Abstract
This study examines the effects of adherence to street codes on school-based violent victimization. In doing so, it separates street values into two distinct orientations: (1) retaliatory norms, and (2) general toughness norms. We analyze survey data from students across four waves of the Rural Substance Abuse and Violence Project using two-level mixed effects Poisson regression. The model specifies random intercepts across individuals to address possible endogeneity of the repeated measurements per respondent. Further, consistent with a fixed-effects approach, we control for differences across the schools in which students were nested. Findings reveal that adherence to retaliatory norms is negatively related to the frequency of serious violent victimization. In contrast, general toughness norms are positively related to serious violent victimization.
Examination of data from the United States indicates that, overall, the school is a key setting for understanding youth victimization. For example, Musu-Gillette et al. (2019, p. 32) present recent data from the National Crime Victimization Survey indicating that in 2017, students age 12 to 18 experienced higher rates of violent victimization at school than away from school (21 vs. 12 per 1,000, respectively). Though much school-based violent victimization appears relatively minor (i.e., simple assaults), serious victimization experiences can interfere with the academic achievement of all students (Juvonen et al., 2000; Schwartz, 2000). Furthermore, it can have profound effects on the mental and physical health of both victims and offenders. These effects include depression, anxiety, school avoidance, and feelings of insecurity—effects that, in turn, appear to increase the risk of maladaptive behavior problems, low self-esteem, loneliness, and suicide across different cultural and national groups (Boivin et al., 1995; Boulton & Underwood, 1992; Boulton & Smith, 1994; Brain, 1997; Rigby, 1998; Schwartz et al., 1998). Therefore, understanding ways to tackle violent victimization at school is, rightfully, a public concern.
In fact, this concern has generated a good deal of recent scholarship examining the correlates of school-based victimization. This work emphasizes the importance of measures such as low self-control, aggressive or deviant behavior, delinquent peer associations, and extracurricular activities in enhancing risk of school-based victimization (Augustine et al., 2002; Burrow & Apel, 2008; Deryol et al., 2017; Peguero & Popp, 2012; Schreck, 1999; Schreck & Fisher, 2004; Schreck et al., 2006; Tillyer et al., 2010, 2011; Turanovic & Pratt, 2014; Turanovic et al., 2015; Wilcox et al., 2009; Zaykowski & Gunter, 2012). In contrast, very few studies have explored the effects of adherence to street codes in the examination of school victimization despite the fact that Anderson’s (1999) ethnographic study suggests that adherence to such codes might be related to both in-school and outside-of-school violent victimization (for an exception see Schreck et al., 2012).
Thus, this study seeks to better understand whether and how adherence to values overlapping with Anderson’s “code of the street” relates to school-based adolescent serious violent victimization. It does so by considering how different types of street codes potentially indicate various aspects of target congruence, and thus relate to victimization risk in unique ways. To that end, the research reported here separates street values into the two distinct street code orientations: (1) adherence to retaliatory norms and (2) adherence to norms embracing “toughness” more generally. Consistent with some of Anderson’s (1999) claims and target congruence theory (Finkelhor & Asdigian, 1996), we suggest that these two distinct orientations might be differentially related to serious violent victimization at school. In particular, we posit that the relationships work in opposite directions, with retaliatory norms negatively related to serious violent victimization and general toughness norms positively related. The relationships among different street codes, risky lifestyles, and serious violent victimization are modeled using multilevel mixed effect Poisson regression analysis with data from 3,840 students within 111 unique school contexts over the course of a four-wave study.
The Code of the Street
In his ethnographic work in inner-city Philadelphia, Anderson (1999) argued that in structurally disadvantaged African American neighborhoods—where many people suffer from persistent poverty, family disruption, and social isolation—there exist norms that are often opposed to mainstream values. Yet, Anderson also recognized that the degree to which the residents in such neighborhoods were detached from mainstream society varied, thus leading to individual variation in the adoption of the code (see also Stewart & Simons, 2006, 2010). According to Anderson, there are two orientations—“decent” and “street”—that exist within neighborhoods like those he observed. In “street families,” parents “may aggressively socialize their children into it [the code] in a normative way” (Anderson, 1999, p. 45). Thus, in street families, young people are particularly vulnerable to becoming completely committed to the code of the street—to internalizing the values reflected in the code. In contrast, decent families are oriented more toward “decency and civility” (Anderson, 1999, p. 36). According to Anderson (1999), most residents in the community he observed were decent and they appreciated mainstream values. Children of decent parents thus learned to respect authority and display non-violent behaviors.
The street families and decent families that Anderson observed coexisted in very close proximity, so decent families also had to teach their children to take care of themselves when challenged. Regarding this issue, Anderson (1999) argued that some parents in decent families encouraged their children to avoid trouble by talking and walking away from those who were aggressive toward them as a tactic to avoid physical fighting. On the other hand, knowing that “turning the other cheek” was not always a winning strategy, many decent parents encouraged their children to stand their ground so that they were not targeted repeatedly for victimization. In other words, decent families would often encourage their children to temporarily adopt aspects of the code—to display toughness and aggression for self-protective purposes. Anderson described such temporary displays as “code-switching” on the part of decent individuals, and he implied that “code-switching” from decent to street was sometimes necessary for survival. Thus, though not completely inclined toward physical aggression, decent people tended to condone defensive or retaliatory violence situationally—as an appropriate response to a personal attack or threat thereof.
Anderson (1999, p. 93) also conveyed that the dynamics of the code exist not only on the disadvantaged street but also in the inner-city school setting, as “school becomes a primary staging area for the campaign for respect.” This staging area is where students, in essence, practice their street values in interactions with one another. For example, it is the school setting where the most deprived kids—those who are likely most committed to the street values—might try to gain respect by putting peers down with verbal harassment and physical violence (Anderson, 1999). Other work confirms the importance of school as a staging area. For example, a student interviewed by Brunson and Miller (2009, p. 197) indicated that violence at school often stemmed from “like who runs it, like you know, who’s dominant between everybody” . . . “it’s like a showcase here, you know, a lot of people, they just want people to watch ‘em” (Brunson & Miller, 2009, p. 197). And, just like on the street, a student should not show fear to someone who wants to mess with her/him in the school’s “staging area.” In this regard, Brunson and Miller’s (2009, p. 197) interviewees indicated that student bystanders often amplify conflicts between classmates by clearly expressing expectations for retaliation when hit: “people will start crowding around. . .[egg] stuff on” . . . “cheer [it] on. Be like, ‘man, you let him whoop you like that? Get up and hit him back!’” (see also Miller, 2008; Swartz, 2012; Swartz et al., 2017). In sum, based on Anderson’s (1999) claims as well as support from subsequent qualitative research, the informal rules of the street also permeate schools, and adherence to the code can thus affect student behavior and experiences in the school setting. In fact, limited previous quantitative research and meta-analyses have also indicated that adherence to street codes is relevant for understanding school-based offending and victimization among students in schools, including those beyond the inner city (Felson et al., 1994; McGloin et al., 2011; Moule & Fox, 2020; Ousey & Wilcox, 2005; Schreck et al., 2012; Swartz et al., 2017).
Retaliatory Norms Versus General Toughness Norms
Street codes are commonly measured as a unidimensional construct, including in some previous studies using the same data analyzed herein (e.g., Ousey & Wilcox, 2005; Stewart et al., 2006; Stewart & Simons, 2006, 2010; Swartz et al., 2017; Taylor et al., 2010). Yet, there is also some evidence that street code measures exhibit weak internal validity, if combined (McNeeley & Wilcox, 2015). Consistent with the latter evidence, we explore further the possibility that previous research on the effects of street values has largely ignored the two distinct types of street values implicit in Anderson’s ethnographic study. First, as alluded to above, one aspect of Anderson’s code of the street involves retaliatory norms (RNs), which largely serve the purpose of self-protection. Sending a message that one will not back down in the face of attack presumably increases invulnerability to future victimization. In contrast, one who fails to react to threats can gain the reputation of a “chump”—and it is the chumps “who most often get picked on, tried or tested, and become victims of robbery and gratuitous violence” (Anderson 1999, p. 131). In fact, reaction to an initial victimization is presumed to help offenders identify victims who are likely to resist or not (see also Felson & Clarke, 1998; Kennedy & Baron, 1993).
In addition to retaliatory norms, with their potential self-protective function, Anderson (1999) also describes another aspect of the code of the street that we term “general toughness norms” (GTNs). According to Anderson, youths—particularly those in street families—learn generally aggressive ways through the observation of hostility and violence on the part of adults. Among those who internalize GTNs, displays of aggression, toughness, and sexual prowess are common ways to assert oneself in the “campaign for respect” that is particularly salient during adolescence (Anderson, 1999, p. 68). These deviant behaviors are only reinforced by peers as respect is won.
Thus, drawing from Anderson’s (1999) work, it seems that there are subtle but important differences between RNs and GTNs within the code of the street—not only in terms of the content of the norms but when and how they are used. The adoption of RNs is likely more situational than the adoption of GTNs. For example, one might approve of the norm of fighting to defend oneself from future victimizations (i.e., supportive of defensive or retaliatory violence), but might not approve of more indiscriminate aggression for the sake of having “juice.” This scenario aligns with Anderson’s description of “decent” individuals who must sometimes code switch and behaviorally express retaliatory norms for self-protection but who also typically adhere to conventional values. On the other hand, GTNs likely represent a more cynical view of law, and a deeper commitment to toughness and aggression, as a general way of life—as an underlying part of social identity—and thus align with Anderson’s description of “street” individuals.
Besides Anderson’s work on the code within disadvantaged urban neighborhoods, other work supports the idea that there are potentially two types of subcultural violent norms. For example, a study of violence among ex-offenders and the general population by Markowitz and Felson (1998) introduced two distinct types of attitudes as part of a “street-like” subculture—“attitudes toward retribution” and “attitudes toward courage.” They argued that these two orientations should be distinguished as they are subtly, yet importantly, different. Specifically, they argued that attitudes toward retribution were used to enact justice and deter future victimization. In contrast, attitudes toward courage were likely to be practiced to gain respect and protect one’s social identity. Thus, their description was quite consistent with RNs and GTNs described above as implicit in Anderson’s work (Markowitz & Felson, 1998; see also Tedeschi & Felson, 1994). Similarly, Agnew (1994) argued that measures of “general approval of violence” (without any particular reason or provocation) should be distinguished from measures of “approval of retaliatory violence.” Accordingly, in his study, he separated survey items tapping approval of general violence (“How wrong it is for someone your age to hit or threaten to hit someone without any reason”) from survey items tapping conditional approval of retaliatory violence (“Is it alright to beat up people if they started the fight or if they call your names”). Descriptive statistics using the sample from the National Youth Survey showed that very few respondents agreed with norms supporting general violence, whereas 39% of respondents agreed with situational retaliation—a disparity that supports the distinction between these two types of norms (Agnew, 1994). Finally, O and Wilcox (2018) examined eight school-level indicators of violent norms, created by aggregating student survey responses within schools. Two and six indicators, respectively, loaded distinctly (i.e., no cross-loading) on two different factors which represented concepts that the authors labeled “retaliatory norms” and “general violent norms.”
RNs and GTNs: Implications for Victimization
There is debate about whether Anderson’s work implies that adherence to street codes should be positively or negatively related to victimization—whether adopting the code promotes respect from would-be attackers, thus serving as self-protection, or whether adopting the code antagonizes would-be attackers, thus enhancing victimization risk (e.g., McNeeley & Wilcox, 2015; Stewart et al., 2006). The interpretation of this debate that we offer here is that these two seemingly contradictory processes are not necessarily mutually exclusive. Instead, the two distinct norms in the code of the street identified above—RNs and GTNs—might be differently related to victimization, with RNs protecting against victimization, and GTNs increasing risk of victimization.
These different effects of distinct types of street values make sense when viewed vis-à-vis target congruence theory (Finkelhor & Asdigian, 1996). In target congruence theory, Finkelhor and Asdigian (1996) argue that personal attributes congruent with offender motivations can affect target suitability. They suggest that target congruence has three dimensions: target vulnerability, target gratifiability, and target antagonism. First, individuals may be vulnerable due to physical characteristics (e.g., small size/stature, limited physical strength) or emotional/psychological traits (e.g., fearfulness, submissiveness) (Elvey et al., 2018; Kulig et al., 2017; O & Wilcox, 2018; Schwartz et al., 1993). Personal attributes might also affect a victim’s gratifiability, which refers to the extent to which a target has some “quality, possession, skill, or attribute that an offender wants to obtain” (Finkelhor & Asdigian, 1996, p. 6). Third, some individual attributes “antagonize” offenders—that is, they ignite anger, jealousy, or impulses in offenders, perhaps due to bias on the part of the offenders (Finkelhor & Asdigian, 1996, p. 6). Biases aside, personal attributes can antagonize offenders because they present noxious stimuli. For instance, the belligerence, short temper, and impulsiveness associated with low self-control might antagonize others, thus increasing risk for violent and property victimization, even net of risky lifestyle factors (Augustine et al., 2002; Deryol et al., 2017; Schreck et al., 2006; Turanovic & Pratt, 2014; Turanovic et al., 2015; Wilcox et al., 2009).
RNs as target invulnerability
Anderson’s description of the display of RNs, as an aspect of the overall code of the street, corresponds with target congruence theory’s dimension of target vulnerability. In particular, Anderson (1999, pp. 130–131) indicates that RNs encourage youths who are being targeted by others to “display heart to engage in a standoff” and that “the message that you are not a pushover must be sent loudly and clearly.” Anderson’s description thus implies that retaliatory reaction, or action that implies one could retaliate (i.e., displaying “retaliation potential”), sends a message of target invulnerability and is an attempt to defend an individual against victimization (see also, Stewart et al., 2006, for discussion).
GTNs as target antagonism and target gratifiability
Anderson (1999) suggested that when young people adopt GTNs, they embrace the idea that general displays of physicality, sexual prowess, aggression, and toughness go hand in hand with respect. In order to gain respect, they are constantly ready to hit others. However, “in turn, they are readily hit by other children” (Anderson, 1999, p. 69). In short, persistent (general) displays of toughness can serve to antagonize others, thus enhancing victimization risk. Additionally, Anderson (1999, p. 75) described a violent give-and-take constantly being played out on the streets in the race for respect, whereby “raising oneself up largely depends on putting someone else down” and that “a profound power transaction occurs” when someone who displays toughness is overtaken (Anderson 1999, p. 126). In this game of give-and-take, “tough” people are particularly gratifiable (i.e., respect-worthy) targets.
Beyond Anderson: RNs, GTNs, and Victimization
Additional qualitative work supports the idea that commitment to retaliatory norms might provide protection against victimization. For example, Baron et al. (2001) interviewed 125 homeless male street youths who indicated that, upon being victimized, they felt that physical aggression was required to protect themselves from further victimization and thus they would exert retaliatory violence to reduce the odds of repeated attacks. Similarly, Rich and Grey’s (2005) interviews with Black male victims of violence between the age of 18 and 30 revealed that 84% of the participants agreed that retaliation functioned to protect from repeat victimization. These victims believed that failure to respond aggressively to physical and psychological harms (i.e., failure to display RNs) created the impression that one was a “sucker” who just sits and takes an act of violence against them—thus sending a message to the previous offenders, or other potential assailants, that he was a suitable target for repeat victimization (see also Felson, 1982).
In contrast, most quantitative studies of the effect of street code on victimization to date provide little support for the idea that RNs protect against victimization while GTNs promote it. However, such differential effects would be difficult to discern, if they existed, because most previous quantitative studies have used measures of street codes that combine items tapping both RNs and GTNs (e.g., Melde et al., 2020; Schreck et al., 2012; Stewart et al., 2006; Zavala & Spohn, 2013). To be clear, the authors of these studies did not acknowledge a possible distinction between RNs and GTNs, and we see this as a potential limitation. Studies using single measures of street codes tend to find a positive relationship between street codes and victimization. For example, Stewart et al. (2006) examined the relationship between a unidimensional measure of street codes (one that combined RN and GTN items) and victimization among a sample of African American adolescents. They found a positive effect of street codes on victimization controlling for neighborhood context, demographic characteristics, and prior victimization experience. Their explanation for this finding was, essentially, that the display of street codes was antagonistic—that being belligerent toward others in order to gain respect might, in turn, provoke violence on the part of the disrespected party as a matter of saving face.
In other key research, Schreck et al. (2012) examined whether the correlates of violent school-based victimization were distinct from the correlates of non-violent school-based victimization. Using student data from the Rural Substance Abuse and Violence Project (RSVP) and a measure of street code orientation that combined RNs and GTNs, Schreck et al. (2012) found that those who adopted the street code were significantly more likely to experience violent victimization compared to nonviolent victimization, though this result was inconsistent across the four waves of data analyzed. In contrast, McNeeley and Wilcox (2015) analyzed data from the Seattle Neighborhoods and Crime Survey while using three separate single-item measures of street codes, noting that, if combined, the resulting index exhibited weak internal consistency. One of the street code items was similar to this study’s conceptualization of RNs, and the other two overlapped with the concept of GTNs. In opposition to the idea that RNs may represent invulnerability, they found that the item tapping retaliatory norms was associated with enhanced risk of victimization, as was an item tapping general toughness norms. The authors explained that adherents of the street code behave in ways that are considered disrespectful and belligerent, thus triggering other street code followers to respond violently—an explanation consistent with the idea that those displaying street codes are perceived as antagonistic and gratifiable targets.
The Present Study
Among the previous studies of the street codes-victimization link to date, only Schreck et al. (2012) have examined street values in relation to school-based adolescent violent victimization. We think continued work in this regard is warranted in order to determine the applicability of street values to understanding risk of violent victimization beyond disadvantaged inner-city communities, much like has already been demonstrated in the street values-violent offending literature (e.g., McGloin et al., 2011; Moule & Fox, 2020; Ousey & Wilcox, 2005; Swartz et al., 2017). While Schreck et al.’s (2012) study is pathbreaking in this regard, it is noteworthy that it did not focus on frequency of victimization, but on the likelihood of experiencing violent versus non-violent victimization. Also, much like the street code literature more generally, Schreck et al. (2012) used a unidimensional measure of street values, thus masking potentially distinct effects of RNs versus GTNs on violent victimization specialization (including victimization at school). Using the same RSVP data as Schreck et al. (2012), this study seeks to unpack the relationship between street codes and school-based serious violent victimization by examining the extent to which RNs and GTNs relate to the frequency of serious violent victimization at school. Specific hypotheses are as follows:
H1: Controlling for risky lifestyle, low self-control, social bonding, and sociodemographic factors, adherence to norms supporting retaliatory violence is negatively related to the frequency of serious violent victimization.
H2: Controlling for risky lifestyle, low self-control, social bonding, and sociodemographic factors, adherence to norms supporting general toughness is positively related to the frequency of serious violent victimization.
Data and Methods
Data
To test our hypotheses, we used the four total waves of student survey data from RSVP. Funded by the National Institute on Drug Abuse (DA-11317) and spanning the years 2001 to 2004, RSVP was designed by the original investigators to examine individual and contextual factors that affected substance use, victimization, and offending among middle and high school students in the state of Kentucky. The original study design for RSVP involved, first, a stratified random sampling of 30 of Kentucky’s 120 counties. Four strata were defined by county population, with stratum one representing the most populous counties in the state and stratum four representing the least populous. Largely relying upon probability-proportionate-to-size (stratum size) sampling techniques, two counties were selected from stratum one, four counties were sampled from stratum two, 11 counties were selected from stratum three, and 13 counties were sampled from stratum four. All public schools containing 7th graders within the sampled counties were asked to participate in the study, with 65 of the 74 eligible schools agreeing. All 9,488 seventh graders within the 65 participating schools in Wave 1 were targeted for the student sample. Active parental consent was obtained for 4,102 of these 7th graders. Completed surveys were obtained from 3,692 students in wave 1; 3,638 students in wave 2; 3,050 students in wave 3; and 3,040 students in wave 4. After listwise deletion of cases with missing data, there were 3,840 students who provided data in at least one wave. A total of 11,559 observations were provided by these students over the course of the four waves and within a total of 111 unique school contexts, as most students crossed from an elementary or middle school to a high school at some point during the study.
Measures: Dependent Variable
The dependent variable of interest in this analysis is a measure of in-school serious violent victimization. For serious violent victimization, students indicated how many times (0 = 0, 1 = 1. . …10 = 10+) they had experienced the following acts during the present school year, on the school grounds, or during school-related activities: (1) things taken by force (i.e., robbery), (2) having a weapon pulled on them (i.e., knife, brass knuckles, and so on, other than gun), and (3) having a gun pulled on them. Responses to the serious violent victimization items were summed for each respondent per wave, resulting in a count outcome ranging from 0 to 30 for serious violent victimization. Descriptive statistics in Table 1 indicate that, the average number of student victimizations per year was 0.73.
Descriptive Statistics for Study Variables.
Note. SD = standard deviation; Min = minimum; Max = maximum. (N = 11,559).
Measures: Independent Variables
Key independent variables included two measures of street codes: retaliatory norms and general toughness norms. For measurement of retaliatory norms, students were asked, in each wave, to indicate the extent to which they agreed (1 = strongly disagree. . . 4 = strongly agree) with the two following statements (Cronbach’s alpha = .73): (1) “it is alright to beat up another person if he/she started the fight,” and (2) “it is alright to beat up another person if he/she called you a dirty name.” For the measurement of general toughness norms, students were asked to indicate level of agreement (1 = strongly disagree. . . 4 = strongly agree) with the following six statements (Cronbach’s alpha = .87): (1) “in order to gain respect from your friends, it is sometimes necessary to beat up on other kids,” (2) “hitting another person is an acceptable way to get him/her to do what you want,” (3) “it is okay to break the law if you can get away with it,” (4) “to get ahead, sometimes you have to do things that seem wrong,” (5) “most things that adults call ‘crime’ don’t really hurt anyone,” and (6) “it is okay to break the law if nobody is hurt by it.” In short, our measure of RNs taps acceptance of situation-specific violence (when provoked), whereas our measure of GTNs measures more general skepticism of law and, coinciding with that, a more indiscriminant use of crime and violence. The scores across the two and six items, respectively, were averaged for each student-wave case.
We acknowledge that previous work often measures “street codes” as a single, unidimensional measure (Stewart & Simons, 2006, 2010; Taylor et al., 2010). 1 However, our theoretical framework and hypotheses (supported by some empirical literature, as discussed above) suggest that RNs and GTNs may be unique, with distinct implications for victimization. Thus, the present study separates street codes into two distinct street values—RNs and GTNs—and examines whether previous work might have masked important distinctions between types of street codes. Given this approach, we conducted Confirmatory Factor Analysis (CFA), with results presented in Figure 1. The CFA supports a two-latent-factor model (RMSEA = 0.048; CFI = 0.993; TLI = 0.983) (Hu & Bentler, 1999) and provides acceptable standardized regression weights for each of the indicators (i.e., factor loadings all >0.50). Given the moderate correlation between RNs and GTNs in the CFA (0.68), collinearity diagnostics (tolerances and VIFs) were assessed, and these values suggest that multicollinearity is unlikely to distort the findings presented herein. 2
Confirmatory factor analysis for street values.
Beyond our key measures of RNs and GTNs, this study draws upon the school victimization literature and includes as correlates of school-based serious violent victimization four variables for measuring risky lifestyle at each wave: delinquency, delinquent peers, school sports, and other school activities (e.g., Augustine et al., 2002; Burrow & Apel, 2008; Deryol et al., 2017; Peguero & Popp, 2012; Schreck & Fisher, 2004; Schreck et al., 2003, 2004, 2006; Tillyer et al., 2011, 2018; Wilcox et al., 2009; Welsh, 2001; Zaykowski & Gunter, 2012). Delinquency was measured as the average of 13 survey items asking students to indicate how often (1 = never. . .5 = daily or almost daily) they engaged in various delinquent activities in the current school year (Cronbach’s alpha = .89). Delinquent peers was calculated by taking the sum of 16 dichotomous items that tapped whether the student’s closest friends (0 = no, 1 = yes) were involved in various delinquent behaviors during the current school year (Cronbach’s alpha = .90). School sports was measured by a single item asking respondents about how often they take part in school sports during the school year (1 = never. . . 5 = everyday). Other school activities was measured by a single item asking respondents about how often they take part in other school activities (e.g., band, student government, yearbook) during the school year (1 = never . . .5 = everyday).
The model presented below also controls for other key correlates of victimization, including low self-control, social bonds, gender, race, and wave. The control variables were measured at each wave in a manner similar to other studies using RSVP (e.g., Ousey et al., 2008; Tillyer et al., 2016, 2018). Low self-control was measured by taking the mean of 11 items (1 = never true. . .4 = always true) that measure the student’s frustration, restlessness, temper, and attention span (Cronbach’s alpha = .91), with higher values reflecting lower levels of self-control. Social bonding in the form of attachment to school was measured by calculating the mean of 6 items that asked students to indicate their agreement (1 = strongly disagree. . . 4 = strongly agree) with statements that tapped positive feelings about school, teachers, and education (Cronbach’s alpha = .70). Student gender (0 = male, 1 = female) and student race (0 = non-white, 1 = white) were measured dichotomously. 3 Further, we also included wave (1 = wave1. . .4 = wave4) as a control variable.
Analytic Strategy
We view student serious violent victimization as logically driven by situational and interactional factors—including the immediate (i.e, contemporaneous) expression of street codes—as opposed to characteristics of individuals 1 year prior to the reported victimization (see also Schreck et al., 2012; Tillyer et al., 2018; Wilcox et al., 2009). Thus, we examined the contemporaneous relationships among RNs, GTNs and serious violent victimization across the 11,559 observations. However, in doing so, we take into account the potential endogeneity of each measurement per individual through estimation of a two-level, random-intercept mixed effect Poisson regression model using the “mepoisson” command in STATA. 4 Since over-time trends in victimization and the correlates thereof is not the focus, all slopes are treated as fixed. However, the random-intercept Poisson model accounts for clustering of observations within individuals and provides more conservative standard errors than using a similar negative binomial model (Wooldridge, 1999). Further, as each individual respondent is nested within a school at each wave, we include school fixed effects—essentially estimating the effects of school code dummy variables that account for school-level differences (using i.schoolcode subcommand). This approach enables the effects for our key independent variables to be interpreted as individual effects net of school-level variation. We also specified robust standard errors by using the “vce (robust)” subcommand.
Results
Table 2 presents the results of our estimation of the count of serious violent victimization experiences. Incident rate ratios (RRs) in Table 2 represent the exponentiated regression coefficients (
Mixed-effects Poisson Regression Model for Student Serious Violent Victimization.
p ≤ .05. **p ≤ .01. ***p ≤ .001.
Table 2 examines the effects of RNs and GTNs on the frequency of serious violent victimization. As hypothesized, Table 2 indicates that adherence to RNs is negatively and significantly associated with the frequency of serious violent victimization. More specifically, the finding indicates that the count of serious violent victimization decreases by approximately 12.7% for every unit increase in adherence to RNs. Also consistent with hypotheses, Table 2 indicates that adherence to GTNs is positively related to the frequency of serious violent victimization—increasing the expected frequency of victimization by 22.9% for each increment increase in adherence to GTNs, controlling for measures of risky lifestyles, low self-control, school attachment, and sociodemographic factors. Overall, RNs and GTNs are related to the frequency of serious violent victimization in opposite directions, as hypothesized.
Regarding the effects of risky lifestyle measures on serious violent victimization, results from Table 2 indicate that those who are more involved in delinquency report more frequent serious violent victimization (RR = 1.520, a 52.0% increase in frequency per unit increase in delinquent behavior). Further, the frequency of serious violent victimization increases by 3.9% for every unit increase in delinquent peer associations. Table 2 also indicates that those who are more frequently involved in school sports report more frequent serious violent victimization (RR = 1.107, a 10.7% increase in frequency per unit increase in involvement in school sports). However, other school activities are not significantly related to the frequency of serious violent victimization. Besides the effects of risky lifestyle measures on serious violent victimization, Table 2 indicate that for every increment in low self-control, the frequency of serious violent victimization increases by 21.2%. In contrast, with each unit increase in school attachment, the frequency of serious violent victimization decreases by 43.7%. Further, females report significantly less frequent victimization relative to males in Table 2. However, race and wave are not significantly related to the serious violent victimization in Table 2.
Discussion and Conclusion
We examined two hypotheses (H1 and H2) suggesting that the two distinct types of street values are related to students’ serious violent victimization in opposite directions. The findings were in support of H1 and H2: adherence to RNs was significantly and negatively related to the frequency of serious violent victimization and adherence to GTNs was positively related to the frequency of serious violent victimization. Overall, our findings are consistent with Anderson’s code of the street theory as well as target congruence theory (Finkelhor & Asdigian, 1996). More specifically, the findings reported here support a broader view of the applicability of Anderson’s Code of the Street thesis—a view that extends beyond the inner-city and into the hallways of schools in communities that span the rural-urban continuum (see also Moule & Fox, 2020; Schreck et al., 2012; Swartz et al., 2017). Moreover, our findings largely support the idea street codes are multifaceted, with different dimensions aligning with different aspects of target congruence. In particular, our results are consistent with the idea that RNs overlap with target vulnerability—that such norms might send a message of target invulnerability. Additionally, our findings support the idea that GTNs overlap with target antagonism and target gratifiability. Thus, displays of GTNs are potentially provocative and make one likely to be perceived as a valuable target in the game of give-and-take that is often behind the campaign for respect. That said, we offer these interpretations with caution and note that additional research (from offenders) is needed to more clearly understand whether displays of RNs and GTNs do actually align with perceived invulnerability, gratifiability, and antagonism in the manner suggested here.
Our findings regarding the different (opposite) effects of RNs and GTNs on victimization offer one explanation for conflicting results across previous studies that examine the relationship between the code and victimization using single indexes that combined these two types of norms. However, we recognize that our very examination of potential multidimensionality with respect to street codes contradicts some previous work. For example, Taylor et al. (2010) concluded that at least one street code scale—that developed by Stewart et al. (2006)—exhibited unidimensionality, with strong internal consistency among street code items across different groups and contexts (though, in an exception, they found that items regarding street codes loaded on two distinct factors for American Indian youths). Given such findings, we exercise caution in suggesting that there may be two distinct dimensions to street codes, each with different implications for serious violent victimization. We encourage further exploration and debate regarding this issue.
In terms of practical implications, the findings reported here suggest that tackling GTNs, and the behavioral implications thereof (e.g., belligerence, antagonism), are particularly promising avenues for reducing school-based victimization. In fact, numerous existing school-based prevention programs do tackle GTNs in a variety of ways. For instance, some of the more successful school-based violence prevention programs focus on student-level behavioral modification, social competency skill-building, and peer sociability—all of which would address the negative behaviors associated with adherence to GTNs (e.g., Gottfredson, 2001; Gottfredson et al., 2012). Other successful programs focus on altering the school environment so that it is a context that more effectively conveys positive norms, manages behavior, and emphasizes academic achievement—thus providing students with positive avenues for achieving respect so that adherence to GTNs are less needed, useful, and/or tolerated (Gottfredson, 2001; Gottfredson et al., 2012). The findings reported here reinforce the value of such programming.
Beyond programming addressing the harmful effects of GTNs, the results reported here regarding RNs imply that in certain instances (i.e., specifically when confronted with an attack), sending a strong message that one is willing to defend oneself can be beneficial. Obviously, care is needed in discussing this implication, as we do not intend to suggest that our findings support unfettered violence in the name of self-protection. Rather, we view the findings reported here as consistent with many self-defense protocols which support defensive action in particularly dangerous situations (i.e., telling potential victims to walk assertively, shout down would-be attackers, fight back, etc.). That said, we recognize that the empirical evidence in support of the value of victim resistance (not directly measured here) is quite mixed (c.f., Lindegaard et al., 2015; Tark & Kleck, 2004).
It is important to qualify our study’s implications and contributions by noting study limitations and the need for additional research. First, the data we used come from students in schools across the rural-urban continuum in one state—Kentucky. As discussed above, we follow upon previous work that has extended the street code framework to samples beyond the inner-city, including to this Kentucky-based sample. We nonetheless recognize that our findings are not necessarily generalizable to students in other geograpic, socioeconoimic, and cultural contexts. For example, in stark contrast to the negative direct effects of RNs on victimization that is observed in the model reported here, Stewart et al. (2006) argued that decent youths’ temporary adoption of street values (i.e., “code switching”) is likely not interpreted (on the part of offenders) as an ability to take care of oneself in neighborhoods where the street culture is particularly pervasive, and thus such youths remain vulnerable. While our analysis controlled for school differences, future work might more explicitly consider whether certain aspects of school or neighborhood climate moderate the effects of individual adherence to RNs and GTNs on vicitmization. 5
Other limitations of the present study involve validity concerns, both in terms of measurement and causality. The RSVP study was originally designed with the intention of measuring norms among the students surveyed, but it was not explicitly directed at distinguishing RNs and GTNs. Thus, the measures available for our secondary analysis aimed at discerning the effects of RNs versus GTNs were limited. Somewhat related, an unintended consequence of delineating a distinction between these orientations, as we have done here, is the implication that adolescents have either a RN orientation or a GTN orientation. In fact, while this study distinguishes RNs and GTNs, it does not preclude the possibility of students adhering strongly (or weakly) to both orientations. Here, we examined the effects of RNs and GTNs, while holding the other constant, but future work may wish to more fully consider the ways these norms work in tandem to influence victimization (i.e., through interaction effects). Additionally, our study’s emphasis on effects of RNs and GTNs on victimization downplays the possible reciprocity between normative orientations and victimization. Given that experiences with victimization might affects one’s adherence to particular norms, future studies might look to examine the norms-victimization relationship in a bi-directional manner.
Lastly, though student data was collected across four waves of the RSVP study, our interest for the purposes of the present study was not on intra-individual change in RNs, GTNs, and violent victimization across waves. Rather, our initial interest was in examining the potentially distinct effects of RNs and GTNs on victimization within any one wave. While we see this as a valuable first-step in examining the links to vicitmization of RNs versus GTNs, we do realize that potentially important developmental, dynamic processes were not addressed through our approach. Thus, explicit theory (and more detailed testing thereof) of time-varying effects of RNs and GTNs on intra-individual trajectories of victimization is still a potentially worthy direction for future research.
Despite limitations, we believe this study offers an important step in explaning school-based violent victimization in light of Elijah Anderson’s Code of the Street thesis. In the process, it highlights the potential utility in recognizing two distinct dimensions of street codes while also emphasizing the particularly harmful effects of embracing general toughness as a way of life.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship and/or publication of this article: Data analyzed for this manuscript were originally collected through funding from the National Institute on Drug Abuse (DA 11317, Richard R. Clayton, PI). The conclusions and opinions of this study are the sole responsibility of the authors.
