RACE, CODE OF THE STREET, AND VIOLENT DELINQUENCY: A MULTILEVEL INVESTIGATION OF NEIGHBORHOOD STREET CULTURE AND INDIVIDUAL NORMS OF VIOLENCE

Abstract

The study outlined in this article drew on Elijah Anderson’s (1999) code of the street perspective to examine the impact of neighborhood street culture on violent delinquency. Using data from more than 700 African American adolescents, we examined 1) whether neighborhood street culture predicts adolescent violence above and beyond an adolescent’s own street code values and 2) whether neighborhood street culture moderates individual-level street code values on adolescent violence. Consistent with Anderson’s hypotheses, neighborhood street culture significantly predicts violent delinquency independent of individual-level street code effects. Additionally, neighborhood street culture moderates individual-level street code values on violence in neighborhoods where the street culture is widespread. In particular, the effect of street code values on violence is enhanced in neighborhoods where the street culture is endorsed widely.

Scholars have devoted renewed interest to the effects of neighborhood context on problem behaviors (Sampson and Lauritsen, 1994; Sampson, Morenoff, and Gannon-Rowley, 2002). Research in this area has shown that neighborhood context influences a variety of outcomes, including victimization (Rountree, Land, and Miethe, 1994), delinquency (Bursik and Grasmick, 1993; Simcha-Fagan and Schwartz, 1986; Simons et al., 2005; Wikström and Loeber, 2000), and violence (Bellair and McNulty, 2005; Morenoff, Sampson, and Raudenbush, 2001; Sampson, Raudenbush, and Earls, 1997; Silver, 2000). This line of research has generated an extensive body of literature underscoring the significant impact that neighborhood structural context has on an array of outcomes (Bursik and Grasmick, 1993; Sampson, Morenoff, and Gannon-Rowley, 2002).

Despite this sizable body of literature underscoring the importance of neighborhood context, research investigating the effect of neighborhood street culture on violence in disadvantaged neighborhoods has been understudied. This gap in the research literature exists despite the increased interest in neighborhood effects generated by Wilson’s (1987) work on the concentration of poverty in disadvantaged neighborhoods. Thus, fundamental questions about neighborhood street culture’s influence on violence remain unanswered. In particular, does neighborhood street culture influence violent delinquency among African American adolescents, net of individual-level street code values? Also, does neighborhood street culture moderate individual-level street code values on violent delinquency among African American adolescents? These questions are paramount to Elijah Anderson’s (1994) “code of the street” thesis derived from his ethnography of inner-city neighborhoods. Anderson (1999: 33) defined the neighborhood street culture as “a set of informal rules governing interpersonal public behavior, particularly violence… [which provide] a rationale allowing those who are inclined to aggression to precipitate violent encounters in an approved way.” Additionally, Anderson (1999: 68–72) viewed adoption of the street code as an individual-level process in which one either fully or partially embraces the norms of the street culture, “or at least learns to comport themselves in accordance with its rules.”

In Code of the Street, Elijah Anderson (1999) outlined a multilevel process in which macrostructural patterns of disadvantage, racial inequality, and limited economic opportunities foster a street culture that is conducive to violence (Bernard, 1990; Krivo and Peterson, 1996; Massey and Denton, 1993; Peterson and Krivo, 1993; Wilson, 1987, 1996). In particular, Anderson (1994) suggested that the aforementioned structural conditions lead to a sense of hopelessness and cynicism about societal rules and their application, thereby resulting in a street culture that undermines mainstream conventional norms (Bruce, Roscigno, and McCall, 1998; Horowitz, 1983; Hughes and Short, 2005; Kubrin and Weitzer, 2003; Miller, 1958; Sampson and Wilson, 1995). He argued that the street culture is an ecological construct that is an emergent property of structurally disadvantaged neighborhoods and that it shapes values that influence violence among adolescents. In this sense, Anderson (1994) contended that a neighborhood contextual component to the street culture serves as an “institutional feature” of disadvantaged neighborhood street life that structures public interactions, especially violence (Anderson, 1999; Matsueda, Drakulich, and Kubrin, 2006). By treating the street culture as a neighborhood-level property, Anderson (1994: 82) argued that the source of violence is not only an individual-level process in which one adopts the street code but also an ecological one that is embedded in the broader social context. Matsueda, Drakulich, and Kubrin (2006: 340) noted that the “code of the streets” is an objective property of the neighborhood that operates above and beyond compositional characteristics of the residents (see also Cao, Adams, and Jensen, 1997: 375–6; Fischer, 1995: 547). Inherent in Anderson’s (1994: 82) model is the notion that individuals’ attitudes, beliefs, and behaviors are, in part, a function of the neighborhood context in which they are positioned. Indeed, Oliver (2006: 919) argued that the street culture, or the “streets,” is an “institution that is as important as the family, the church, and the educational system…” For some residents, the street culture influences individual’s “psychosocial development,” “life course trajectories” (Oliver, 2006), and “cognitive landscapes” (Sampson and Bartusch, 1998: 800) that shape individual beliefs, values, and behaviors. Yet, Anderson’s (1994, 1999) neighborhood street culture and violence proposition has not been the subject of many investigations, even though his argument attempted to explain the structural and cultural foundations of violence in disadvantaged neighborhoods.

Many studies in this area have relied on large macrolevel designs and have contained no explicit measures of a locale’s commitment to violence (Ellison and McCall, 1989; Gastil, 1971; Hackney, 1969; Loftin and Hill, 1974; Messner, 1983; Parker and Smith, 1979; Reed, 1971). Instead, these aggregate-level studies have assumed a locale’s commitment to violence based on a significant relationship observed between violence rates and racial composition or regional indicator variables, with little attention directed at individual-level attitudes about violence. Moreover, other studies that use commitment to violence as a predictor of offending have focused almost exclusively on microlevel analyses (Ball-Rokeach, 1973; Doerner, 1979; Erlanger, 1974; Heimer, 1997; Markowitz and Felson, 1998; Matsueda, 1989). These studies rarely have focused attention on a locale’s neighborhood street culture, which is a property of disadvantaged contexts that could influence violence either directly, indirectly, or as a moderating interaction between individual-level correlates (Anderson, 1994; Ousey and Wilcox, 2005; Simcha-Fagan and Schwartz, 1986).

Although neighborhood street culture’s influence on violence has been understudied, a handful of quantitative assessments have been conducted on Anderson’s thesis that examined whether neighborhood structural inequality gives rise to individual-level street code values that legitimatize the use of violence. For example, using a sample of approximately 700 African American adolescents, Stewart and Simons (2006) found that neighborhoods with high levels of structural disadvantage and violence led adolescents to adopt street code values in which they believed that the use of interpersonal violence is justified to gain respect. Additionally, the adoption of individual-level street code beliefs was predictive of violence, even after controlling for neighborhood disadvantage, family characteristics, and experiences with racial discrimination. In another study, using the same data set noted in the preceding discussion, Stewart, Schreck, and Simons (2006) observed that adolescents who lived in violent neighborhoods and embraced street code values were at a heightened risk of being involved in violence as a victim. Specifically, these researchers found that adoption of individual-level street code values increased the risk of violent victimization and that this risk was amplified for those adolescents who resided in high crime settings (Stewart, Schreck, and Simons, 2006). Similarly, Stewart, Simons, and Conger (2002) found that individual-level adoption of street code values was an important predictor of violence among the aforementioned sample of African American youth. The effect of individual-level street code values was significant even after accounting for the level of neighborhood social problems and socioeconomic status within the neighborhood. Furthermore, drawing on a national sample of youth, Brezina et al. (2004) found that disadvantaged urban youth who held street code beliefs were more likely to engage in violence. These studies have offered insight into the relationship between individual-level street code beliefs and violence. However, what is absent from these studies is a consideration of how contextual variation in neighborhood street culture, which is a property of some structurally disadvantaged settings, influences violence above and beyond individual street code values as well as other compositional factors.

Despite theoretical arguments that explicitly advance an ecological and structural approach to an oppositional street culture that influences violence (Anderson, 1990, 1994; Baumer et al., 2003; Cloward and Ohlin, 1960; Cohen, 1955; Horowitz, 1983; Shaw and McKay, 1969), few attempts have been made to measure the presence of a neighborhood-level oppositional street culture, although it has implications for the causes of violence in structurally disadvantaged settings. As Sampson and Bartusch (1998: 780–1) pointed out, “remarkably little research in this area takes a contextual perspective,” thereby failing to situate oppositional norms of violence as part of the broader neighborhood context. Thus, our knowledge in this area is limited. Even less pronounced in this literature is a direct conceptualization and test of neighborhood street culture’s influence on violence as Anderson (1999) proposed. The lack of attention to neighborhood street culture is surprising because Anderson’s propositions have provided an opportunity to extend our understanding of possible neighborhood processes that influence African American adolescent violence in disadvantaged areas.

In the current article, we contributed to the literature on neighborhoods and crime by addressing the above shortcomings. We assessed two hypotheses derived from Elijah Anderson’s (1999) code of the street perspective. First, we examined whether neighborhood street culture increases an adolescent’s violence after controlling for his or her personal commitment to the code of the street. Second, we explored whether neighborhood street culture moderates individual-level street code values on adolescent violence. We evaluated these hypotheses with longitudinal data from the Family and Community Health Study (FACHS), which consists of 763 African American adolescents from 71 neighborhoods. These data provide a unique opportunity to examine the effect of neighborhood street culture’s influence on violence while controlling for individual-level characteristics across a wide variety of ecological contexts.

NEIGHBORHOOD STREET CULTURE AND VIOLENCE

In Code of the Street, Elijah Anderson (1999) discussed the complexities of inner-city life for residents in structurally disadvantaged neighborhoods. He highlighted the structural and cultural factors leading to violence. Anderson (1999) argued that the high rates of poverty, joblessness, violence, racial discrimination, alienation, mistrust of police, and hopelessness that characterize many disadvantaged settings have led to a neighborhood street culture (see also Bruce, Roscigno, and McCall, 1998; Jargowsky, 1997; Kubrin and Weitzer, 2003; Oliver, 1994; Sampson and Wilson, 1995). The street culture is an objective property of structurally disadvantaged, disorganized, and socially isolated neighborhoods, and it serves to influence individual-level public interactions (Anderson, 1999: 145). This oppositional street culture emanates from the poverty of innercity neighborhoods and its ever-present threat to survival, and it espouses violence in particular situations. With the negative structural changes in the inner-city, disadvantaged neighborhoods, “the trust and perceptions of decency that once prevailed in the community are increasingly absent” and are replaced by “street values, represented by… violence and crime…” (Anderson, 1999: 145). Thus, the street culture articulates powerful norms that govern interpersonal public behaviors, especially with regard to violence. As such, the neighborhood street culture outlines the proper way to present oneself in a manner that demands respect and deters acts of victimization from others as well as to negotiate respect on the street (Anderson, 1999: 10).

The street culture emerges as an adaptation to adverse economic conditions and as a means of acquiring status and self-worth in disadvantaged areas. Anderson (1999) argued that much of the violence among disadvantaged inner-city residents revolves around the desire to acquire status by being perceived as “tough” or “violent.” At the heart of the street culture is an emphasis on respect, toughness, retribution, and ultimately, violence. Anderson suggested that the street culture emphasizes maintaining the respect of others, toughness, and exacting retribution when someone disrespects you through the use of violence. In this context, the street culture maintains that a violent response when someone initiates disrespect is necessary to elicit respect from others who follow the street code (Baron, Kennedy, and Forde, 2001; Hughes and Short, 2005; Luckenbill, 1977). When challenges develop or transgressions occur, violence is viewed as acceptable, appropriate, and even obligatory (Anderson, 1999; Cohen and Nisbett, 1994; Hughes and Short, 2005; Luckenbill and Doyle, 1989; Rich and Grey, 2005; Wolfgang and Ferracuti, 1967). Put differently, the neighborhood street culture rewards the use of toughness, aggressiveness, risk taking, and violence because it allows residents to achieve social status among peers.

An additional complication for residents in structurally disadvantaged areas where the street culture is manifested is the limited access to law (Anderson, 1999; Black, 1983; Kubrin and Weitzer, 2003; Oliver, 1994; Sullivan, 1989; Vélez, 2001). Anderson (1999: 66) suggested that residents in disadvantaged neighborhoods view the criminal justice system as unfair and discriminatory against minorities (Tyler and Huo, 2002; Unnever, 2008). Thus, residents often are reluctant to contact the police because of negative police and citizen interactions (Brunson, 2007; Carr, Napolitano, and Keating, 2007; Weitzer, 1999). As Carr, Napolitano, and Keating (2007) observed, youth in high-crime neighborhoods reported being stopped for no good reason, harassed, treated roughly, as well as encountering dishonest and lackadaisical police, which led to beliefs of procedural injustice and cultural attenuation (Sunshine and Tyler, 2003; Warner, 2003). Anderson (1999: 10) noted that the street culture emerges where the influence of the police ends and personal responsibility for one’s safety begins, which leads residents to resort to a kind of “people’s law” based on “street justice”. As Anderson (1999: 307) pointed out, the code maintains that, “a man goes for himself, takes up for himself, and calls on no one else to fight his battles.” Consequently, disputes are settled informally, violently, and without the intervention of responsible authorities like parents, police, or teachers (Black, 1983).

In sum, Anderson’s argument suggested that variation exists across settings in the street culture that influences violence. However, few studies have assessed a neighborhood-based oppositional street culture on violence in disadvantaged settings despite Anderson’s presentation of a compelling explanation. Although little research directly has addressed the relevance of Anderson’s (1999) propositions with regard to the interplay between neighborhood street culture and violence, we drew from several pieces of work. Most notable in this regard was the work of Sampson and Bartusch (1998) who found no support for neighborhood-level variation in tolerance for deviance by race. However, they did find evidence of “legal cynicism” in disadvantaged neighborhoods where residents were ambivalent toward the law and social norms in the context of extreme structural disadvantage. In a related study, Warner (2003) found that conventional cultural strength measured at the neighborhood level was an important predictor of informal social control. Furthermore, she observed that conventional cultural strength was attenuated in structurally disadvantaged neighborhoods with active drug markets. Although these studies did not assess how neighborhood-level oppositional and conventional cultural norms influenced violence, they did suggest that more disadvantaged settings have difficulties sustaining important prosocial norms (Bursik, 1988; Bursik and Grasmick, 1993).

In one of the few studies to assess neighborhood subculture and violence, Browning, Feinberg, and Dietz (2004) found that a neighborhood-based tolerance of deviance, which tapped whether residents tolerated teenagers smoking cigarettes, using marijuana, drinking alcohol, and fighting, did not significantly predict violent victimization or homicide rates. Moreover, three multilevel studies have been performed that contextualize school culture’s effect on violence, net of individual-level violent attitudes (Bernburg and Thorlindsson, 2005; Felson et al., 1994; Ousey and Wilcox, 2005). These studies of school culture are important because they represent one of the staging areas (besides the neighborhood) in which adolescents can display code-related behaviors (Brunson and Miller, 2009). However, the results of these studies are mixed with regard to contextualized school culture and violent beliefs as predictors of violence and delinquency. For example, Ousey and Wilcox (2005) found that individual-level violence values were predictive of violence, whereas school culture of violence was not. In their classic study, Felson et al. (1994) found that high schools with a culture of violence, as well as individual attitudes about violence, were predictive of violent offending among high-school males. Similarly, Bernburg and Thorlindsson (2005) observed that both school culture norms of violence, as well as individual-level norms, were predictive of violence.

Although not directly assessing Anderson’s street culture and violence proposition, the aforementioned studies do highlight the importance of contextualizing oppositional norms in various kinds of environments (i.e., neighborhoods and schools). In the current article, we followed the lead of these studies and examined a neighborhood-based street culture of violence. Based on Anderson’s (1999) arguments, our first hypothesis was that neighborhood street culture would be related significantly to violent delinquency above and beyond individual-level street code values. We examined this idea while controlling for several other individual-level characteristics as well as neighborhood factors such as economic disadvantage and the level of homicide.

NEIGHBORHOOD STREET CULTURE, STREET CODE VALUES, AND VIOLENCE

As noted, Anderson (1999) distinguished between neighborhood street culture, which is a property of structurally disadvantaged settings, and the adoption of street code values, which is an individual-level process. Although the neighborhood street culture provides a set of informal rules that shape public interactions around violence, the adoption of street code values represents individuals embracing values, beliefs, and behaviors that are consistent with the rules or norms of the neighborhood street culture. Anderson’s thesis suggested that residents’ beliefs and behaviors are likely to be structured and influenced by the neighborhood environment in which they are situated and live their day-to-day lives.

Anderson (1994: 82) described the conflict between the following types of cultural orientations in structurally depressed inner-city neighborhoods: “decent” and “street.” The decent cultural orientation embraces mainstream values, whereas the street cultural orientation constitutes an oppositional culture “whose norms are often consciously opposed to those of mainstream society.” In structurally disorganized areas, decent cultural orientations clash with street culture orientations where violence and disorder are an expected part of everyday life (Anderson, 1978; Kobrin, 1951; Kornhauser, 1978; Liebow, 1967; Shaw and McKay, 1969). Although only a small fraction of the population fully embrace the street culture orientation, it applies to everyone.

These arguments suggest that in a disadvantaged and racially isolated context where a neighborhood street culture is widespread, adolescents need to acquire knowledge consistent with the street culture for survival and defensive purposes (Anderson, 1994: 82). This conclusion indicates that “knowledge of the code is thus largely defensive, and it is literally necessary for operating in public” (Anderson, 1999: 33). As Anderson (1999) noted, many youth come from “decent” families and are committed to mainstream values that place little importance on violence as a primary goal. However, even though they personally are not committed to the street code, they must become familiar with and responsive to the rules of the street culture when navigating a dangerous and violent social context (Anderson, 1999: 92). As Sharkey (2006) argued, street efficacy is important for navigating in dangerous neighborhoods and for avoiding violence. Sharkey (2006) found that adolescents who showed a high level of street efficacy or street knowledge could avoid violence and delinquent peers. Thus, having knowledge of the streets is necessary for operating in these environments, and the conditions that encourage adoption of the code are pronounced in disadvantaged communities (Anderson, 1999; Hughes and Short, 2005; Kubrin, 2006; Stewart and Simons, 2006; Wilkinson, 2003; Wilson, 1987).

According to Anderson (1999), the adoption of street code values is important in neighborhoods where the street culture orientation system is widely prevalent. Consequently, individuals develop a range of social identities that are consistent with the street culture to manage the threats and demands of a context maintained by violence (Anderson, 1999; Fagan and Wilkinson, 1997; Oliver, 1994). The more violent one’s social identity is, the more respect and street credibility he or she has among peers following the street culture (Kubrin, 2005). Indeed, Anderson (1999: 131) observed that adolescents often created altercations with the primary focus of building respect on the streets and of letting others know they are not a chump. For example, Wilkinson (2003) found that young men committed robberies as a way to build or maintain street credibility and to impress their peers. As more people engage in street culture-related behaviors, the level of violence escalates, and the number of people who rely on violence for defensive purposes increases (Anderson, 1999; Canada, 1995; Fagan and Wilkinson, 1997).

Those individuals who follow the street culture must be ready to use violence and let others on the street know that transgressions have consequences (Felson, 1982; Tedeschi and Felson, 1994). To let transgressions go unchallenged—even small ones—demonstrates that one is soft and weak; therefore, all transgressions must be avenged (Anderson, 1999; Courtwright, 1996; Hughes and Short, 2005; Jacobs, 2004; Jacobs and Wright, 2006; Rich and Grey, 2005). As Jacobs (2004: 297) pointed out, “word on the street travels fast and reputational damage can be severe and longlasting,” leading one to pay the price of diminished respect and credibility on the street. In their study of homicide victimizations in disadvantaged St. Louis neighborhoods, Kubrin and Weitzer (2003) observed that a disproportionate number of homicides were retaliatory, based in part on street code values. They found that offenders were encouraged by neighborhood street culture norms to retaliate or risk losing respect among peers. Additionally, narratives from individuals involved in the homicides suggested that disputes and disrespect had to be settled in a violent street manner.

The ideas presented by Anderson (1999) also suggested that the neighborhood street culture moderates the effect of individual-level street code values on violence. Anderson documented how the neighborhood street culture creates a fertile backdrop in which street code values that influence violence are created, strengthened, and reinforced through social interactions in neighborhoods where the street culture is endorsed intensely. As a result, adolescents are more likely to come into contact with others in their neighborhood who also are influenced by the street culture, thereby reinforcing and strengthening one’s personal commitment to street code values that lead to violence. In this vein, adolescents residing in such neighborhood contexts where the street culture is accepted widely and street code values are adopted widely are likely to see violence as a way of life. This is especially the case for street-oriented adolescents who dominate public space. Indeed, street-oriented individuals enforce and reinforce the rules of the street culture through various interactions within the neighborhood and place emphasis on toughness and violence. Adolescents who are committed strongly to the street code are at high risk for engaging in violence. However, Anderson (1994, 1999) suggested that these individuals are especially violent when they are residing in areas where the street culture is highly prevalent and that aggression, therefore, is accepted widely. In other words, neighborhood street culture tends to amplify the violence-provoking effect of personal commitment to the street code. Following this logic, in our second hypothesis, we expected the effect of street code values to be associated strongly with violent delinquency in settings where strong evidence is found of a neighborhood street culture.

DATA AND METHODS

SAMPLE

Our research was based on waves 1 and 2 of the FACHS, a multisite (Georgia and Iowa) investigation of neighborhood and family effects on health and development (Simons et al., 2002). FACHS was designed to identify neighborhood and family processes that contribute to African American children’s development in families living in a wide variety of community settings. To facilitate this objective, sample members were recruited from neighborhoods, defined here as census tracts, that varied on demographic characteristics, specifically racial composition (i.e., percent Black) and economic level (i.e., percent of families with children living below the poverty line).¹ Specifically, using 1990 census data, tracts were identified for both Iowa and Georgia in which the percent of African American families was high enough to make recruitment economically practical (10 percent or higher), and in which the percent of families with children living below the poverty line ranged from 10 to 100 percent. From these criteria, 71 usable census tracts were identified, and the FACHS sample was selected from these areas.² In Georgia, families were selected from 36 census tracts from metropolitan Atlanta areas, such as South Atlanta, East Atlanta, Southeast Atlanta, and Athens, that varied in terms of economic status and ethnic composition. In Iowa, the 35 census tracts that met the study criteria were located in two metropolitan communities—Waterloo and Des Moines. In both research sites, families were drawn randomly from rosters and contacted to determine their interest in participation.³ Interviews were completed with 72 percent of the eligible Iowa families and with a little more than 60 percent of the eligible Georgia families who could be located, which is comparable with other community studies of families using intensive measurement procedures (Capaldi and Patterson, 1987; Conger and Elder, 1994). Respondents were reimbursed for participating in the study. Primary caregivers received $100, and target children received $70. The reimbursement levels reflected the different amounts of time required of each family member for participation.

Before data collection began, four focus groups in Georgia and four in Iowa examined and critiqued the self-report instruments. Each group was composed of 10 African American families who lived in neighborhoods similar to those from which the study participants were recruited. Group members suggested modification of items that they perceived to be culturally insensitive, intrusive, or unclear. After the focus groups’ revisions were incorporated into the instruments, the protocol was pilot tested on 16 families—8 from each site. Researchers took extensive notes on the pilot test participants’ reactions to the questionnaires and offered suggestions for more changes. The focus groups and pilot tests did not indicate a need for changes in any of the instruments used in the present article.

The first wave of the FACHS data was collected in 1997 from 867 African American children ages 10–13 years old (400 boys and 467 girls; 462 from Iowa and 405 from Georgia), their primary caregiver, and a secondary caregiver when one was present in the home. In the second wave of data, 763 children (ages 12–15) and their caregivers were interviewed again in 1999. Our analysis was based on 763 of these participants who had complete data on the variables of interest.⁴ We focused on waves 1 and 2 because this span was a period for escalating rates of delinquency (Moffitt, 1993; Sampson and Laub, 1993). Given the sampling design, these subjects represented a sample of African American youth from the two research sites that came from extremely poor-to-middle class families and who resided in neighborhoods that exhibited significant variability in economic status, racial composition, and other factors, which were sampling features that are well suited for studying neighborhood effects (e.g., Jencks and Mayer, 1990).

MEASURES

Dependent Variables

To take advantage of the longitudinal nature of the data and to increase confidence in temporal ordering, we predicted violent delinquency_T2 measured at time 2, while controlling for a previous measure of violent delinquency at time 1. Thus, our final model of table 1 provided a conservative test of the proposed research hypotheses because of the strong correlation between violent delinquency at time 1 and 2 (r = .63). As a result of incorporating a prior measure of the outcome variable, little residual variance remained in the dependent variable for other covariates to explain, which yielded very conservative estimates. We also estimated all independent variables at time 1. The descriptive statistics for the study variables are provided in appendix A.

Table 1.

Multilevel Models of Violent Delinquency_T2 Regressed on Neighborhood Street Culture, Street Code, and Controls

	Model 1			Model 2			Model 3			Model 4
	B	SE	t Ratio	B	SE	t Ratio	B	SE	t Ratio	B	SE	t Ratio
Neighborhood level
Street culture	.34*	.07	4.86	.29*	.07	4.14	.19*	.07	2.71	.13*	.06	2.17
Disadvantage	−	−	−	.15*	.05	3.00	.12*	.05	2.40	.10*	.05	2.00
Homicide rate	−	−	−	.16*	.06	2.67	.15*	.06	2.50	.12*	.06	2.00
Individual street code
Street code	−	−	−	−	−	−	.35*	.08	4.38	.22*	.08	2.75
Individual level
Family SES	−.10	.07	−1.43	−.10	.07	−1.43	−.10	.07	−1.43	−.10	.07	−1.43
Family structure (1 = two)	−.07	.05	−1.40	−.07	.05	−1.40	−.07	.05	−1.40	−.06	.05	−1.20
Target gender (1 = male)	.18*	.05	3.60	.18*	.05	3.60	.18*	.05	3.60	.13*	.05	2.60
Parental offending	.17*	.07	2.43	.17*	.07	2.43	.17*	.07	2.43	.15*	.07	2.14
Parent arrested	.03	.03	1.00	.03	.03	1.00	.03	.03	1.00	.03	.03	1.00
Adolescent monitoring	−.19*	.06	−3.17	−.19*	.06	−3.17	−.19*	.06	−3.17	−.16*	.06	−2.67
Violent peers	.22*	.06	3.67	.22*	.06	3.67	.21*	.06	3.50	.19*	.06	3.17
Prior arrest	.07	.05	1.40	.07	.05	1.40	.07	.05	1.40	.05	.05	1.00
School attachment	−.19*	.07	−2.71	−.19*	.07	−2.71	−.19*	.07	−2.71	−.17*	.07	.84
Urban (1 = urban)	−.02	.06	−.33	−.02	.06	−.33	−.02	.06	−.33	−.02	.06	−.33
South (1 = south)	−.04	.06	−.67	−.04	.06	−.67	−.04	.06	−.67	−.04	.06	−.67
Racial discrimination	.19*	.05	3.80	.19*	.05	3.80	.15*	.05	3.00	.11*	.05	2.20
Prior violent delinquency	−	−	−	−	−	−	−	−	−	1.92*	.22	8.73
Neighborhood variance explained		27.90%			42.80%			48.60%			62.40%

NOTE: N = 71 neighborhoods; N = 763 individuals.

^*p < .05.

Appendix A.

Descriptive Statistics and Correlations for the Study Variables

	Mean	SD	Violent DelinquencyT2 Correlation
Neighborhood level
Street culture	14.56	4.37	.22*
Disadvantage	9.51	4.07	.17*
Homicide rate	.15a	.47	.19*
Individual street code
Street code	16.15	3.64	.25*
Individual level
Family SES	13.41	4.14	−.05
Family structure (1 = two)	.52	.50	−.03
Target gender (1 = male)	.46	.50	.16*
Parental offending	2.92	3.45	.08*
Parent arrested (1 = arrested)	.20	.47	.19*
Adolescent monitoring	18.61	3.54	−.18*
Violent peers	4.21	1.72	.23*
Prior arrest (1 = arrest)	.11	.29	.14*
School attachment	28.41	5.48	−.15*
Urban (1 = urban)	.52	.48	.05
South (1 = south)	.49	.46	.08*
Racial discrimination	22.13	6.64	.12*
Prior violent delinquency	.17	.39	.63*
Dependent variables
Violent delinquencyT2	.28	.45	−

NOTE: N=71 neighborhoods; N=763 individuals.

^*p < .05,

^aper 1,000.

Violent delinquency_T2

The violence construct is dichotomous in nature and was measured at times 1 and 2 using eight questions that assessed violent offending. Respondents answered a series of questions regarding whether during the preceding year they had engaged in various violent acts such as physical assault, threatening others, bullying people, using a weapon, robbing others, and other aggressive behaviors in their neighborhood. If adolescents reported engaging in any of the aforementioned acts of violence during the past year, then they were coded 1 = engaging in behavior (0 = did not engage in behavior).⁵ Approximately 28 percent of the sample reported engaging in violent delinquency_T2 at time 2 and approximately 17 percent at time 1.

Independent Variables

Neighborhood street culture

A nine-item, self-report scale measured this construct at time 1. Primary caregivers were asked to indicate the extent to which street code values that support the use of violence operated in their neighborhoods (1 = strongly disagree to 4 = strongly agree). The questions included the following:

With reference to your neighborhood…

1. When someone disrespects you, it is important that you use physical force or aggression to teach him or her not to disrespect you.

2. If someone uses violence against you, it is important that you use violence against him or her to get even.

3. People will take advantage of you if you don’t let them know how tough you are.

4. People do not respect a person who is afraid to fight physically for his/her rights.

5. Sometimes you need to threaten people in order to get them to treat you fairly.

6. It is important to show others that you cannot be intimidated.

7. People tend to respect a person who is tough and aggressive.

8. Sometimes you have to use physical force or violence to defend your rights.

9. Arguing or fighting with other people usually makes matters worse rather than better (question recoded).⁶

By using primary caregivers’ reports, we could reduce the problem of shared method variance between individual-level street code and neighborhood street culture by using multiple informants to assess these study constructs (Lorenz et al., 1991).

To construct a neighborhood measure of street culture, we followed the procedure used by Raudenbush and Sampson (1999) for constructing neighborhood-level scales from individual-level survey-based data. We used a three-level hierarchical linear model in which the estimates from a multilevel measurement model adjust for measurement error and generate separate variance components for scale items, individuals, and neighborhoods. Thus, the level 1 model represented variation among scale items within individuals. The level 2 model represented variation among individuals within neighborhoods, and the level 3 model represented variation across neighborhoods. A neighborhood-level scale reliability was estimated, and in our case, the reliability is .72, indicating a reliable measurement of our construct (see Raudenbush and Bryk, 2002: 228–51).⁷

Neighborhood disadvantage

The following census variables were used to form this construct at time 1: proportion of households that were female-headed, proportion of persons on public assistance, proportion of households below the poverty level, proportion of persons unemployed, and proportion of persons who are African American. This construct reflected economic disadvantage in racially segregated African American neighborhoods. Previous studies have used some combination of these variables to assess community socioeconomic status (SES) (Baumer et al., 2000; Sampson, Raudenbush, and Earls, 1997). These variables are strongly correlated, and principal components and alpha factor analyses indicated that these variables loaded (>.72) on a single factor in our sample. The items were standardized and combined to form a measure of disadvantage. We added a constant (10) to the term that eliminated negative values. The alpha coefficient was .89.

Neighborhood homicide rate

This variable was measured using reported incidents of homicide from police records for each neighborhood in 1996 and 1997. Neighborhood homicide captures variation in the violent crime rate for each neighborhood. Homicide is a rare event, so we combined the years of 1996 and 1997 to reduce measurement error and to stabilize rates. We analyzed the violent crime rate per 1,000 neighborhood residents.⁸

Adopting the street code

A seven-item, self-report scale measured this construct at times 1 and 2. Adolescents were asked to indicate the extent to which it was justifiable or advantageous to use violence (1 = strongly disagree to 4 = strongly agree). The questions included the following:

1. When someone disrespects you, it is important that you use physical force or aggression to teach him or her not to disrespect you.

2. If someone uses violence against you, it is important that you use violence against him or her to get even.

3. People will take advantage of you if you don’t let them know how tough you are.

4. People do not respect a person who is afraid to fight physically for his/her rights.

5. Sometimes you need to threaten people in order to get them to treat you fairly.

6. It is important to show others that you cannot be intimidated.

7. People tend to respect a person who is tough and aggressive.

The responses were summed to obtain a total score concerning the extent to which the respondent held beliefs that were consistent with adopting a street code. The alpha coefficient was .78.

Control Variables

Family SES was measured by primary caregiver education level and family income. These two items were standardized and summed to form a composite measure of family SES. Family structure is a dichotomous variable denoting households in which two caregivers are present in the home, in comparison with single caregiver homes (1 = two caregiver family and 0 = one caregiver family). Target adolescent gender is a dichotomous variable in which males were assigned a value of 1. Parental offending is measured using 15 questions that asked primary caregivers whether during the preceding year they had engaged in various violent acts, such as physical assault, threatening others, and using a weapon. The alpha coefficient was .78. Parent arrested assesses whether the primary caregiver had been arrested in the past year. Adolescent monitoring was measured by five questions (answered by the primary caregiver) that focused on child monitoring (e.g., “How often do you know who your child is with when he/she is away from home?”). The alpha coefficient was .81. Violent peers was measured by three items adapted from the National Youth Survey (Elliott, Huizinga, and Menard, 1989), which asked respondents how many of their close friends had engaged in violent acts. We summed the responses to the items to obtain a total score regarding the extent to which the respondents’ friends engaged in violent behavior. The alpha coefficient for the scale was .68. Prior arrest assesses whether the target child had been arrested in the past year.

School attachment was measured by a 12-item scale that indicated the extent to which the respondents care about school and have positive feelings for school. The items were summed to create an index of school attachment. The alpha coefficient was .79. Urban is a dichotomous variable indicating neighborhoods located in urban areas with nonurban neighborhoods as the reference group. South is a dichotomous variable indicating neighborhoods located in the southern United States with Midwestern neighborhoods as the reference group designation. Racial discrimination uses 13-items to measure the extent to which adolescents were the victim of racial discrimination (Landrine and Klonoff, 1996). The items assess the frequency (1 = never to 4 = several times) with which various discriminatory events were experienced during the past year (i.e., “How often has someone yelled a racial slur or racial insult at you just because you are African American?” and “How often has someone threatened you physically just because you are African American?”). The alpha coefficient for the scale was .85.

ANALYTIC STRATEGY

We used multilevel modeling techniques to examine the effects of neighborhood- and individual-level factors on street code values and adolescent violence. Multilevel models were appropriate in this case because we were interested in individual outcomes that we hypothesized are affected by neighborhood- and individual-level characteristics. Multilevel modeling has become customary for estimating contextual effects when individuals are clustered within neighborhoods (Goldstein, 2003; Kreft and De Leeuw, 1999; Rabe-Hesketh and Skrondal, 2008; Raudenbush and Bryk, 2002). These models explicitly recognized that individuals within a particular neighborhood might be more similar to one another than individuals in another neighborhood and, therefore, might not provide independent observations.

In such situations, the residual errors likely are to be correlated within neighborhoods in nested data, which violates the assumption of independence of observations fundamental in traditional ordinary least-squares analysis (Raudenbush and Bryk, 2002). Consequently, failure to account for nonindependence of observations can result in standard errors that are biased downward, increasing the chances of reaching incorrect conclusions (Goldstein, 2003; Kreft and De Leeuw, 1999; Raudenbush and Bryk, 2002). Multilevel models avoid violating the assumption of independence of observations that the traditional ordinary least-squares analysis commits in analyzing hierarchical data, and they produce correct estimates of standard errors (Raudenbush and Bryk, 2002).

Furthermore, multilevel modeling techniques also allow for the simultaneous investigation of both individual- and neighborhood-level variance components on the outcome variable of interest (e.g., street code values or violent delinquency) while maintaining the appropriate level of analysis for the independent variables. To estimate our theoretical models, we used the multilevel function in the STATA 10 program (StataCorp, College Station, TX; Rabe-Hesketh and Skrondal, 2008).⁹

RESULTS

MULTILEVEL MULTIVARIATE ANALYSES

Although not shown, a preliminary step in multilevel analyses involved fitting an unconditional, random analysis of variance (ANOVA) model (i.e., a model with no predictors or control variables). We estimated an unconditional model for our dependent variable—violent delinquency_T2. The variance component was significant, indicating that violent delinquency varied significantly across neighborhoods. Furthermore, the estimated reliability for violent delinquency ranged from .67 to .65 across models. These results suggest that we can generate reliable true score estimates. In the analyses that followed, each of the individual-level covariates was grand mean centered. Doing so allowed us to estimate neighborhood contextual effects without the confounding impact of compositional effects (Britt, 2000; Haynie, Silver, and Teasdale, 2006).¹⁰

We now turn to one of our primary research hypotheses and assess whether neighborhood street culture predicts violent delinquency. In this series of models, we specified violent delinquency_T2 as the dependent variable. At the neighborhood-level, we assessed whether neighborhood street culture is a significant predictor of violence. If Anderson’s (1999) prediction is correct, then neighborhood street culture should influence violence above and beyond compositional effects. In model 1 of table 1, the findings suggest that neighborhood street culture significantly predicted violent delinquency. This finding suggests that violence is higher in neighborhoods where the street culture is more pronounced. In other words, adolescents who live in areas where the neighborhood street culture is well established are more likely to engage in violence. Although this finding is consistent with Anderson’s (1999) thesis, it is necessary to account for other predictors that might render this relationship spurious. In models 2 and 3 of table 1, a series of regressions are presented that account for additional neighborhood- and individual-level factors that might mediate the relationship between neighborhood street culture and violence.

Model 2 introduces two additional neighborhood predictors (disadvantage and homicide rate) that have been shown to be associated with violent offending.¹¹ Although neighborhood disadvantage and neighborhood homicide rate are associated significantly with violent delinquency, neighborhood street culture remains a strong and significant predictor of individual-level interpersonal violence. In model 3, we added individual street code values to the predictive equation. Once the street code measure was entered into the model, it was a significant predictor of violent delinquency and reduced some of neighborhood street culture’s effect on violence. In accordance with our expectations, though, neighborhood street culture remained a significant predictor of violence, suggesting that the street culture operates above the effects of individual street code values.

To assess whether our findings are biased by the omission of prior violent offending, we included an additional control variable—prior violent delinquency_T1. Model 4 shows that even after controlling for prior violence and other important covariates, neighborhood street culture and adoption of the street code remain significant predictors of violent delinquency. Adolescents who reside in a neighborhood context where the street culture is widespread increase their chances of offending by 14 percent ([e (.13) = 1.14 − 1.00] × 100). Adopting the street code increases the chances of violent offending by 25 percent. Furthermore, of all the control variables included in model 4, nine were significant. Adolescent males who are exposed to parental criminality, associate with violent peers, experience discrimination, have a history of offending, and live in neighborhoods characterized by social problems are more likely to engage in violence. However, adolescents who have parents that monitor their activities and who are attached to school are less likely to engage in violence. Overall, this model accounted for about 62 percent of the neighborhood variance.¹²

We now turn to our second research hypothesis, which tested whether neighborhood street culture moderates the effect of individual-level street code values on violence. We first estimated a random-coefficient model by allowing the slope for street code values to vary across neighborhoods while holding all other predictors fixed (Rountree, Land, and Miethe, 1994). The results reveal that significant variation exists across neighborhoods in adolescents’ street code values (variance component = .0691, χ² = 102.40, p <.05). To explain the variance in the street code slope, we incorporated a cross-level interaction term by specifying the slope for individual-level street code values as a function of neighborhood street culture. The cross-level interaction results are presented in model 1 of table 2. A significant and positive interaction coefficient was found between neighborhood street culture and street code on violence. This finding indicates that the effect of an adolescent’s street code values on violence is more intense in neighborhoods where the street culture is most pronounced and widespread.¹³

Table 2.

Multilevel Models of Violent Delinquency_T2 Regressed on Cross-Level Interaction

	Model 1
	B	SE	t Ratio
Neighborhood level
Street culture	.12*	.05	2.40
Disadvantage	.06	.05	1.20
Homicide rate	.08	.06	1.33
Individual street code
Street code	.15*	.07	2.14
Cross-level interaction
Street code × neighborhood street culture	.23*	.08	2.87
Individual level
Family SES	−.11	.07	−1.57
Family structure (1 = two)	−.10	.06	−1.67
Target gender (1 = male)	.13*	.05	2.60
Parental offending	.15*	.07	2.14
Parent arrested (1 = arrested)	.03	.03	1.00
Adolescent monitoring	−.16*	.06	−2.67
Violent peers	.19*	.06	3.17
Prior arrest (1 = arrest)	.05	.04	1.25
School attachment	−.17*	.07	−2.43
Urban (1 = urban)	−.02	.06	−.33
South (1 = south)	−.04	.06	−.67
Racial discrimination	.11*	.05	2.20
Prior violent delinquency	1.82*	.22	8.27
Neighborhood variance explained		62.10%

NOTE: N = 71 neighborhoods; N = 763 individuals.

^*p < .05.

To illustrate the interaction results, we plotted the predicted probabilities. In particular, we plotted the influence of street code values on adolescent violence at two distinct levels of neighborhood street culture. As depicted in figure 1, adolescents with high street code values and who live in high street culture neighborhoods have higher probabilities of engaging in violent delinquency.¹⁴ The predicted probability of violence is lower in settings where investment in the street code and the neighborhood street culture are lower than average. These results are consistent with Anderson’s (1999) predictions.

Figure 1.

The Effect of Street Code on Violent Delinquency by Neighborhood Street Culture

DISCUSSION

A substantial body of research has examined neighborhood context as a key predictor of adolescent problem behaviors (Sampson, Morenoff, and Gannon-Rowley, 2002). However, little empirical attention has been given to the effect of neighborhood street culture norms on African American adolescent violence and the possibility of its moderating effect on individual-level street code values. In the current study, we drew on Elijah Anderson’s (1999) code of the street perspective to address these issues. First, we examined whether residing in a neighborhood where the street culture is highly prevalent predicts adolescent violence beyond the effect of individual-level street code values. Second, we analyzed whether neighborhood street culture conditions individual-level street code values on adolescent violence. Our results were consistent with Anderson’s ethnographic description of the code of the street and adolescent violence.

Anderson (1997, 1999) described a multilevel process in which the code of the street is an emergent property of disadvantaged, violent, socially isolated neighborhoods that impacts violence by shaping street code values. Once the street culture is established, it serves as an “institutional feature” of disadvantaged neighborhood street life that structures individual-level public interactions around the code (Anderson, 1999; Matsueda, Drakulich, and Kubrin, 2006; Oliver, 2006).

Our analyses revealed support for this position. Consistent with our hypotheses and Anderson’s (1999) ethnographic description, our multilevel analyses indicated that neighborhood street culture is a significant predictor of violent delinquency. Although neighborhood street culture’s effect on violence was reduced when individual-level commitment to street code was added to the equations, neighborhood street culture maintained its significant effect on violence. This outcome is consistent with Anderson’s (1999) contention that following the code of the street is important in neighborhoods where this value system is prevalent. Persons interacting in an area where the street culture is accepted widely put pressure on residents to follow the street code to navigate this dangerous milieu (Anderson, 1999). As Felson et al. (1994) argued, a culture of violence places pressure on all members to use violence or seem ready to use violence regardless of their personal beliefs (see also Bernburg and Thorlindsson, 2005). When neighborhood street culture values are widespread, residents might perceive violence as a way to protect their social identity by earning or maintaining respect among relevant social actors (Bernburg and Thorlindsson, 2005; Ellison, 1991; Felson, 1978; Tedeschi and Felson, 1994). Thus, adolescent residents use violence to construct a presentation of self that is conducive with street expectations and is a mechanism to promote an image or reputation based on toughness and aggression (Fagan and Wilkinson, 1997; Markowitz and Felson, 1998; Wilkinson, 2003). Our finding that neighborhood street culture has a direct influence on violence above the effect of individual-level street code values is consistent with the idea that the constant threats to identity posed in neighborhoods where the street culture is pronounced pressures everyone to have knowledge of the code and to be ready to posture or to use violence to save face and protect oneself (Anderson, 1999).

Furthermore, we observed a significant interaction between neighborhood street culture and individual-level street code values that influences adolescent violence (Sampson and Lauritsen, 1994). This finding revealed that the positive effect of individual-level street code values on violent delinquency is increased when an adolescent lives in a neighborhood where the street culture is endorsed. This pattern of results adds to the growing literature that describes how neighborhood structural characteristics combined with deviant cultural and situational codes lead to the perpetuation of violence (Anderson, 1994; Bernard, 1990; Bruce, Roscigno, and McCall, 1998; Canada, 1995; Kubrin and Weitzer, 2003; Luckenbill and Doyle, 1989; Sampson and Wilson, 1995; Stark, 1987). Similar patterns have been observed in parallel settings such as schools (Bernburg and Thorlindsson, 2005; Felson et al., 1994) and prisons (Harer and Steffensmeier, 1996), indicating that context matters for shaping and moderating beliefs, values, and behavior patterns.

The broader implications of our research underscore the importance of neighborhood context on adolescent outcomes. We observed that neighborhood street culture had a direct influence (contextual) and a conditional influence on adolescent behavior. In this sense, our findings are in line with Swidler’s (1986) “culture in action” thesis, which argued that within a social structure, culture influences actions by shaping a repertoire or “tool kit” of habits, skills, and styles from which actors select differing pieces for constructing “strategies of action” (also see Hannerz, 1969). This theory provides a framework of knowledge for actors on what to do in different kinds of circumstances (also see Hannerz, 1969). Indeed, in Anderson’s (1999) code of the street thesis, he highlighted how street code values are moderated by various kinds of neighborhood structural and racial inequalities. And, in neighborhoods where the street culture is widespread and pronounced, this street culture puts pressure on everyone to conform. Thus, individuals are forced to draw on their repertoire or tool kit to determine a particular course of action within the context of a restricted range of options generated by an inequitable structural system.

Moreover, the results of our study are strongly indicative of an integrated structural and cultural perspective on adolescent violence. However, more research is required. Future research should explore how the presence of “old heads” (role models) mediate or moderate the effects of neighborhood street culture on adolescent violence in disadvantaged neighborhoods. Anderson (1999: 180) observed that male role models serve as a potential source of social control in the larger African American community. Indeed, Parker and Reckdenwald (2008) found that the presence of African American male role models lowered African American adolescent violence and mediated the effect of structural disadvantage on violence. These results suggest that African American male role models are important mechanisms of social control in stratified urban environments where the street culture is prevalent. The presence of traditional male role models to mediate neighborhood structural effects on violence raises a fundamental question: Can the presence of role models mediate the effect of neighborhood street culture on adolescent violence? Furthermore, future research also should incorporate more direct measures of intervening mechanisms between neighborhood street culture and violence.

Although our research generally is supportive of Anderson’s (1999) thesis, and we are confident in our results, our study has some limitations. First, because we focused on a single racial group (African Americans), we could not determine whether neighborhood street culture has a similar effect on violence across other racial/ethnic adolescent populations. Thus, we are limited in the generalizability of our results to other populations. Future research could examine whether racial and ethnic invariance exists across populations with regard to neighborhood street culture and violent delinquency. Second, in this study, we focused on nonlethal forms of violence. We could not capture lethal forms of violence such as homicide; thereby, it is unclear whether neighborhood street culture affects such outcomes. Thus, additional research is required between the code of the street and lethal violence (Kubrin and Weitzer, 2003). Third, given the difficulty in operationalizing some concepts, we cannot dismiss the possibility that better measures might yield different results. For example, we could not measure the more general questions about how respondents felt their neighbors would respond with regard to neighborhood street culture norms. Our questions were restricted to respondents’ own experiences with neighborhood street culture norms. Furthermore, our code of the street measure was limited to attitudinal and belief items as well as to our neighborhood street culture measure. It is difficult for us to prove definitively that the relationships among neighborhood street culture, street code, and violent delinquency are not capturing, in part, the influence of an omitted variable. Moreover, because we used two waves of data, we could not assess the long-term consequences of involvement in the street culture. Indeed, Hagan (1991) found that identification with a delinquent subculture has a negative effect on early adult trajectories that stratified labor market attainment for working-class males. This finding suggests that involvement in oppositional street cultures can undermine prosocial adult trajectories throughout the life course (Hagan, 1997). Unfortunately, we could not assess these stratified outcome possibilities in early adulthood with the current data. Despite these limitations, we believe that our findings are plausible in that adolescents following the street culture have a higher probability of offending.

In sum, the results of our study support Anderson’s (1999) ethnographic, multilevel code of the street thesis on adolescent violence. We hope that our findings help shed light on a complex issue. In addition, we hope that the current study will encourage future investigations into the various ways that the structure and culture of neighborhoods affect adolescent violence in African American communities.

Appendix B.

Multilevel Spatial Regression Models of Street Code_T2 and Violent Delinquency_T2 Regressed on Neighborhood Street Culture and Street Code

	Model 1 Violent DelinquencyT2			Model 2 Violent DelinquencyT2
	B	SE	t Ratio	B	SE	t Ratio
Neighborhood level
Street culture	.12*	.05	2.40	.10*	.05	2.00
Disadvantage	.09*	.04	2.25	.05	.05	1.00
Homicide rate	.11*	.05	2.20	.07	.06	1.17
Spatial lag	.12*	.06	2.00	.10*	.05	2.00
Individual street code
Street code	.20*	.09	2.22	.14*	.05	2.80
Cross-level interaction
Neighborhood street culture       ×street code	—	—	—	.24*	.08	3.00

NOTE: N = 71 neighborhoods; N = 763 individuals.

^*p < .05.

Biographies

Eric A. Stewart is an associate professor of criminology and criminal justice at Florida State University and a member of the Racial Democracy, Crime and Justice Network. His research interests include racial inequality and criminal outcomes, crime throughout the life course, and multilevel processes on adolescent development.

Ronald L. Simons is the Distinguished Professor of Research in the Department of Sociology and a research fellow in the Institute for Behavioral Research at the University of Georgia. Much of his research has focused on the manner in which family processes, peer influences, and community factors combine to influence deviant behavior across the life course. He also has completed work on domestic violence and the effect of racial discrimination on child development.