Abstract
According to general strain theory’s (GST) vicarious strain hypothesis, individuals’ behavior is influenced by witnessing or having knowledge of others’ stressful experiences, especially those within their network. Drawing on Agnew’s GST, the current study examines the relationship between peer victimization and violent offending, as well as the potential intervening factors in this relationship. The primary research questions are: (a) does vicarious strain occur through having friends who are victimized? And (b) what role does violent peers play in the coping process? Data for the analysis were obtained from the National Longitudinal Study of Adolescents to Adult Health’s friendship network data. Results from multivariate analysis show that victimization of ones’ friend (or peer victimization) is associated with depression as well as personal involvement in violent offending. This relationship, however, is fully mediated by depression. Regarding deviant peers, a three-way interaction effect was observed, specifically peer victimization, involvement with peers, and associating violent peers was a significant predictor of violent offending. Findings from the current study highlight the importance of considering the role of deviant peers in the coping process. Results lend support to Agnew’s GST. Implications for theory and research are discussed.
Keywords: peer victimization, depression, violence, deviant peers
In the United States, criminal victimization remains a social problem that severely impacts the lives of individuals. As a result, victimization continues to be a concern for researchers, educators, and clinicians. According to recent data from the National Crime Victimization Survey (NCVS), in 2016 there were approximately 5.7 million violent victimizations (e.g., rape, sexual assault, robbery, aggravated and simple assault) experienced by individuals aged 12 and older (Morgan & Kena, 2017). The overall violent victimization rate that year was 21.1 victimizations per 1,000 persons age 12 and older. While victimization impacts society as a whole and is not a problem concerning only the United States (see Enzmann et al., 2010; Tseloni et al., 2010), young people are more likely to be victimized compared to adults. Data from the Office of Juvenile Justice and Delinquency Prevention (OJJDP) shows that most acts of violent victimization were experienced by young people (between the ages of 12 and 24 years old) when compared to their adult counterparts (age 25 and over).
Victimization has detrimental effects on the lives of those who experience it, potentially increasing the risk of maladaptive coping. Studies have found that violent victimization is positively associated with depression (Hawker & Boulton, 2000; Manasse & Ganem, 2009; Mitchell et al., 2007), substance use behaviors (Luk et al., 2010; Mitchell et al., 2007; Tharp-Taylor et al., 2009), and delinquency (Hay & Evans, 2006; Manasse & Ganem, 2009). Criminologists have attempted to provide an explanation for how victimization may increase individuals’ involvement in criminal and deviant behavior. Most research examining the relationship between victimization and crime has focused on the direct effect of experienced victimization and less on its vicarious effect. This study contributes to general strain theory (GST) literature by specifically investigating the relationship between vicarious victimization and violent offending.
Agnew’s (1992) GST provides insight into the relationship between victimization and crime. The main argument of GST is that strain and stressors lead to negative emotions (e.g., anger, frustration, and depression), ultimately fostering criminal and deviant coping. Agnew (2001) argues that victimization is a key strain that is likely to lead to crime as it is viewed as unjust, high in magnitude, often occurs in areas where social control is low, and creates pressure or incentive to engage in crime. According to GST, there are three ways that strains can impact someone’s life. Strain can be experienced (e.g., being victimized), anticipated (e.g., expectation of strain to continue or occur in the future), or vicariously experienced (e.g., witnessing strain experienced by others) (see Agnew, 2002). Most GST research investigating the strain–crime link has focused on the experienced variety, providing little scrutiny to the anticipated and vicarious strain hypotheses (see Agnew, 2002, 2006; Baron, 2009; Lin et al., 2011; McGrath et al., 2012; Vogel & Keith, 2015). This study is primarily concerned with the vicarious strain hypothesis, specifically arguing that the effect of peer victimization negatively impacts the behavior of those within the same network.
GST’s vicarious strain hypothesis have received empirical support. Findings suggests that the victimization experiences of others, especially family and friends, significantly impact the behavior of those who witness it (see Agnew, 2002; Baron, 2009; Lin et al., 2011; McGrath et al., 2012; Vogel & Keith, 2015). The current study seeks to further investigate this hypothesis by exploring the role of deviant peers in the coping process.
There are several mechanisms that contribute to the strain–crime relationship, one of which is deviant peers. There is a strong relationship between deviant peer affiliation and criminal and deviant behavior. Associating with deviant peers has been shown to increase personal involvement in negative conduct, such as crime and delinquency (Fergusson et al., 2002; Haynie, 2002; Haynie & Osgood, 2005), and substance use (Fergusson et al., 2002; Van Ryzin et al., 2012; Walker & Holtfreter, 2019). These findings remain consistent across perceptual (i.e., individuals’ perceptions of peer delinquency) and actual measures (i.e., friendship network analysis) of deviant peers (see Hoeben et al., 2016). Being that deviant peers promote criminal and deviant behavior, it is important investigate the role of peers in the coping process. In light of GST, associating with deviant peers has been argued to influence maladaptive coping (see Agnew & White, 1992; Mazerolle & Maahs, 2000; Walker & Holtfreter, 2019). Deviant peers may impact the relationship between vicarious victimization and crime.
To date, the vicarious strain hypothesis has received minimal empirical scrutiny. Research assessing GST’s vicarious strain hypothesis have typically relied on individuals’ perceptions of the victimization among those within their network (see Agnew, 2006; Baron, 2009; Lin et al., 2011). Guided by GST and using friendship network data (i.e., respondents’ closest male and female friends) from the National Longitudinal Survey of Adolescent to Adult Health (Add Health), the current study examines the relationship between vicarious strain (i.e., peer victimization) and violent offending.
LITERATURE REVIEW
General Strain Theory
GST is one of many theories that provide an explanation for criminal and deviant behavior, primarily focusing on negative relationships with others. Agnew (1992) defines negative relationships as “relationships in which others are not treating the individual as he or she would like to be treated” (p. 50). Although GST has focused on negative relationships as a primary stressor, there are other sources of strains that can also increase involvement in crime and delinquency. Agnew (1992) asserts that there are three sources of strain: (a) the failure to achieve positive valued goals (e.g., failing to graduate from college or obtaining employment); (b) the presence of negative stimuli (e.g., experiencing victimization); and (c) the removal/absence of positive stimuli (e.g., death of a family member or close friend, or divorce from a spouse).
Emotions play a critical role in Agnew’s (1992) GST. Agnew argues that individuals who experience strain are likely to experience negative emotions, such as anger, frustration, depression, and anxiety. Moreover, he asserts that anger is the most critical emotional reaction to strain and is most conducive to criminal and delinquent behavior. Agnew differentiates between self- or inner-directed (e.g., depression) and other- or outer-directed emotions (e.g., anger), hypothesizing that outer-directed emotions are likely to produce outer-directed deviance, such as interpersonal aggression and violence. To the contrary, self- or inner-directed emotions are likely to result in self-directed acts like substance use and eating disorders. Scholars typically focus on illegitimate coping (e.g., crime/delinquency and substance use) when testing GST (e.g., Agnew, 2006; Agnew et al., 2002; Agnew & White, 1992; Capowich et al., 2001; Mazerolle et al., 2003; Walker & Holtfreter, 2019). In later works, Agnew expanded on GST by further specifying that there are three types of strain.
Agnew (2002) posits that there are three types of strain, which are experienced, vicarious, and anticipated strain. Experienced strain refers to personal experiences, specifically with the presentation of negative stimuli (e.g., victimization). Agnew (2006) argues that the experience strain hypothesis is the most tested component of GST (see Agnew et al., 2002; Aseltine et al., 2000; Baron, 2004; Broidy, 2001; Hay & Meldrum, 2010; Moon & Morash, 2017). Vicarious strain refers to the witnessing or knowing of actual strains experienced by people either in close proximity or within one’s network, especially family, close friends, or those within the same community (Broidy & Agnew, 1997). Strain experienced by an individual within a particular group may vicariously affect others within that specific network. This is especially true for networks where there is a high concern for the welfare of others. Lastly, anticipated strain occurs when there is an expectation for strain to continue and/or occur in the future (Agnew, 2002). This particular type of strain is typically concerned with an individual’s negative outlook or expectations of the future.
A considerable amount of GST research has examined the experienced strain hypothesis while seldom testing the effect of vicarious strains on negative outcomes. This dearth in literature is problematic, especially when considering that the strain of criminal victimization does not only affect those who experience the trauma, but also the lives of others. In light of GST, the next section discusses research examining the vicarious strain hypothesis.
Vicarious Strain
The vicarious strain hypothesis suggests that strain or negative treatment experienced by others may affect the emotions and behavior of individuals who witness it, especially those within the same network (e.g., family members, friends, and community residents). Agnew (2002) asserts that there are multiple ways vicarious strain presents itself to an individual: (a) they may directly witness the strain endured by others, such as an assault, (b) they may hear others experience strain (e.g., gunshots, screams), or (c) they may learn about the strain endured by others, directly (e.g., from victims) and/or indirectly (e.g., media). There are certain characteristics of vicarious strain that are most likely to result in delinquency. More specifically, strains that occur to close others, to the members of groups to which individuals belong or identify with, are in close physical proximity, are not successfully resolved, or strains that are contagious (i.e., strains that have a high probability of affecting the individual) elevates the risk of criminal and deviant behavior (Agnew, 2002). In short, there are several ways vicarious strain promotes maladaptive coping.
There is variation in how vicarious strains have been operationalized. Vicarious strain is usually measured using perceptual measures asking respondents about their friends’ victimization (see Baron, 2009) or asking respondents if they have witnessed violence (e.g., seeing someone being victimized; see Lin et al., 2011). GST provides little guidance on a desirable approach for measuring vicarious strain. Consequently, most studies have relied on perceptual measures. This may be due to data limitations and lack of theoretical clarity. To date, there has been limited research examining vicarious strain hypothesis using friendship network data (see Vogel & Keith, 2015). Building on prior research, the current study deviates from the norm by also using friendship network data to assess the GST’s vicarious strain hypothesis. This source of data provides an actual measure of the victimization experienced by respondents’ friends. It also allows for testing whether having a more intense relationship with those victimized peers make the vicarious effect stronger.
Victimization and Vicarious Strain
Criminal victimization is associated with emotional, social, and behavioral problems, especially among the adolescent population. Victimization increases the risk of depression (Barchia & Bussey, 2010; Hawker & Boulton, 2000; Mitchell et al., 2007), substance use (Carson et al., 2008; Luk et al., 2010; Mitchell et al., 2007; Sullivan et al., 2006; Tharp-Taylor et al., 2009), and crime and delinquency (Baron, 2004; Manasse & Ganem, 2009; Mitchell et al., 2007; Sullivan et al., 2006). Moreover, there are long-term consequences of victimization. Widom et al. (2006) found that childhood victimization increases the risk of illicit drug use in middle adulthood. The effects of victimization are not limited to those who experience it, but also impact the lives of those who witness it.
Vicarious strains, like victimization, have been hypothesized to influence maladaptive coping. Agnew (2001) argues that vicarious victimization may upset individuals because people they care about have been harmed, ultimately pressuring them to cope in a deviant manner. He provides reasons why vicarious strain may promote delinquency. Specifically, individuals may engage in delinquency to (a) prevent additional harm to those they care about, (b) seek revenge against those they believe are responsible for the harm, and (c) to alleviate negative emotions. Studies testing GST’s vicarious strain hypothesis have found that witnessing victimization is associated with criminal and deviant behavior (e.g., Agnew, 2002; Baron, 2009; Lin et al., 2011; McGrath et al., 2012; Zavala & Spohn, 2013). Agnew’s (2002) study was one of the first to explore the vicarious effects of victimization. He found that vicarious victimization, specifically victimization of friends and family, was associated with an increase in delinquency.
Baron’s (2009) study examined the role of violent personal, vicarious, and anticipated victimization on youths’ violent offending. In this particular study, vicarious strain was measured by asking respondents how many of their friends were a victim of an assault, minor assault, or had threat or force used against them to get things. Baron found that all three types of strains were significantly associated with self-reported violent offending. In addition, the relationship between vicarious strain and violence was moderated by low self-control. Lin et al.’s (2011) study tested the vicarious strain hypothesis and its effect on violent/property crime and drug use. Here vicarious strain was operationalized by asking adolescents if they ever saw someone being stabbed with a knife, sexually assaulted/raped, robbed/mugged, threatened with a weapon, being hit or kicked/beaten, or getting shot. They found that violent vicarious strain was significantly related to self-reported delinquency. Similarly, Eitle and Turner (2002) examined the effect of witnessing community violence, learning about the violent victimization of a significant other (e.g., domestic violence), and being exposed to traumatic news (e.g., hearing about a friend being jumped) on criminal behavior. They found that all sources of vicarious strain increased the risk for criminal behavior, thus lending support to Agnew’s vicarious strain hypothesis. In sum, there are several studies showing that individuals who have witnessed others experience victimization, either indirectly or directly, are more likely to engage in maladaptive coping behavior. However, this is fairly uncommon as most adolescents’ exposure to victimization may be through hearing about things that occurred to their friends especially within networks where the level of involvement is high and there is a high concern for the welfare of others.
Majority of studies so far examining the vicarious strain hypothesis have either relied on perceptual measures or do not investigate the nature of the relationship that individuals have with those who are victimized. For example, respondents are typically asked to report on their friends level of victimization (e.g., “how many of their friends had been the victim of a serious assault; a minor assault; and having someone use threats or force against them to get things”) (see Baron, 2009; Eitle & Turner, 2002), or to report on how often they witnessed victimization, such as seeing somebody get shot or stabbed, beat up, or have their property stolen (e.g., Agnew, 2002; Kort-Butler, 2010; McGrath et al., 2012; Zavala & Spohn, 2013). Agnew argues that strain will have a greater effect on an individual’s behavior when people he or she cares about are harmed. Since vicarious victimization occurs by having friends who are victimized, the use of friendship network data is beneficial as it allows researchers to capture the actual victimization taking place within a network and the level of involvement that respondents have with their peers. Similarly, it is also possible to assess respondents friends’ levels of criminal and deviant behavior. Vogel and Keith’s (2015) study found support for the vicarious strain hypothesis using friendship network data, specifically vicarious peer victimization was positively associated with violent offending. While that is the case, they only consider the victimization experience of one friend within respondents’ network and do not examine whether the level of involvement with their peer conditions the relationship between vicarious victimization and violent offending. Building on this line of research assessing the relationship between peer victimization and violent offending, the current study considers the victimization experiences of all individuals within respondents’ direct network as well as the potential intervening mechanisms in this relationship.
Vicarious Strain and Deviant Peers
Deviant peers are associated with a host of negative outcomes. Research shows that associating with deviant peers increases involvement in substance use, as well as criminal and delinquent behaviors (see Hoeben et al., 2016). GST suggests that deviant peers will promote criminal and deviant behavior among those who experience strain by (a) supplying a form of support for criminal behavior, (b) considering certain acts as an appropriate response to strain, and (c) serving as instigators. Agnew (1992, 2006) argues that those with deviant peers are more likely to have access to deviant coping strategies and view deviance as an attractive or appropriate response to stressful situations. Agnew and White (1992) found that deviant peers moderated the relationship between strain and drug use. Similarly, Mazerolle and Maahs’s (2000) findings suggests that the relationship between strain and delinquency is conditioned by delinquent peers. Walker and Holtfreter (2019) study found that deviant peers promoted substance use involvement among females who experienced teenage pregnancy and depression. While prior work has examined whether the effect of experienced strain (e.g., personal victimization) and associating with deviant peers impacts maladaptive coping, it remains unknown whether the effect of vicarious strain (e.g., peer victimization) on violent offending is influenced by deviant peers.
Vicarious Strain and Network Characteristics
When assessing the effect of vicarious strain on maladaptive coping, friendship network data can provide a better understanding of the peer victimization and crime relationship. According to McGloin and Shermer (2009), network data provides information on a social group’s cohesion, a person’s position within their group, and the level of social interaction among those within a network. Arguably, peer victimization would have a greater effect on individuals who are well connected with their network, specifically those who are highly involved with their peers.
CURRENT FOCUS
The vicarious strain hypothesis is one of the most under studied components of GST. With previous research on vicarious strain in mind and using friendship network data from the National Longitudinal Survey of Adolescent to Adult Health (Add Health), the present study examines the relationship between vicarious strain and maladaptive coping. The current study explored the following hypotheses:
H1: Consistent with GST, it is expected that peer victimization will be a significant predictor of depression. Agnew (1992) argues that strain is associated with negative emotions, such as anger, frustration, anxiety, and depression.
H2: It is hypothesized that peer victimization is positively associated with personal involvement in violent offending. Prior research assessing GST’s vicarious strain hypothesis found that vicarious victimization is associated with personal involvement in criminal and delinquent behavior (e.g., Agnew, 2002; Baron, 2009; Lin et al., 2011; Zavala & Spohn, 2013).
H3: Regarding GST’s assumptions, depression is hypothesized to mediate the relationship between peer victimization and violent offending (Agnew, 1992).
H4: It is expected that the effect of peer victimization on violent offending is conditioned by violent peers. Arguably, deviant peers will elevate the effect of peer victimization on violent behavior by either supplying a form of support for criminal behavior, considering certain acts as an appropriate response to strain, or serving as instigators (see Agnew, 2006).
H5: Having a strong relationship with peers (i.e., spending time with peers) may amplify the vicarious strain effect on violent offending, especially among those who associate with violent peers. Therefore, it hypothesized that individuals who are highly involved with their peers who have experienced victimization and also associate with violent peers will be more likely to cope in a criminal manner.
METHODS
This study uses data from waves I and II of the National Longitudinal Survey of Adolescent to Adult Health (Add Health). Add Health is a nationally representative, school-based sample of nearly 20,000 adolescents in the United States from grades 7 through 12 between 1994 and 1995. The survey was administered using a stratified sample design, stratifying schools by region, school size and type, urbanicity, race and ethnicity, and grade level. A total of 132 schools participated in data collection, of which 80 were high schools and the remaining 52 were middle schools. The in-school data were collected from students who were present on the day surveys were administered. During this time, respondents had the opportunity to elicit friendship nominations from school rosters. Individuals who participated in the in-school survey were selected from the school roster to participate in the in-home data collection, specifically an in-home survey and interview. Participants were surveyed using computer-assisted personal interviews and audio computer-assisted self-interviews for sensitive questions. Information collected includes, but is not limited to, the respondent’s social and demographic characteristics, risky behaviors, criminal activities, victimization experiences, self-efficacy, education and occupation of parents, and health status.
In addition to baseline behavioral measures at wave I, measures were used from wave II to investigate the relationship between peer victimization and violent offending. Wave II data collection was conducted between April and August of 1996. All adolescents in grades 7 through 11 at wave I were included in wave II interviews, ultimately yielding a sample of 14,738 respondents. Measures were consistent across waves I and II.
The Add Health survey design is unique, allowing researchers to collect data on social networks. This study is limited to a portion of the Add Health data collected from particular schools; these data are referred to as the “saturation sample” (n = 2,728). A total of 16 schools (i.e., 2 large and 14 small schools) were selected from rural and urban areas to take part in the in-home interviews. To enable analysis of social networks, the saturation sample only include schools where all students were surveyed. Individuals within the selected schools were allowed to nominate up to 10 friends, specifically five of their closest male friends and five of their closest female friends. In some cases, respondents did not nominate any friends, ultimately resulting in missing network data. For the purpose of the current study, only respondents with complete friendship network data were analyzed. After excluding cases with missing data using listwise deletion, the current analysis is based on interviews with 1,848 respondents.
Dependent Variable
Violent Offending.
While the Add Health study was primarily concerned with assessing the health behaviors, risks, and status of adolescents, four items assessed adolescents’ involvement in violent offending. Items were measured on a three-point scale (0 = never, 1= once, 2 = more than once), respondents were asked during the past 12 months, how often did they get into a serious physical fight, hurt someone badly enough to need bandages or care from a doctor or nurse, pulled a knife or gun on someone, or shot or stabbed someone. Each item was recoded dichotomously, indicating whether the respondent had engaged in that activity. Similar to Haynie and Payne (2004), the final violent offending scale is a binary response (0 = no violence and 1 = engaged in violence) indicating whether the respondent participated in any of the four acts. Approximately 19% of the sample reported engaging in violence at wave II.
Independent Variables
Using in-school friendship network data that allows respondents to nominate up to 10 friends, adolescents’ friendship networks are determined by all individuals the respondent identifies and nominates as a close friend. Defining each participant’s network enables researchers to identify the portion of friends in the network who have been violently victimized. To assess peer victimization, the current study uses a common four-item violent victimization scale drawn from wave I (see Guterman et al., 2002; Schreck & Fisher, 2004). Respondents were asked during the past 12 months, how often have “someone pulled a knife or gun on you,” “someone shot you,” “someone cut or stabbed you,” or “you were jumped.” Responses ranged from 0 (never) to 2 (more than once). Each item was recoded dichotomously where 0 = not at all and 1 = at least once. The items are then summed with higher values representing higher levels of victimization (mean = .29, SD = .55). To capture the absolute level of victimization occurring in the network, the victimization index is calculated for each network member and summed across all members. For each respondent, peer victimization was measured averaging the amount of victimization experienced by those within respondents’ network.
Intervening Variables
One of the most commonly used scales by social scientists to measure depressive symptoms in the general population is the Center for Epidemiological Studies Depression Scale (CES-D). This scale has been deemed reliable and valid (Radloff, 1977). The current study uses a 16-item scale consisting of measures adapted from CES-D to assess respondent’s levels of depression. During waves I and II data collection, respondents were asked within the past 7 days how often: (a) “you were bothered by things that usually don’t bother you”; (b) “you didn’t feel like eating, or your appetite was poor”; (c) “you felt that you could not shake off the blues, even with help from your family and your friends”; (d) “you felt that you were just as good as other people”; (e) “you had trouble keeping your mind on what you were doing”; (f) “you felt depressed”; (g) “you felt hopeful about the future”; (h) “you thought your life had been a failure”; (i) “you felt fearful”; (j) “you were happy”; (k) “you talked less than usual”; (l) “you felt lonely”; (m) “people were unfriendly to you”; (n) “you enjoyed life”; (o) “you felt sad”; and (p) “you felt that people disliked you.” Responses to each question ranged from 0 (never or rarely) to 3 (most of time or all of the time). Items d, g, j, and n were recoded so that higher values indicated greater levels of depression. A summated scale using the 16-items was constructed where higher values reflect greater levels of depression (wave I: mean = 13.84, SD = 6.38; wave II: mean = 13.76, SD = 6.39). The scale has high internal consistency in the study sample (α = .84; .85, respectively).
Violent peers measure was operationalized using the same four items used to measure violent offending. Items were measured on a three-point scale (0 = never, 1 = once, 2 = more than once), respondents were asked during the past 12 months, how often did they “get into a serious physical fight,” “hurt someone badly enough to need bandages or care from a doctor or nurse,” “pulled a knife or gun on someone,” or “shot or stabbed someone.” Each item was recoded dichotomously, indicating whether the respondent had engaged in that activity. The final violent offending scale is a binary response of whether the respondent participated in any of the four acts (0 = no violence, 1 = engaged in violence). Similar to previous research (see McGloin, 2009; McGloin & Shermer, 2009), the violent peers measure was constructed by taking the mean value of offending items for the respondent’s friendship send network (mean = .15, SD = .26).
To assess respondent involvement with peers, for each friend nominated at wave I the respondent was asked two questions: (a) whether he or she had met the friend after school to hang out during the past week, and (b) whether the respondent spent time with the friend during the past weekend. Each item was coded dichotomously (0 = no, 1 = yes). Similar to previous research (e.g., Haynie & Osgood, 2005; McGloin & Shermer, 2009), involvement was measured by summing the two responses across friends and dividing this sum by the square root of the number of friends (mean = 3.39, SD = 2.51). Higher values reflect greater levels of involvement with peers.
Control Variables
Consistent with prior research (see McGloin & Shermer, 2009; Walker & Holtfreter, 2019), low self-control was operationalized using six items. Respondents were asked how much did they agree with the following statements, during the past week: (a) “when you have a problem to solve, one of the first things you do is get as many facts about the problem as possible,” (b) “when you are attempting to find a solution to a problem, you usually try to think of as many different ways to approach the problem as possible,” (c) “when making decisions, you generally use a systematic method for judging comparing alternatives,” and (d) “after carrying out a solution to a problem, you usually try to analyze what went right and what went wrong” (1 = strongly agree, 2 = agree, 3 = neither agree nor disagree, 4 = disagree, 5 = strongly disagree). The last two items asked respondents during the past school year, how often have you had trouble “getting your homework done” (0 = never to 4 = everyday), and “paying attention in school” (0 = never to 4 = everyday). Being that the response sets varied among the questions, the items were standardized prior to creating a summated scale, with higher values reflecting lower levels of self-control (α = 0.67).
Similar to Wight et al. (2006), social support was operationalized using seven items that asked respondents how much they felt “that adults care about you,” “your teachers care about you,” “your parents care about you,” “your friends care about you,” “people in your family understand you,” “your family pays attention to you,” and “you and your family have fun together” (0 = not at all, 1 = very little, 2 = somewhat, 3 = quite a bit, and 4 = very much). The items were summed so that higher values indicate greater levels of social support (α = .78).
In an attempt to guard against spuriousness, age, sex, and race/ethnicity are statistically controlled for in the multivariate analyses. Age was measured in years (mean = 16.57, SD = 1.51). Regarding sex, male was a binary measure where 0 = female and 1 = male (50% male). Race/ethnicity was controlled for using a single dichotomous variable, specifically White (0 = no, 1= yes), with other races/ethnicities as the reference category. Lastly, a baseline measure violent offending was controlled for. Approximately 34% of the sample reported engaging in violence at wave 1.
ANALYTIC STRATEGY
To empirically assess the relationship between peer victimization and individuals’ self-reported violent offending, the analysis proceeds in several stages. First, a test of multicollinearity was conducted, which revealed that there were no problems (VIF range = 1.01–1.66, mean VIF = 1.22). Second, descriptive statistics and bivariate correlations for study variables are calculated. Bivariate correlations are presented to detect significant relationships between the dependent and key independent variables. Third, to assess the relationship between peer victimization and depression, a negative binomial regression model was estimated due to the overdispersion in the depression measure (mean = 13.84, variance = 40.67). Fourth, the Bruesch–Pagan test was used to check for heteroscedasticity in the negative binomial regression models. The results indicated the presence of heteroscedasticity; therefore, robust standard errors (SEs) were estimated. Due the dichotomous nature of the violent offending variable, logistic regressions models were estimated.
The analyses to be presented required the creation of multiple interaction terms to investigate the moderation effects of deviant peers. To examine whether violent peers conditions the relationship between peer victimization and violent offending (hypothesis 4), a two-way interaction term was created between peer victimization and violent peers. A three-way interaction term (peer victimization × involvement with peers × violent peers) was also created to test the hypothesis that respondents who are highly involved with their peers who have experienced violent victimization and also associate with violent peers will be more likely to cope in a criminal manner (hypothesis 5). To accurately analyze a three-way interaction, it requires the inclusion of various two-way interactions between the variables that make up the three-way interaction term (Baron, 2011). All variables were mean centered prior to creating interaction terms to alleviate problems associated with multicollinearity (Aiken & West, 1991). To account for the survey design, clustering (i.e., by school) effects were estimated for all models. All analyses were estimated using STATA 15 (StataCorp, College Station, TX).
RESULTS
Table 1 summarizes descriptive statistics and presents Pearson’s correlation coefficients for each of the variables included in the study. Based on the descriptive statistics, the average age is 16.53 years and approximately 52% of the sample is White. Regarding the dependent variable, approximately 34% of the sample reported involvement in violent offending at wave I and 19% of respondents at wave II.
TABLE 1.
Descriptive Statistics and Bivariate Correlations for Study Variables (N = 1,848)
| M | SD | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Violent offending | 0.19 | 1.00 | ||||||||||||
| 2 | Peer victimization | 0.29 | 0.55 | 0.12* | 1.00 | ||||||||||
| 3 | Depression | 13.84 | 6.38 | 0.10* | 0.08* | 1.00 | |||||||||
| 4 | Violent peers | 0.15 | 0.26 | 0.05* | −0.04 | 0.00 | 1.00 | ||||||||
| 5 | Involvement | 3.39 | 2.51 | 0.10* | 0.06* | 0.10* | 0.01 | 1.00 | |||||||
| 6 | Low self-control | −0.02 | 3.67 | 0.10* | 0.03 | 0.20* | 0.03 | 0.02 | 1.00 | ||||||
| 7 | Social support | 21.17 | 4.13 | −0.12* | −0.09* | −0.28* | 0.00 | −0.07* | −0.25* | 1.00 | |||||
| 8 | Age | 16.53 | 1.49 | 0.00 | 0.13* | 0.20* | 0.01 | 0.29* | −0.01 | −0.08* | 1.00 | ||||
| 9 | Male | 0.50 | 0.20* | 0.12* | −0.13* | 0.03 | 0.00 | 0.07* | −0.06* | 0.09* | 1.00 | ||||
| 10 | White | 0.52 | −0.09* | −0.14* | −0.19* | −0.01 | −0.07* | 0.12* | 0.01 | −0.23* | 0.01 | 1.00 | |||
| 11 | Depression (WII) | 13.76 | 6.39 | 0.14* | 0.08* | 0.58* | −0.01 | 0.08* | 0.14* | −0.39* | 0.16* | −0.09* | −0.18* | 1.00 | |
| 12 | Violent offending (WI) | 0.34 | 0.41* | 0.15* | 0.11* | 0.03 | 0.07* | 0.13* | −0.12* | −0.00 | 0.25* | −0.11* | 0.10* | 1.00 |
Source. National Longitudinal Study of Adolescent to Adult Health.
Note. Entries are Pearson’s correlation coefficients.
p < .05 (two-tailed test).
There are several significant correlations between the variables of interest. All of the Pearson’s coefficients are less than an absolute value of 0.50 indicating that there are no problems of collinearity. The correlation between the peer victimization and violent offending was positive (r = 0.12, p < .05). Violent offending was also correlated with depression (r = 0.14, p < .05), violent peers (r = 0.05, p < .05), involvement with peers (r = 0.10, p < .05), low self-control (r = 0.10, p < .05), and social support (r = −0.12, p < .05), all in the expected direction. Compared to prior research, the correlation between deviant peers and violent offending in the current study is fairly weak (see Henry et al., 2001; Nofziger & Kurtz, 2005). One possible explanation for this could be the use of friendship network data, which arguably provides a more accurate amount of violence taking place within a network, rather than relying on respondents’ perceptions of peer behavior. Depression (r = 0.08, p < .05) was among the several variables that was positively correlated with peer victimization. The significant and positive correlations between peer victimization (or friends’ victimization) and deviant peers is consistent with research on the victim–offender overlap, which suggest that individuals who are victimized are likely to engage in criminal behavior, and vice versa (see Lauritsen et al., 1991; Sampson & Lauritsen, 1990).
To assess whether the associations between variables of interest persist net of control variables, multivariate analyses were conducted. First, the model in Table 2 assesses whether peer victimization is associated with depression. In support of hypothesis 1, results indicate that peer victimization is related to depression (b = 0.03, p < .01). Other significant predictors of depression include age (b = 0.05, p < .001), male (b = −0.16, p < .001), White (b = −0.16, p < .001), low self-control (b = 0.02, p < .001), and social support (b = − 0.02, p < .001). Moving forward, the focus shifts to examine the direct and indirect effects of peer victimization on violent offending.
TABLE 2.
Effect of Peer Victimization on Depression (N = 1,848)
| Variable | b | SE |
|---|---|---|
| Peer victimization | 0.03** | 0.01 |
| Age | 0.05*** | 0.01 |
| Male | −0.16*** | 0.03 |
| White | −0.16*** | 0.02 |
| Low self-control | 0.02*** | 0.00 |
| Social support | −0.02*** | 0.00 |
| Violent peers | 0.00 | 0.05 |
| Wald χ2 = 1492.09*** | ||
| McFadden’s R2 = 0.03 | ||
Note. SE = robust standard errors; b = negative binomial regression coefficients.
p < .05.
p < .01.
p < .001.
Table 3 displays the results of logistic regression models assessing the relationship between peer victimization and violent offending. Model 1 presents a logistic regression estimates examining the direct relationship between peer victimization and violent offending without the inclusion of the mediating variable (depression). Here, the interest is in whether vicarious strain occurs through having friends who are victimized. As expected, net of control variables, peer victimization is positive and significantly associated with violent offending (b = 0.22, p < .05). This finding supports hypothesis 2. Other significant predictors of violent offending include: male (b = 0.64, p < .001), white (b = −0.37, p < .001), low self-control (b = 0.03, p < .01), social support (b = −0.05, p < .001), and prior violent offending (b = 1.89, p < .001).
TABLE 3.
Effect of Peer Victimization on Violent Offending (N = 1,848)
| Variable | Model 1 | Model 2 | ||
|---|---|---|---|---|
| b | SE | b | SE | |
| Peer victimization | 0.22* | 0.10 | 0.21 | 0.12 |
| Age | −0.07 | 0.04 | −0.09* | 0.04 |
| Male | 0.64*** | 0.16 | 0.73*** | 0.14 |
| White | −0.37*** | 0.09 | −0.28** | 0.09 |
| Low self-control | 0.03** | 0.01 | 0.03* | 0.01 |
| Social support | −0.05*** | 0.01 | −0.02* | 0.01 |
| Violent peers | 0.39 | 0.28 | 0.42 | 0.25 |
| Violent offending (Wave I) | 1.89*** | 0.10 | 1.87*** | 0.10 |
| Depression | - | - | 0.05*** | 0.01 |
| Wald χ2 = 1331.59*** | Wald χ2 = 3499.78*** | |||
| McFadden’s R2 = 0.19 | McFadden’s R2 = 0.20 | |||
Note. Entries are unstandardized logistic regression coefficients (b), and robust standard errors (SE).
p < .05.
p < .01.
p < .001.
According to GST, negative emotions should mediate the relationship between strain and criminal behavior. Model 2 tests the mediating effect of depression on the relationship between peer victimization and violent offending. The model shows that depression is associated with violent offending (b = 0.05, p < .001), net of control variables. Consistent with GST, the direct effect of peer victimization on violent offending is reduced to nonsignificant after adding depression into the model. In other words, depression fully mediated the relationship between peer victimization and violent offending. Overall, these observations provide support for hypothesis 3. Regarding control variables, age, male, white, low self-control, social support, and prior violent offending were significant predictors of violent offending in the expected direction. Surprisingly, a significant effect for violent peers was not observed in either model. This finding is inconsistent with prior research (see Hoeben et al., 2016) suggesting that deviant peers is a robust predictor of criminal and deviant behavior.
Table 4 presents a series of logistic regressions examining moderating effects of violent peers and respondents involvement with peers on the relationship between peer victimization on violent offending. Model 1 test the assumption that violent peers will condition the relationship between peer victimization and violent offending. Contrary to this expectation, no evidence of moderation was found. Lastly, it is hypothesized that individuals who are highly involved with their peers, specifically violent peers, and have friends within their network who have been violently victimization (or peer victimization) would be at an increased risk of engaging in violent offending. Model 2 test this assumption by using a three-way interaction term (peer victimization × involvement with peers × violent peers). In support of hypothesis 5, individuals were likely to engage in violent offending when they were highly involved with friends who were violently victimized and associated with violent peers (b = 0.31, p < .05), net of control variables. The focus now shifts to the implications of the results for theory, research, and practice.
TABLE 4.
Effect of Peer Victimization and Involvement With Peers on Violent Offending (N = 1,848)
| Variable | Model 1 | Model 2 | ||
|---|---|---|---|---|
| b | SE | b | SE | |
| Peer victimization | 0.23* | 0.10 | 0.24* | 0.11 |
| Age | −0.11 | 0.04 | −0.12** | 0.04 |
| Male | 0.66*** | 0.15 | 0.66*** | 0.16 |
| White | −0.35*** | 0.10 | −0.36*** | 0.10 |
| Low self-control | 0.03** | 0.01 | 0.03** | 0.01 |
| Social support | −0.05*** | 0.01 | −0.05*** | 0.01 |
| Violent peers | 0.38 | 0.31 | 0.43 | 0.29 |
| Violent offending (Wave I) | 1.87*** | 0.10 | 1.88*** | 0.09 |
| Involvement | 0.09*** | 0.01 | 0.09*** | 0.01 |
| Peer victimization × violent peers | 0.47 | 0.30 | 0.41 | 0.29 |
| Peer victimization × involvement | - | - | −0.05* | 0.02 |
| Involvement × violent peers | - | - | −0.13** | 0.04 |
| Peer victimization × involvement × violent peers | - | - | 0.31* | 0.12 |
| Wald χ2 = 9576.85*** | Wald χ2 = 148158.72*** | |||
| McFadden’s R2 = 0.20 | McFadden’s R2 = 0.20 | |||
Note. Entries are unstandardized logistic regression coefficients (b), and robust standard errors (SE).
p < .05.
p < .01.
p < .001.
DISCUSSION AND CONCLUSIONS
Building on prior research (see Agnew 2002; Baron, 2009; Lin et al., 2011; Vogel & Keith, 2015), this study used friendship network data to operationalize vicarious strain (i.e., peer victimization) and investigate its effect on violent offending. Past research has relied on respondents’ perceptions to operationalize vicarious strain, which does not consider the level of involvement individuals have with their friends. Friendship network data not only provides an accurate measure of the amount of victimization and offending taking place within a network, but also the amount of time being spent with friends. Within the GST framework, victimization is a key strain that is likely to promote maladaptive coping. The effect of victimization is not limited to those who experience trauma, but also impacts the lives of those who witness it (e.g., family, friends, peer, and others).
Agnew (2001) asserts that vicarious victimization may upset younger individuals because the people they care about have been harmed, ultimately pressuring them to cope in a delinquent manner. Vicarious strain is likely to lead to violence as people attempt to prevent additional harm to those they care about, seek revenge against those they believe are responsible for the harm, and to alleviate negative emotions. What is lacking in the GST literature is the role deviant peers play in the coping process of those who are exposed to strain vicariously. Deviant peer association is a strong predictor of crime and delinquency (Hoeben et al., 2016). This study examined whether vicarious strain occurs through having friends who are victimized, and if that relationship is conditioned by violent peers, and level of involvement within a network. The findings from this research contribute to GST literature and uncovers avenues for future research.
Agnew (1992) argues that strain increases the likelihood of negative emotions (e.g., anger and frustration). In turn, these emotions create the pressure for corrective action, with deviance being one possible solution. In support of GST’s assumption that strain is associated with negative emotions, the results show that peer victimization was positively and significantly associated with depression ultimately providing support for hypothesis 1.
GST suggests that there should be a direct relationship between strain and maladaptive coping. Indeed, peer victimization was directly associated with violent offending. In support of hypothesis 2 and consistent with prior research assessing GST’s vicarious strain hypothesis, the victimization of one’s friend is significantly and positively associated with personal involvement in violent offending. Agnew argues that this relationship between strain and maladaptive coping should be mediated by negative emotions. As for hypothesis 3, the results suggest that depression fully mediates the relationship between peer victimization and violent offending. While that is the case, Agnew (1992, 2006) argues that negative emotions, such as anger and frustration, are likely to promote a criminal response to strain. The analysis in this study was limited in that only one form of negative emotions, specifically depression, was considered. Perhaps the relationship between peer victimization and violent offending is also impacted by crime provoking emotions, such as anger. Following Daniels and Holtfreter (2018), future research examining vicarious strain should investigate other forms of negative emotions (e.g., anger, frustration, anxiety, and envy) when testing GST’s vicarious stain hypothesis.
GST suggests that people are likely to respond to strain in a criminal manner if they associate with deviant peers (see Agnew, 1992, 2006). As previously mentioned, GST suggests that deviant peers promote criminal and deviant behavior among those who experience strain specifically by providing a form of support for criminal behavior, considering certain acts as an appropriate response to strain, and acting as instigators. Interestingly, having violent peers within one’s network did not significantly impact their personal behavior. However, the current study examined the moderating effect of deviant peers on the relationship between vicarious victimization and violent offending. No evidence of moderation was found, providing no support for hypothesis 4. In other words, having friends who were violently victimized and having violent peers did not have a significant effect on violent offending.
Agnew argues that strain will have a greater effect on an individual’s behavior when people he or she cares about are harmed. To further assess the effect of deviant peers in the coping process, a three-way interaction (peer victimization × involvement with peers × violent peers) effect was examined See Figure 1. In support of the final hypothesis, a three-way interaction effect was observed, suggesting that individuals who were highly involved with their peers who had experienced victimization and also associated with violent peers were at an increased risk for engaging in violence.
Figure 1.

Peer victimization × involvement with peers × violent peers on violent offending.
The findings presented in the study suggests that vicarious strain occurs through having friends who are victimized supporting Agnew’s vicarious strain hypothesis. More specifically, the victimization of individuals’ friends is associated with depression and violent offending. Moreover, high involvement with peers who had experienced violent victimization increased the risk of engaging in violence for individuals who also associated with violent peers. Deviant peers association is a robust predictor of crime and delinquency, which influence how individuals cope with strain. To gain a better understanding of the coping process, future studies testing key arguments of GST should continue investigating deviant peers as a potential intervening factor.
Another avenue for future research is the use of friendship network data to test key assumptions of GST. While there are benefits of using traditional perceptual measures of friends’ victimization, for example, using respondents’ reports of their friends’ experiences ensures that the victimization was salient enough in their mind to be recalled, there are also potential limitations. Relying on respondents’ perception of their friends’ victimization experiences and delinquent/criminal involvement presents two concerns, which are: (a) individuals may exaggerate the amount of victimization experienced by their peers as well as their peers’ delinquent and criminal involvement, or (b) their friends’ victimization and deviant behavior may be a reflection of their own. In addition, prior studies assessing GST’s vicarious strain hypothesis use measures that ask respondents to report on the victimizations of strangers, specifically asking respondents to report on how often they witnessed victimization, such as seeing somebody get shot or stabbed, beat up, or have their property stolen (e.g., Agnew, 2002; Kort-Butler, 2010; McGrath et al., 2012; Zavala & Spohn, 2013). Such measures, however, do not assess the level of involvement (or nature of relationship) that individuals have with victims.
It is also important to acknowledge the limitation of using direct reports of friends’ victimization experiences. Relying on friends’ report of their victimizations experiences assumes the respondent knows about the victimization, which may not be the case. To address this limitation, the present study relied on network data to capture the level of involvement individuals had with their peers. It is assumed that the more involved individuals are with their friends, the more likely they would not only be aware of their friends’ victimization experiences but also their friends’ experiences and behavior will have stronger impact on their personal behavior. Our findings support this assumption (i.e., hypothesis 5). Regarding coping responses, it is important to note that the current study assess the effect of peer victimization on violent offending. Future research should consider other coping outcomes (e.g., substance use and nonviolent offending) when investigating the vicarious strain hypothesis.
With regards to implications, a focus should be placed on the cycle of violence. Since violence can be cyclical, school officials and health care providers (e.g., counselors, social workers, and psychologists) should be aware that violent victimization does not only affect the lives of those who experience trauma, but also individuals who have a relationship with those who are victimized. For example, it was found that the victimization of friends is associated with personal violence. Counselors, social workers, and other health professionals should consider providing support for victims and their close relatives and friends to help them cope with tragedies (e.g., victimization) in their everyday lives. In addition, perhaps interventions in academic and professional institutions, such as peer counseling and victim–offender mediation, could be put into practice in an effort to promote prosocial forms of coping and reduce the tendency toward deviance and retaliation.
NOTE
To examine the relationship between peer victimization and depression, wave I measures were used. In reasoning, peer victimization was measured over a 12-month period (i.e., within the past 12 months prior to baseline interview), while depression was measured asking respondents within the past seven days how often they exhibit certain behaviors or emotions. The current study is focused on the assessing negative emotions closest to time strain was experienced.
Funding.
The author(s) received no specific grant or financial support for the research, authorship, and/or publication of this article.
Footnotes
Disclosure. The authors have no relevant financial interest or affiliations with any commercial interests related to the subjects discussed within this article.
REFERENCES
- Agnew R (1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30(1), 47–88. doi: 10.1111/j.1745-9125.1992.tb01093.x [DOI] [Google Scholar]
- Agnew R (2001). Building on the foundation of general strain theory: Specifying the types of strain most likely to lead to crime and delinquency. Journal of Research in Crime and Delinquency, 38(4), 319–361. doi: 10.1177/0022427801038004001 [DOI] [Google Scholar]
- Agnew R (2002). Experienced, vicarious, and anticipated strain: An exploratory study on physical victimization and delinquency. Justice Quarterly, 19(4), 603–632. doi: 10.1080/07418820200095371 [DOI] [Google Scholar]
- Agnew R (2006). Pressured into crime: An overview of general strain theory. Roxbury. [Google Scholar]
- Agnew R, Brezina T, Wright JP, & Cullen FT (2002). Strain, personality traits, and delinquency: Extending general strain theory. Criminology, 40(1), 43–72. doi: 10.1111/j.1745-9125.2002.tb00949.x [DOI] [Google Scholar]
- Agnew R, & White HR (1992). An empirical test of general strain theory. Criminology, 30(4), 475–500. doi: 10.1111/j.1745-9125.1992.tb01113.x [DOI] [Google Scholar]
- Aiken LS, & West SG (1991). Multiple regression: Testing and interpreting interactions. Sage. [Google Scholar]
- Aseltine RH Jr., Gore S, & Gordon J (2000). Life stress, anger and anxiety, and delinquency: An empirical test of general strain theory. Journal of Health and Social Behavior, 41(3), 256–275. doi: 10.2307/2676320 [DOI] [PubMed] [Google Scholar]
- Barchia K, & Bussey K (2010). The psychological impact of peer victimization: Exploring social-cognitive mediators of depression. Journal of Adolescence, 33(5), 615–623. doi: 10.1016/j.adolescence.2009.12.002 [DOI] [PubMed] [Google Scholar]
- Baron SW (2004). General strain, street youth and crime: A test of Agnew’s revised theory. Criminology, 42(2), 457–484. doi: 10.1111/j.1745-9125.2004.tb00526.x [DOI] [Google Scholar]
- Baron SW (2009). Street youths’ violent responses to violent personal, vicarious, and anticipated strain. Journal of Criminal Justice, 37(5), 442–451. doi: 10.1016/j.jcrimjus.2009.07.004 [DOI] [Google Scholar]
- Baron SW (2011). Street youths and the proximate and contingent causes of instrumental crime: Untangling anomie theory. Justice Quarterly, 28(3), 413–436. doi: 10.1080/07418825.2010.512871 [DOI] [Google Scholar]
- Broidy L (2001). A test of general strain theory. Criminology, 39(1), 9–36. doi: 10.1111/j.1745-9125.2001.tb00915.x [DOI] [Google Scholar]
- Broidy L, & Agnew R (1997). Gender and crime: A general strain theory perspective. Journal of Research in Crime and Delinquency, 34(3), 275–306. doi: 10.1177/0022427897034003001 [DOI] [Google Scholar]
- Capowich GE, Mazerolle P, & Piquero A (2001). General strain theory, situational anger, and social networks: An assessment of conditioning influences. Journal of Criminal Justice, 29(5), 445–461. doi: 10.1016/S0047-2352(01)00101-5 [DOI] [Google Scholar]
- Carson DC, Sullivan CJ, Cochran JK, & Lersch KM (2008). General strain theory and the relationship between early victimization and drug use. Deviant Behavior, 30(1), 54–88. doi: 10.1080/01639620802050023 [DOI] [Google Scholar]
- Daniels AZ, & Holtfreter K (2018). Moving beyond anger and depression: The effects of anxiety and envy on maladaptive coping. Deviant Behavior, 40(3), 334–352. doi: 10.1080/01639625.2017.1422457 [DOI] [Google Scholar]
- Eitle D, & Turner RJ (2002). Exposure to community violence and young adult crime: The effects of witnessing violence, traumatic victimization, and other stressful life events. Journal of Research in Crime and Delinquency, 39(2), 214–237. doi: 10.1177/002242780203900204 [DOI] [Google Scholar]
- Enzmann D, Marshall IH, Killias M, Junger-Tas J, Steketee M, & Gruszczynska B (2010). Self-reported youth delinquency in Europe and beyond: First results of the Second International Self-Report Delinquency Study in the context of police and victimization data. European Journal of Criminology, 7(2), 159–183. doi: 10.1177/1477370809358018 [DOI] [Google Scholar]
- Fergusson DM, Swain-Campbell NR, & Horwood LJ (2002). Deviant peer affiliations, crime and substance use: A fixed effects regression analysis. Journal of Abnormal Child Psychology, 30(4), 419–430. doi: 10.1023/A:1015774125952 [DOI] [PubMed] [Google Scholar]
- Guterman NB, Hahm HC, & Cameron M (2002). Adolescent victimization and subsequent use of mental health counseling services. Journal of Adolescent Health, 30(5), 336–345. doi: 10.1016/S1054-139X(01)00406-2 [DOI] [PubMed] [Google Scholar]
- Harris KM, C.T Halpern E, Whitsel E, Hussey J, Tabor J, Entzel P & Udry JR (2009). The National Longitudinal Study of Adolescent to Adult Health: Research Design [WWW document]. https://addhealth.cpc.unc.edu//documentation/study-design.
- Hawker DS, & Boulton MJ (2000). Twenty years’ research on peer victimization and psychosocial maladjustment: A meta-analytic review of cross-sectional studies. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 41(4), 441–455. doi: 10.1111/1469-7610.00629 [DOI] [PubMed] [Google Scholar]
- Hay C, & Evans MM (2006). Violent victimization and involvement in delinquency: Examining predictions from general strain theory. Journal of Criminal Justice, 34(3), 261–274. doi: 10.1016/j.jcrimjus.2006.03.005 [DOI] [Google Scholar]
- Hay C, & Meldrum R (2010). Bullying victimization and adolescent self-harm: Testing hypotheses from general strain theory. Journal of Youth and Adolescence, 39(5), 446–459. doi: 10.1007/s10964-009-9502-0 [DOI] [PubMed] [Google Scholar]
- Haynie DL (2002). Friendship networks and delinquency: The relative nature of peer delinquency. Journal of Quantitative Criminology, 18(2), 99–134. doi: 10.1023/A:1015227414929 [DOI] [Google Scholar]
- Haynie DL, & Osgood DW (2005). Reconsidering peers and delinquency: How do peers matter? Social Forces, 84(2), 1109–1130. doi: 10.1353/sof.2006.0018 [DOI] [Google Scholar]
- Haynie DL, & Payne DC (2006). Race, friendship networks, and violent delinquency. Criminology, 44(4), 775–805. doi: 10.1111/j.1745-9125.2006.00063.x [DOI] [Google Scholar]
- Henry DB, Tolan PH, & Gorman-Smith D (2001). Longitudinal family and peer group effects on violence and nonviolent delinquency. Journal of Clinical Child Psychology, 30(2), 172–186. doi: 10.1207/S15374424JCCP3002_5 [DOI] [PubMed] [Google Scholar]
- Hoeben EM, Meldrum RC, Walker D, & Young JT (2016). The role of peer delinquency and unstructured socializing in explaining delinquency and substance use: A state-of-the-art review. Journal of Criminal Justice, 47, 108–122. doi: 10.1016/j.jcrimjus.2016.08.001 [DOI] [Google Scholar]
- Kort-Butler LA (2010). Experienced and vicarious victimization: Do social support and self-esteem prevent delinquent responses? Journal of Criminal Justice, 38(4), 496–505. doi: 10.1016/j.jcrimjus.2010.04.019 [DOI] [Google Scholar]
- Lauritsen JL, Sampson RJ, & Laub JH (1991). The link between offending and victimization among adolescents. Criminology, 29(2), 265–292. doi: 10.1111/j.1745-9125.1991.tb01067.x [DOI] [Google Scholar]
- Lin WH, Cochran JK, & Mieczkowski T (2011). Direct and vicarious violent victimization and juvenile delinquency: An application of general strain theory. Sociological Inquiry, 81(2), 195–222. doi: 10.1111/j.1475-682X.2011.00368.x [DOI] [PubMed] [Google Scholar]
- Luk JW, Wang J, & Simons-Morton BG (2010). Bullying victimization and substance use among US adolescents: Mediation by depression. Prevention Science, 11(4), 355–359. doi: 10.1007/s11121-010-0179-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Manasse ME, & Ganem NM (2009). Victimization as a cause of delinquency: The role of depression and gender. Journal of Criminal Justice, 37(4), 371–378. doi: 10.1016/j.jcrimjus.2009.06.004 [DOI] [Google Scholar]
- Mazerolle P, & Maahs J (2000). General strain and delinquency: An alternative examination of conditioning influences. Justice Quarterly, 17(4), 753–778. doi: 10.1080/07418820000094751 [DOI] [Google Scholar]
- Mazerolle P, Piquero AR, & Capowich GE (2003). Examining the links between strain, situational and dispositional anger, and crime: Further specifying and testing general strain theory. Youth & Society, 35(2), 131–157. doi: 10.1177/0044118X03255029 [DOI] [Google Scholar]
- McGloin J (2009). Delinquency balance: Revisiting peer influence. Criminology, 47(2), 439–477. doi: 10.1111/j.1745-9125.2009.00146.x [DOI] [Google Scholar]
- McGloin JM, & Shermer L (2009). Self-control and deviant peer network structure. Journal of Research in Crime and Delinquency, 46(1), 35–72. doi: 10.1177/0022427808326585 [DOI] [Google Scholar]
- McGrath SA, Marcum CD, & Copes H (2012). The effects of experienced, vicarious, and anticipated strain on violence and drug use among inmates. American Journal of Criminal Justice, 37(1), 60–75. doi: 10.1007/s12103-011-9127-1 [DOI] [Google Scholar]
- Mitchell KJ, Ybarra M, & Finkelhor D (2007). The relative importance of online victimization in understanding depression, delinquency, and substance use. Child Maltreatment, 12(4), 314–324. doi: 10.1177/1077559507305996 [DOI] [PubMed] [Google Scholar]
- Moon B, & Morash M (2017). Gender and general strain theory: A comparison of strains, mediating, and moderating effects explaining three types of delinquency. Youth & Society, 49(4), 484–504. doi: 10.1177/0044118X14541877 [DOI] [Google Scholar]
- Morgan ER, & Kena G (2017). Criminal victimization, 2016. Bureau of Justice Statistics. [Google Scholar]
- Nofziger S, & Kurtz D (2005). Violent lives: A lifestyle model linking exposure to violence to juvenile violent offending. Journal of Research in Crime and Delinquency, 42(1), 3–26. doi: 10.1177/0022427803262061 [DOI] [Google Scholar]
- Radloff LS (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1(3), 385–401. doi: 10.1177/014662167700100306 [DOI] [Google Scholar]
- Sampson RJ, & Lauritsen j. L. (1990). Deviant lifestyles, proximity to crime, and the offender-victim link in personal violence. Journal of Research in Crime and Delinquency, 27(2), 110–139. [Google Scholar]
- Schreck CJ, & Fisher BS (2004). Specifying the influence of family and peers on violent victimization: Extending routine activities and lifestyles theories. Journal of Interpersonal Violence, 19(9), 1021–1041. doi: 10.1177/0886260504268002 [DOI] [PubMed] [Google Scholar]
- Sullivan TN, Farrell AD, & Kliewer W (2006). Peer victimization in early adolescence: Association between physical and relational victimization and drug use, aggression, and delinquent behaviors among urban middle school students. Development and Psychopathology, 18(1), 119–137. doi: 10.1017/S095457940606007X [DOI] [PubMed] [Google Scholar]
- Tharp-Taylor S, Haviland A, & D’Amico EJ (2009). Victimization from mental and physical bullying and substance use in early adolescence. Addictive Behaviors, 34(6–7), 561–567. doi: 10.1016/j.addbeh.2009.03.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tseloni A, Mailley J, Farrell G, & Tilley N (2010). Exploring the international decline in crime rates. European Journal of Criminology, 7(5), 375–394. doi: 10.1177/1477370810367014 [DOI] [Google Scholar]
- Van, Ryzin. J,M, Fosco GM, & Dishion TJ (2012). Family and peer predictors of substance use from early adolescence to early adulthood: An 11-year prospective analysis. Addictive Behaviors, 37(12), 1314–1324. doi: 10.1016/j.addbeh.2012.06.020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vogel M, & Keith S (2015). Vicarious peer victimization and adolescent violence: Unpacking the effects of social learning, general strain, and peer group selection.d.eviant Behavior. 36(10), 834–852. doi: 10.1080/01639625.2014.977187 [DOI] [Google Scholar]
- Walker D, & Holtfreter K (2019). Teen pregnancy, depression, and substance abuse: The conditioning effect of deviant peers. Deviant Behavior, 1–16. doi: 10.1080/01639625.2019.1666610 [DOI] [PMC free article] [PubMed] [Google Scholar]
- Widom CS, Marmorstein NR, & Raskin White H (2006). Childhood victimization and illicit drug use in middle adulthood. Psychology of Addictive Behaviors, 20(4), 394–403. doi: 10.1037/0893-164X.20.4.394 [DOI] [PubMed] [Google Scholar]
- Wight RG, Botticello AL, & Aneshensel CS (2006). Socioeconomic context, social support, and adolescent mental health: A multilevel investigation. Journal of Youth and Adolescence, 35(1), 111–126. doi: 10.1007/s10964-005-9009-2 [DOI] [Google Scholar]
- Zavala E, & Spohn RE (2013). The role of vicarious and anticipated strain on the overlap of violent perpetration and victimization: A test of general strain theory. American Journal of Criminal Justice, 38(1), 119–140. doi: 10.1007/s12103-012-9163-5 [DOI] [Google Scholar]
