels (M = .42) (Hipp 2007, 674). Finally, we employ an indicator of immigration that is consistent with prior social disorganization research on homicides. Recent immigrants is operationalized as the percentage of the neighborhood population that is foreign-born and arrived in the 1990s (Martinez, Stowell, and Cancino 2008).

Control variables

Given the nested (e.g., offender-victim robbery counts within tracts) and spatial (e.g., geographical tract adjacency) nature of the data, we include three control variables. Downtown (1 = downtown census tract, 0 = other census tract) controls for high-volume traffic by tourists and residents frequenting the low-population base center-city entertainment area (Smith, Frazee, and Davison 2000). We account for racial composition by controlling for the percentage of the tract that is black (i.e., percentage black). Percentage black is included to determine the independent effect of racial composition, net of heterogeneity (see Hipp 2007, 675). To control for spatial autocorrelation (Morenoff, Sampson, and Raudenbush 2001) and robbery diffusion (Smith, Frazee, and Davison 2000), we construct spatial lags using a two-stage estimation procedure (e.g., Tolnay, Deane, and Beck 1996) for each of the dependent variables to reflect weighted averages of intragroup/intergroup robbery counts in all tracts surrounding a particular tract.

Analytic strategy

Although more common than other forms of violence, robbery tends to yield low counts when disaggregated by offender-victim race/ethnicity at the census tract level. Due to the highly skewed distribution of outcomes and the overdispersion of all robbery count-dependent variables, we employed a Poisson-based negative binomial estimation procedure (Osgood 2000). The negative binomial strategy translates rare events into rates by including an offset measure based on the natural logarithm of the population at risk; for our purposes, we base this on the race/ethnicity of the victim. In the following analyses, intragroup/intergroup robbery counts (i.e., rates) are regressed on the independent variables across 242 San Antonio census tracts.

Findings

Before proceeding with the multivariate results, we draw attention to several descriptive patterns. During the period under investigation, the most common robbery type involved Latinos victimized by fellow Latinos (65 percent). Black victimization of Latinos was the second most common type (14 percent), followed by black intragroup (12 percent) and Latino-on-black (9 percent) robbery. Descriptive statistics, zero-order correlation matrices, and diagnostics are available upon request.

Table 1 presents four negative binomial regression models for intragroup/intergroup robbery outcomes. Our results indicate that neighborhood disadvantage significantly influences black-on-black (model 1) and Latino-on-black (model 4) robbery. Neighborhoods with higher levels of disadvantage experience significantly more black robbery victimization. Residential instability is positively and significantly related to all types of robbery.

Table 1.

Negative binomial regressions for intragroup/intergroup robbery, san antonio census tract (N = 242)

	Model 1: Black on Black		Model 2: Latino on Latino		Model 3: Black on Latino		Model 4: Latino on Black
	B	SE	B	SE	B	SE	B	SE
Disadvantage index	0.454	0.149**	0.143	0.112	0.291	0.158	0.703	0.201***
Residential instability	0.174	0.065**	0.233	0.052***	0.244	0.075**	0.177	0.094*
Crime-prone population	−0.027	0.021	0.003	0.015	−0.044	0.025	−0.014	0.031
Racial/ethnic heterogeneity	0.905	0.750	0.734	0.616	3.240	0.721***	−2.451	1.021*
Recent immigrants	0.047	0.038	0.008	0.021	0.082	0.036*	0.024	0.043
Downtown	0.331	0.167*	0.215	0.121	0.249	0.190	0.170	0.239
Percentage black	−0.013	0.007	0.005	0.007	0.022	0.009*	−0.062	0.011***
Spatial lag	0.903	0.274***	0.442	0.065***	1.219	0.384**	1.427	0.674*
Intercept	−7.560	0.544***	−10.168	0.631***	−11.140	0.673***	−5.807	1.012***

NOTE: Logged race/ethnic group of victim included as regression offset.

^*p < .05.

^**p < .01.

^***p < .001.

As for racial/ethnic heterogeneity, our measure of group contact, an interesting pattern of results emerges for intergroup violence. Neighborhoods characterized by higher racial/ethnic heterogeneity and black racial composition exhibit significantly higher black-on-Latino robbery (model 3). In contrast, more heterogeneous communities and those with larger black populations have significantly lower rates of Latino-on-black robbery (model 4). Thus, in more heterogeneous and more predominantly black neighborhoods, blacks are more likely to rob Latinos while Latinos are less likely to rob blacks. Interestingly, racial/ethnic heterogeneity is not associated with intragroup robbery for either blacks or Latinos.

Unlike previous research showing that immigrant neighborhoods are inversely related to racial/ethnic group-specific violent victimization rates (Martinez 2002), San Antonio neighborhoods with more recent immigrants show significantly more black-on-Latino robbery (model 3). This pattern suggests that recent Latino immigrants may be more vulnerable to victimization by blacks, especially in (or around) historically black neighborhoods. More broadly, the observed null effect of immigration is inconsistent with expectations of social disorganization theory (Lee 2003). Our expectation that neighborhoods with higher levels of recent immigrants would provide a “protective” buffer against intragroup/intergroup robbery is unsupported. Regarding the spatial lags, all four types of robbery tend to be spatially dependent.

Discussion and Conclusion

Motivated by social disorganization theory, this study examined Latino versus black within- and between-group violence. We sought to broaden current knowledge regarding criminal violence in three ways. First, we drew attention to robbery (a financially motivated violent crime occurring between strangers) among similarly situated racial and ethnic groups. Second, we disaggregated robberies by offender and victim race/ethnicity, which allowed us to compare intergroup with intragroup violence. Finally, analyses of tract-level data permitted a more refined understanding of how local social structure influences different types of robbery. Consistent with previous studies, we found that robbery tends to be largely an intragroup phenomenon (Wadsworth and Kurbin 2004). In line with expectations, we also found both similarities and differences in the correlates of intragroup/intergroup robbery across San Antonio census tracts as well as mixed support for the social disorganization perspective. Consistent with social disorganization claims about uniformity, residential instability is positively associated with all types of robbery. In contrast, the effect of disadvantage is not uniform for robbery types; this factor is a stronger predictor of black than Latino victimization. Furthermore, our findings reveal disparate impacts of racial/ethnic heterogeneity, percentage recent immigrant, and percentage black robberies. These factors all impact intergroup but not intragroup robbery yet have different patterns (e.g., directions) of influence on the former type of violence. We interpret these patterns as supporting our contentions that social structural factors can impact intragroup/intergroup violence differently and that similar structural contexts may lead to divergent violent outcomes.

Prior literature has offered some food for thought for interpreting the disparate effects of social structure that we observed. Based on the work of McPherson, Smith-Lovin, and Cook (2001) and Hernandez (2007), it appears that the consequences of neighborhood heterogeneity may generate an unease (i.e., social distance) between Latinos and blacks that then promotes intergroup robbery by blacks against Latinos. This association may be due, in part, to changing neighborhood demographics as well as racial/ethnic conflict. Black intergroup robbery may increase as Latinos move into older black neighborhoods. As Latinos gradually settle into predominantly black neighborhoods (suggested by decreases in residential segregation), residents may respond by committing acts of violence to “defend” what is perceived as proprietary social space (i.e., see Suttles’s [1968] defended community thesis).

Yet, it is puzzling that heterogeneity and racial composition produce opposite results for Latino intergroup robbery. Apparently, more heterogeneous neighborhoods and those with more blacks do not translate into the types of responses whereby Latinos increase their robbery victimization of blacks. Research that explores in greater detail the circumstances of Latino versus black robberies is needed to shed additional light on these contrasting results.

While blacks and Latinos in San Antonio tend to reside in seemingly structurally similar neighborhoods, these similarities manifest in different intragroup/intergroup victimization risks. Some results are consistent with the expectations of the “Latino Paradox,” which holds that social outcomes for Latinos are better than those of other groups living in similar structural locations. We found support for this perspective given that Latino victimization is not influenced by neighborhood disadvantage. The impact of other factors (i.e., the null effect of immigration) is less conclusive regarding this claim. Thus, our findings signal a need for additional research on the mechanisms by which broad social conditions influence different types of violence.

Finally, our results have other implications for future violence research. For example, the current study illustrates the benefits to theory testing of employing data disaggregated by the race/ethnicity of both offenders and victims. The use of such data casts new light on neighborhood-level intergroup dynamics and informs our understanding of neighborhood processes. This approach should be taken with a variety of problematic neighborhood outcomes including additional types of crime. This study also draws attention to the salience of introducing Latinos as offenders and victims and should encourage others to examine neighborhood-level intragroup/intergroup Latino violence. In the end, as the American urban landscape changes, incorporating Latinos and other racial/ethnic groups in complex and nuanced ways into research on the interrelationships between place, social structure, and crime will broaden and deepen our understanding of these social phenomena.

Biographies

Jeffrey M. Cancino is an associate professor in the Department of Criminal Justice at Texas State University– San Marcos. Recent publications have appeared in Social Science Quarterly, Journal of Ethnicity in Criminal Justice, Journal of Criminal Justice, and Homicide Studies. His current research focuses on Latino violence in San Antonio.

Ramiro Martinez Jr. is a professor of criminal justice at Florida International University. He received the American Society of Criminology’s Division on People of Color and Crime’s Coramae Richey Mann Award, the Florida International University Faculty Award for Excellence in Research, and the American Sociological Association Latina\o Section Award for Distinguished Contributions to Research. He authored Latino Homicide: Immigration, Violence and Community (Routledge 2002) and edited Immigration and Crime: Race, Ethnicity, and Violence (New York University Press 2006).

Jacob I. Stowell is an assistant professor in the Department of Criminal Justice and Criminology at the University of Massachusetts, Lowell. His work has appeared in journals such as the American Journal of Sociology, Social Science Quarterly, and Aggression and Violent Behavior. His current research focuses on the temporal influence of immigration on patterns of lethal and nonlethal violence.