Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Feb 21.
Published in final edited form as: Law Hum Behav. 2019 Apr 18;43(3):250–262. doi: 10.1037/lhb0000327

Proximal Predictors of Gun Violence Among Adolescent Males Involved in Crime

Zachary R Rowan 1, Carol A Schubert 2, Thomas A Loughran 3, Edward P Mulvey 4, Dustin A Pardini 5
PMCID: PMC8860145  NIHMSID: NIHMS1604686  PMID: 30998029

Abstract

The growing public health and legal concerns regarding gun violence has led to a call for research that investigates risk factors for gun violence across a variety of domains. Individual and sociocontextual risk factors have been associated with violence more broadly, and in some instances gun-carrying, however no prior research has investigated the role of these factors in explaining gun violence using longitudinal data. The current study utilized a subsample (N = 161) from the Pathways to Desistance Study, which is a longitudinal sample of serious adolescent offenders to evaluate interindividual and intraindividual differences in relevant risk factors of gun violence. Results suggest that there are a few key proximal individual-level and sociocontextual predictors for gun violence, including witnessing nongun violence, future orientation, and perceived personal rewards to crime. Findings demonstrate the salience of exposure to violence in contributing to gun violence and identify levers of action for public policy.

Keywords: gun violence, risk factors, adolescent violence

The marked rise in youth violence in the United States during the early 1990s elevated the issue of gun violence as both a public health and criminal justice concern. Although there has been a recent decline in male youth reporting weapon carrying (13.7% to 8.7%; Centers for Disease Control and Prevention [CDC], 2017a) and general declines in overall firearm violence since the 1990s (Planty & Truman, 2013), nearly 36,000 people still die every year because of gun-related violence, with two more persons injured for every person killed (CDC, 2017b). This violence is concentrated in certain locales and groups, with young adults, males, and minorities disproportionately affected (Wintemute, 2015). Yet there is a substantial lack of research investigating the precursors and correlates of gun violence. As a result, the Institute of Medicine and National Research Council (Leshner, Altevogt, Lee, & McCoy, 2013) and the American Psychological Association (2013) have called for research designed to identify risk factors for gun violence across multiple domains. To address this void, the current study adopts an ecological approach to understand and predict gun violence as it recognizes the influence of multiple contexts (i.e., individual characteristics, social networks, and the community) in explaining this behavior (Farrell & Flannery, 2006; Hawkins et al., 2000).

In general, studies discussing gun violence focus on factors related to gun carrying or factors related to the etiology of violence more broadly (e.g., Lizotte, Krohn, Howell, Tobin, & Howard, 2000; Resnick, Ireland, & Borowsky, 2004; Spano, Pridemore, & Bolland, 2012). This body of work emphasizes the constellation of individual and sociocontextual risk factors that interact and develop over time to explain violence, as compared to the role of any one specific risk factor (e.g., Dodge, Greenberg, Malone, & Conduct Problems Prevention Research Group, 2008). Prior work using the Pathways to Desistance Study has evaluated predictors of homicide (DeLisi, Piquero, & Cardwell, 2016) and has also adopted a similar ecological model to consider the risk factors of gun carrying (Beardslee, Docherty, Mulvey, Schubert, & Pardini, 2018). It is notable, however, that there is limited work using longitudinal data to identify and specifically examine risks for gun violence and, importantly, test how the development of risk factors within-individuals predicts gun violence during adolescence and young adulthood. The paucity of research on gun violence is not entirely surprising when one considers the fact that since 1995 Congress stopped federal funding to the Centers for Disease Control to study gun violence (Kaplan, 2018). Thus, efforts to bypass existing governmental barriers to study gun violence by using publicly available data offer an exciting opportunity to address this void and offer valuable empirically based findings to counteract some prevalent myths currently driving public discourse on this topic.

The current study evaluates proximal individual and sociocontextual risk factors of firearm use in a longitudinal sample of high-risk adolescents with a history of engaging in serious criminal offenses. In doing so, it contributes to research investigating gun violence in two ways. First, it examines a range of individual and sociocontextual risk factors for gun violence in a highly policy relevant group; serious adolescent offenders at high risk for involvement in harm associated with criminal gun violence. Second, it evaluates whether intraindividual changes in key risk factors increase the likelihood that these adolescents are involved in gun violence. This offers a methodologically rigorous test of how changes risk factors within individuals’ leads to involvement in gun violence. As a result, this study offers insight into the factors that could be most relevant to focus upon as antecedents of gun violence in prevention or intervention efforts with adolescents at increased likelihood to engage in this behavior.

Most extant research in this area has focused on predicting between-person differences in gun carrying, the assumed primary precursor to involvement in gun violence. For instance, Hemenway, Prothrow-Stith, Bergstein, Ander, and Kennedy (1996) surveyed seventh and 10th grade studens and found that substance use (i.e., cigarette, alcohol), poor academic performance, and being male significantly predicted gun carrying. Gun carrying has also been shown to be related to participation in gangs, drug selling, and whether an individual’s peer group owns a weapon for protection (e.g., Hayes & Hemenway, 1999; Lizotte, Howard, Krohn, & Thornberry, 1997). At the contextual level, Molnar, Miller, Azrael, and Buka (2004) provided evidence, using the Project on Human Development in Chicago Neighborhoods data, that the social and physical disorder of a neighborhood were also positively associated with carrying concealed firearms. Although research investigating patterns of gun carrying is an important contribution toward reducing gun violence, individuals who engage in firearm use likely represent a small subset of those that opt to carry.

Individual-Level Risk Factors

Individual level risk factors for the use of firearms are often assumed to be some subset of predictors of violence (or guncarrying) more broadly. One of the most widely discussed predictors of aggressive or violent behavior later in life is a history of early aggression (e.g., Huesmann, Eron, Lefkowitz, & Walder, 1984; Nagin & Tremblay, 1999; Thornberry, Huizinga, & Loeber, 1995). The stability of aggression over key developmental periods contributes toward individuals acquiring a hostile attribution bias and normative beliefs that condone the use of aggressive behavior across a range of situations (e.g., Dodge & Frame, 1982; Hawkins et al., 2000). Thus, among youth carrying a weapon, those that exhibit higher levels of hostility or aggression may be more likely to respond to a perceived hostile situation by resorting to gun violence (Reid, Richards, Loughran, & Mulvey, 2017). Relatedly, highly impulsive adolescents who lack the capacity to account for future consequences may be more likely to use a weapon (e.g., Hawkins et al., 2000). For example, Piquero, MacDonald, Dobrin, Daigle, & Cullen (2005) evaluated the role of self-control in predicting violent offending and homicide victimization and found that self-control was significantly related to both outcomes. Although offenses other than those involving a weapon may contribute to homicide victimization, the lethality of gun use and association of guns with homicides is suggestive of the potential interdependence between elements of low-self-control, engaging in violence, and involvement in situations where weapons are likely (e.g., Hepburn & Hemenway, 2004; Piquero et al., 2005; Siegel, Ross, & King, 2013). In the context of a violent transaction, individuals who more severely discount future actions and consequences may be less motivated to refrain from weapon use. Each of these factors is part of a constellation of cognitive factors, many of which are developmental in nature, that are demonstrated to have a relationship to risky behavior during adolescence and young adulthood.

Mental illness and substance abuse are often considered as factors related to gun use (McGinty, Webster, Jarlenski, & Barry, 2014; Monahan, 1992). However, most people with mental illness (exclusive of substance use) are no more likely to specifically engage in gun violence (e.g., Elbogen & Johnson, 2009; Steadman, Monahan, Pinals, Vesselinov, & Robbins, 2015; Swanson, McGinty, Fazel, & Mays, 2015; Van Dorn, Volavka, & Johnson, 2012). Substance abuse, on the other hand, tends to be more strongly linked to involvement in gun violence. Prior cross-sectional research demonstrates that heavy alcohol drinking and alcohol abuse increases the likelihood of gun possession (e.g., Loh et al., 2010), self-inflicted gun injury/suicide (e.g., Miller, Hemenway, & Wechsler, 2002), and homicide (e.g., Hohl et al., 2017).

The use of a firearm may also contribute to the development of an individual’s social identity and status because of the intrinsic rewards attached to using a weapon (e.g., Fagan & Wilkinson, 1998) and thus could increase the likelihood of gun use. For instance, Anderson (1999) argued that the code of the street often demanded that individuals demonstrate masculinity, toughness, and even aggressive or violent behavior in order to acquire status and respect. Access to a gun (and the use of one) could be a symbolic—and practical—display of power that would contribute to maintaining status for adolescents (Anderson, 1999; Fagan & Wilkinson, 1998; Katz, 1988). Individuals who view criminal behavior as an inherently rewarding endeavor may be more likely to use a gun (Loughran, Reid, Collins, & Mulvey, 2016). Indeed, responsiveness to rewards associated with criminal behavior has been shown to be an important contributor to the understanding of offender-decision making, although prior work has not explicitly tested whether the perceived rewards to crime contribute to gun use, as they appear to do for more general criminal involvement (e.g., Loughran, Nguyen, Piquero, & Fagan, 2013; Nguyen & Bouchard, 2013).

Sociocontextual Risk Factors

It is important to also recognize that individuals are situated within social contexts, which add an additional dimension to understanding risks for criminal involvement (e.g., Elliott et al., 1996; Sampson, Raudenbush, & Earls, 1997). An adolescent’s social context, for example, can provide higher levels of opportunities for antisocial behavior or lower levels of social control. In line with these types of influences, prior research has shown the importance of constructing and maintaining a more structured prosocial set of activities as part of the process of reducing criminal involvement (Maruna, 2000). Completing school and obtaining gainful employment (e.g., Mesters, van der Geest, & Bijleveld, 2016; Sampson & Laub, 1993; Savolainen, 2009; Uggen, 2000) affect both the types of situations individuals encounter daily and the level of commitment to the prosocial world, making criminal behavior costlier (e.g., Laub & Sampson, 2003; Sampson & Laub, 1993). Relatedly, individuals who spend more time socializing in unstructured settings generally report higher levels of offending and violence (e.g., Maimon & Browning, 2010; Osgood & Anderson, 2001; Osgood, Wilson, O’Malley, Bachman, & Johnston, 1996). For example, Maimon and Browning (2010) demonstrated, using the Project on Human Development in Chicago Neighborhoods (PHDCN) data, that there was variation across neighborhoods in unstructured socializing with peers and that unstructured socializing with peers was significantly related to violent behavior.

Given the evidence to suggest that there are differences in the concentration of violence across neighborhoods (e.g., Braga, Papachristos, & Hureau, 2010; Weisburd, 2015), the variation in exposure to violence may lead individuals to be more likely to carry a weapon and engage in violent behavior (e.g., Bingenheimer, Brennan, & Earls, 2005; Guerra, Huesmann, & Spindler, 2003; Loeber et al., 2005). Elevated levels of violent crime within a neighborhood may motivate individuals to obtain a weapon for protection and also increase the situations which facilitate (or demand) the use of weapons (e.g., Cao, Cullen, & Link, 1997; Spano & Bolland, 2013). For example, Bingenheimer et al. (2005) used a propensity score stratification strategy with data on adolescents from Chicago to examine the relationship between exposure to violence and subsequent violent offending. They found that adolescents exposed to violence nearly doubled the probability of engaging in violence compared to those adolescents who were not exposed. It is also possible that exposure to violence is related to gang involvement and a subsequent increased probability of involvement in violent behavior. The nexus between these factors, however, is still unclear, although some longitudinal research has indicated that involvement in gangs contributes to increasing levels of violence and the use of a gun (e.g., Braga et al., 2010; Lizotte et al., 2000; Spano & Bolland, 2011; Spano et al., 2012).

Consideration of individual-level risk factors within a broader ecological approach can provide a more comprehensive depiction of the factors which facilitate the use of firearms. Further, the current study extends existing work that explores the precursors to gun-carrying by explicitly focusing on the proximal factors associated with gun violence (e.g., Beardslee et al., 2018). The sample is limited to a highly selective group of males who reported shooting or shooting at someone (Self-Report of Offending Scale; Huizinga, Esbensen, & Weiher, 1991) in this 3-year period. Despite the vast literature on violence broadly and a smaller body of work focused on gun carrying, there is limited research explicitly evaluating the risks of gun violence. Our objective is to be inclusive of key individual and sociocontextual factors that are theoretically important to gun violence and grounded in prior work on violence broadly and gun carrying.

Existing work using the Pathways data (e.g., Baskin & Sommers, 2014; Cardwell & Piquero, 2018; Reid et al., 2017) and more general reviews (e.g., Farrington & Loeber, 2000) have devoted considerable attention to the study of risk factors for violence and gun-carrying. For instance, Cardwell and Piquero (2018) adopt a similar risk-factor approach to understand predictors of violence during adolescence and evaluate the stability of involvement of violence into young adulthood. Cardwell and Piquero (2018) conclude that gang involvement and exposure to violence were significantly associated with odds of subsequent violent behavior. Similarly, Baskin and Sommers (2014) utilize Pathways data to observe how exposure to community violence is related to different trajectories of violent behavior and conclude that exposure to violence led individuals to be more embedded in patterns of violent criminal behavior. The goal of the current study is not to provide a comparative analysis of risk factors for gun violence perpetrators and gun carriers (or to violence more generally) but it is to prioritize an understanding of the risks associated specifically with gun violence that can advance the broader literature and provide additional insight into the consistency in findings across empirical work. The Pathways to Desistance Study offers a valuable source of data to evaluate the effects of proximal time-varying risk factors in a high-risk sample and enables a longitudinal consideration of how within-individual changes in risk factors predict gun use.

Method

The Pathways to Desistance (“Pathways”) study is a comprehensive longitudinal study of serious adolescent offenders from two urban locales (Philadelphia, PA and Phoenix, AZ). The Pathways study followed a sample of 1,354 serious adolescent offenders that were enrolled into the study from adolescence into early adulthood (a 7-year period) to examine the effects of changes in developmental, social contextual, and intervention-related experiences on criminal offending in early adulthood. Recruitment occurred from November 2000 through January 2003 in the juvenile court system in Maricopa County (Phoenix), Arizona or Philadelphia County, Pennsylvania. After providing informed consent, youth enrolled in the Pathways study participated in a baseline interview followed by a series of 10 interviews (“time-point interviews”) which occurred at 6-month intervals for the first three years and annually thereafter through seven years. Information regarding the theoretical foundation for the study can be found in Mulvey et al. (2004); details regarding recruitment, a description of the full sample, and the study methodology are discussed in Schubert et al. (2004). The study was reviewed and approved by the Institutional Review Boards of the University of Pittsburgh (the study coordinating center), Arizona State University, and Temple University (data collection sites). More information about the study can also be found at www.pathwaysstudy.pitt.edu.

Participants

The current study involves a 3-year examination of selective group of males who reported shooting or shooting at someone (Self-Report of Offending Scale; Huizinga et al., 1991) in this 3-year period (n = 163). We concentrated on the initial 3-year period because the recall period for these interview waves was uniformly 6 months, ensuring that the covariates examined were reasonably proximal to gun violence. Furthermore, the initial 3-year period was chosen because the sample would, on average, range from ages 16 to 19, thus providing a view of an active and important period of development.

Several other exclusion criteria were used to isolate the analytic sample. As noted earlier, males are disproportionately exposed to gun- and nongun violence and also at high risk for carrying guns and shooting others (e.g., CDC, 2017b; Hemenway et al., 1996). Indeed, 14% (163/1170) of the male study participants reported shooting a gun in this 3-year period compared to less than 4% (7/184) of the female study participants. Given that the majority of gun violence is perpetrated by males and the small number of female shooters present in the data, we restrict our sample to only include males. The sample was further limited by restricting the recall periods examined to those in which the youth was in the community (not in placement or jail) for at least 10% of the recall period time. To have spent at least 10% of the recall period in the community may seem like a small amount of time, but because of the relatively high frequency of time out of the community in this sample (Schubert & Mulvey, 2014) it was important to balance concerns about biasing the analysis with ensuring adequate opportunity to engage in gun violence. This was also done to ensure that the proximal factors of interest reflected the “usual” circumstances for the youth rather than time in a more constricted institutional environment (time in a facility limits access to drugs and alcohol as well as opportunities for gun violence; see Chassin, Knight, Vargas-Chanes, Losoya, & Naranjo, 2009; Piquero et al., 2001). This latter criterion eliminated two of a possible 163 males from the sample; these two males had no recall periods with the minimal required time in the community.

The final sample for the study included 161 unique males observed for a total possible 966 recall periods; however, listwise deletion of cases that have missing values on key variables results in a maximum of 626 cases (mean number of observations per adolescent = 3.88; SD = 1.50). Table 1 provides descriptive statistics on the analytic sample used. These males were, on average, age 16.53 (SD = 1.02) at the time of the baseline interview and were predominately minority youth. On average, these youth had 3.39 (SD = 2.27) petitions to court before the baseline interview and 85% were rearrested within the 3 years.

Table 1.

Descriptive Statistics on Analytic Sample

Variable M (SD) or %
Age (mean years at baseline) 16.53 (1.02)
Race/ethnicity
 White 12.4
 Black 43.5
 Hispanic 37.9
 Other 6.2
Site
 Philadelphia 48.0
 Phoenix 52.0
Prior petitions to court (mean at baseline) 3.39 (2.27)
Gun violence (mean frequency at baseline) 6.45 (16.76)
Open in a new tab

As noted, all these young men reported shooting a gun in at least one wave between the time of the baseline interview and the end of the three years; 24 (14.9%) of these youth also reported shooting a gun prior to the baseline interview. The majority of the sample (73.3%) reported shooting a gun in just one recall period postbase-line, whereas 19.3% reported shooting in two waves and the remaining 7.4% reported shooting in three or four recall periods. Youth most frequently reported shooting or shooting at someone across all 3 years just one time and on average reported shooting or shooting at someone 6.45 times (SD = 16.76). Two of these youth were deceased before the end of the 3-year period, with gun violence the cause of death for one of them.

Measures

As noted earlier, one of the primary goals of study was to examine individual and sociocontextual risk factors around the time of a shooting incident. Extant literature was used to identify a reasonable subset of factors from the Pathways interview battery to include in the analyses. Table 2 provides descriptive statistics on the overall mean, between person variance, and within person variance of the measures used in the analyses.

Table 2.

Descriptive Statistics on Measures Across Study Period

SD
Measure M Between Within Minimum Maximum
Hostility .78 .65 .53 0 4
Poor future orientation 2.54 .44 .39 1 4
Personal rewards to crime 2.95 2.16 1.55 0 10
Alcohol use 1.68 .99 .98 0 4
Time in ungainful activity .57 .31 .30 0 1
Neighborhood disadvantage .01 .87 .44 −1.68 3.36
Witness nongun violence 1.51 .99 1.04 0 6
Victim nongun violence .39 .54 .53 0 6
Gang involvement (% involved) .28 0 1
Open in a new tab

Individual factors.

Hostility was captured with a subscale of the Brief Symptom Inventory that included five items in which participants rate the extent to which they have been bothered (0 = not at all to 4 = extremely; α = .75 at baseline) in the last week by various symptoms (e.g., “Having urges to break or smash things”; Derogatis & Melisaratos, 1983). The hostility indicator used in the analysis reflected the mean of the five items. This subscale was selected to capture an individual’s proclivity toward aggression, which has been associated with subsequent violent behavior (e.g., Dodge & Frame, 1982; Hawkins et al., 2000). Given existing evidence suggesting that individual’s that engage in a greater amount of discounting of the future are more likely to be involved in situations involving violence and/or a weapon, youth’s orientation to the future was assessed using a single scale (e.g., Hepburn & Hemenway, 2004; Piquero et al., 2005; Siegel et al., 2013), the Future Outlook Inventory (Cauffman & Woolard, 1999). The Future Outlook Inventory includes 15 items that participants rank from 1 (never true) to 4 (always true; α = .68 at baseline) how strongly statements reflect how they usually are (e.g., “Before making a decision, I weigh the good vs. the bad”). This inventory specifically captures the decision-making priorities of individuals (i.e., assessment of future consequences), which is relevant to an understanding of the decision to engage in a serious behavior such as gun violence. The items were reverse coded and a summary score representing the mean across the 15 items was calculated, with higher scores indicate poorer future orientation. Frequency of use of alcohol was recoded to reflect level of use from 0 to 4 (0 = abstinence, 1 = 1–5 times in recall period, 2 = 1–3 times per month, 3 = 1–3 times per week, 4 = 4–5 times per week or daily; see the Substance Use Inventory by Sher, 1987). Based on the potential intrinsic rewards associated with gun use, personal rewards from crime (e.g., “How much ‘thrill’ or ‘rush’ is it to break into a store or home?”; = .88 at baseline) was assessed using the mean of a seven-item measure on a scale of 0 to 10 (0 = no fun or kick at all to 10 = a great deal of fun or kick) and was developed for the Pathways study based on the work of Nagin and Paternoster (1994).

Sociocontextual factors.

Time not spent in gainful activity reflects the proportion of the recall period during which the youth was not working at least part time and/or attending school without missing 5 or more days a month. Gang involvement during the recall period was assessed in a series of questions to identify whether a participant reported being a member of a gang (1) or not in a gang (0) in the recall period (e.g., Thornberry, Lizotte, Krohn, Farnworth, & Jang, 1994). The degree of neighborhood disadvantage was rated for each recall period using the primary community address (community address at which the youth lived the longest during the relevant recall period). The neighborhood disadvantage indicator reflects a census block composite score of a range of factors including percentage poverty, percentage unemployment, percentage on welfare, percentage of single-parent households (Health Innovation Program, 2000). Exposure to nongun violence was assessed using a modified version of the Exposure to Violence Inventory (Selner-O’Hagan, Kindlon, Buka, Raudenbush, & Earls, 1998). Six items assessing exposure to nongun violence captured whether the youth had been beaten up, raped, or chased by someone who wanted to seriously hurt them during the recall period and whether they had seen others victimized in the same manner. Two variables reflect a summed variety score of the victimization experiences that were witnessed or experienced. A correlation matrix of the measures used across three years is included in an appendix (see the Appendix).

Analysis

The study aims were addressed in two main sets of analyses. Variables that had research evidence for their association with gun violence were tested for their relationship to gun use in both a between person specification (i.e., pooled logistic regression) and within person specification (fixed effects logistic regression). Across both modeling strategies, risk factors were measured contemporaneously to the gun use outcome to capture the proximal relationship. Though this constrains our ability to establish temporal ordering, specifically in the latter model, we are careful to consider this limitation in our interpretation. Importantly, our estimation strategy eliminates confounding from all fixed measures, and hence, we only include time varying constructs at each wave as within-person predictors. In the event of missing data across each of the models, we implemented listwise deletion.

To first estimate the cross-sectional relationship between these factors and gun use across individuals, a binary outcome (shoot/did not shoot) was regressed on the individual and sociocontextual risk factors listed in the preceding text, using a pooled logistic regression model with clustered robust standard errors. The second goal of the study was to estimate within-individual changes in proximal risk factors for gun use. This strategy is suited to the current research question because it controls for all observed and nonobserved time-stable factors that vary between individuals (e.g., race, environmental influences not observed; Allison, 2009). As a result, we are able to address concerns associated with time-stable omitted variable bias that may explain observed relationships between within person changes in proximal risk factors and gun violence. Additionally, we explore a fundamentally different question regarding how intraindividual changes in proximal risk factors are associated with gun use, which provides an opportunity to understand how shifts in individual and sociocontextual risk factors precipitate adolescent involvement in gun violence.

Because the magnitude of odds ratios for continuous variables in logistic regression model are impacted by the scaling of the independent variable, all continuous predictors were z-score transformed prior to the analyses to place them on comparable metric. Additionally, because estimates from pooled logistic regression models are based on a combination of between- and within-individual variation, whereas fixed effects models focus solely on within-individual variation, the predictors were standardized differently for these two models. Specifically, the combined between/within standard deviation across the time series was used to calculate z scores for the pooled logistic models, and the within-person standard deviation was used to calculate z scores for the fixed effects estimation. Following this transformation, model parameters represent the estimated change in the odds of the outcome per one standard deviation increase in the predictor variable. These standardized odds ratios can be converted back to the original predictor scaling using the between/within and within-individual standard deviation values reported in Table 2 (i.e., ORraw = exp(ln[ORstandardized]/SD).

Results

We first present results from a pooled cross-sectional logistic regression that includes all individual and sociocontextual covariates. All results are presented as odds ratios. Of note, we present models with additional specifications to address the fact that many youths in the selected sample were missing data on the measures of hostility symptoms (112 recall period observations) and neighborhood disadvantage (60 recall period observations). We estimate models excluding these measures to avoid substantial listwise deletion resulting in further reductions in sample size.

Table 3 reports the pooled logistic regression results for several model specifications. Among the individual factors, a standard deviation increase in poor future orientation increases the odds of engaging in gun violence by 1.45 times (SE = .12, p = .003, 95% CI [1.13, 1.85]). Similarly, individuals with higher perceived personal rewards to engaging in crime are 1.42 times more likely to engage in gun violence (SE = .13, p = .005, 95% CI [1.11, 1.82]). There is no statistically significant relationship between an individual’s hostility or alcohol use and the likelihood of engaging in gun violence. Several sociocontextual factors are also significantly related to gun violence. For every additional standard deviation increase in the months that individuals indicate they were not either working or in school, the likelihood of engaging in gun violence increases by about 1.47 times (SE = .12, p = .002, 95% CI [1.16, 1.86]). Witnessing nongun violence consistently increases the likelihood of gun violence. Specifically, individuals are 2.63 times more likely to engage in gun violence for every standard deviation increase in the experience of witnessing nongun violence (SE = .12, p < .001, 95% CI [2.07, 3.34]). Similarly, victimization by nongun violence significantly increases the likelihood of engaging in gun violence by about 1.36 times for each additional standard deviation increase in the number of victimizations reported (SE = .11, p = .004, 95% CI [1.10, 1.67]). Neighborhood disadvantage and gang involvement did not significantly predict the likelihood of engaging in gun violence. Due to the large percentage of missing data on the measures of hostility and neighborhood disadvantage, models were run excluding these measures to determine whether the results from the full model were not influenced by missing data. As reported in Table 3, the results are generally robust across each modeling specification.

Table 3.

Pooled Logistic Regression Predicting Gun Violence

Full model: Model 1a Hostility removed: Model 1b Hostility and ND removed: Model 1c
Measure OR 95% CI OR 95% CI OR 95% CI
Individual factors
Hostility .85 .67–1.07
Poor future orientation 1.45** 1.13–1.85 1.45*** 1.17–1.79 1.46*** 1.19–1.79
Alcohol use .92 .73–1.14 1.01 .83–1.70 1.02 .84–1.21
Personal rewards for crime 1.42** 1.11–1.82 1.37** 1.11–1.70 1.25** 1.01–1.54
Sociocontextual factors
Time in ungainful activity 1.47** 1.16–1.86 1.37** 1.10–1.72 1.38** 1.10–1.71
Neighborhood disadvantage (ND) 1.10 .88–1.39 1.17 .96–1.44
Witness nongun violence 2.63*** 2.07–3.34 2.31*** 1.86–2.86 2.27*** 1.86–2.78
Victimization nongun violence 1.36* 1.10–1.67 1.29** 1.07–1.56 1.26** 1.05–1.51
Gang involvement 1.16 .74–1.83 1.05 .69–1.83 1.08 .73–1.59
Open in a new tab

Note. OR = odds ratio; CI = confidence interval.

*

p < .05.

**

p < .01.

***

p < .001.

Table 4 reports similar models using a fixed-effects estimation strategy. These models focus on the association between within-individual changes in the study risk factors and gun violence across time, thereby eliminating any time-stable factors that vary between-individuals as potential confounds. These models are useful for testing the hypothesis that an adolescent is more prone to engage in gun violence during periods when their exposure to specific risk factors increases or decreases across a time series. Several statistically significant findings emerge. At the individual level, only within-person increases in perceived personal rewards to crime is associated with increases in the likelihood of gun violence. Specifically, when youth experienced a standard deviation increase in their perceived personal rewards to crime, their odds of engaging in gun violence increased by 1.32 times (SE = .13, p = .029, 95% CI [1.03, 1.70]). There were no statistically significant relationships between gun violence and within-person changes in hostility, future orientation, and alcohol use. At the sociocontextual level, results indicate that when youth experienced a standard deviation increase in the proportion of months not gainfully employed or in school, there is a 1.32 times greater likelihood of engaging in gun violence (SE = .12, p = .024, 95% CI [1.43, 1.69]). Additionally, during periods when adolescents experience a one standard deviation increase in their typical frequency of witnessing nongun violence, their likelihood of committing gun violence increases by 1.93 times (SE = .16, p < .001, 95% CI [1.43, 2.62]). It is also that case that when youth experience a standard deviation increase in being victimized through nongun violence, it related to an increase in the likelihood of engaging in gun violence by approximately 1.37 times (SE = .14, p < .030, 95% CI[1.04, 1.82]). Within-person changes in neighborhood disadvantage did not significantly predict the likelihood of gun violence. Similar findings emerge in the supplemental models that exclude measures of hostility and neighborhood disadvantage due to the higher level of missingness in these measures.

Table 4.

Fixed Effects Logistic Regression Predicting Gun Violence

Full model: Model 1a Hostility removed: Model 1b Hostility and ND removed: Model 1c
Measure OR 95% CI OR Measure OR 95% CI
Individual factors
Hostility 1.14 .90–1.44
Poor future orientation 1.06 .82–1.37 1.11 .91–1.36 1.15 .95–1.38
Alcohol use 1.00 .78–1.29 1.14 .92–1.41 1.09 .90–1.32
Personal rewards to crime 1.32* 1.03–1.70 1.31* 1.06–1.05 1.32** 1.09–1.60
Sociocontextual factors
Time in ungainful activity 1.32* 1.43–1.69 1.25* 1.01–1.54 1.20 .99–1.45
Neighborhood disadvantage (ND) 1.07 .83–1.37 1.05 .83–1.31
Witness non-gun violence 1.93*** 1.43–2.62 1.85*** 1.46–2.35 1.90*** 1.52–2.38
Victimization non-gun violence 1.37* 1.04–1.82 1.33* 1.05–1.70 1.26* 1.01–1.58
Gang involvement
Open in a new tab

Note. OR = odds ratio; CI = confidence interval.

*

p < .05.

**

p < .01.

***

p < .001.

Discussion

To date, gun violence research has been primarily informed by examining the precursors to gun access/gun carrying or inferred from research investigating violence more broadly. Additionally, prior research has largely relied on cross-sectional data that limits the capacity to adequately address longitudinal relationships between risk factors and gun violence. The current study takes a different approach. In addition to examining the interindividual predictors related to gun use, it also examines the effects of within-person change in sociocontextual and individual factors associated with gun violence within 6-month periods over the course of three years in a sample of actual gun users. A “micro examination” such as this has the potential to guide the focus of intervention efforts by uncovering points of change within-individuals that increase the likelihood that individuals engage in gun violence. Several messages emerge from this work.

The only consistent individual-level factor that was significantly associated with gun violence across both modeling strategies was the perceived personal rewards to crime. This suggests that the responsiveness of adolescents to thrills and the social rewards during this developmental period in their lives may be achieved through the use of firearms or associated with situations where gun violence occurs (e.g., Loughran et al., 2016). Of note in the pooled logistic regression model, individuals with higher levels of poor future orientation were significantly more likely to engage in of gun violence. This is supportive of the notion that individuals who are more apt to prioritize the present and discount the future are less likely to consider the serious consequences of gun violence and encounter opportunities where violence is likely to occur (e.g., Hepburn & Hemenway, 2004; Piquero et al., 2005; Siegel et al., 2013) The lack of other consistent findings regarding individual-level risk factors within the fixed-effects models may also be related to a broader sample selection issue. The selection criteria used to generate the sample may have produced substantially reduced variation among individual-level characteristics across time. The differences in the findings across the cross-sectional logistic regression and fixed-effects regression models (i.e., future orientation) reinforce the fact that selection effects may be driving some of the observed differences. As seen in Table 2, the between-person variance in hostility and future orientation is slightly larger than the within-person variance in these measures. To the extent that these factors remain relatively stable within individuals captured in our sample could help explain the differences in the findings across the modeling specifications.

Across both sets of analyses and particularly within the fixed-effects estimation strategy, several sociocontextual were significantly associated with a shooting incident. Specifically, a greater proportion of months not gainfully employed or in school, and the frequency of victimization and witnessing of nongun violence were significantly related to the likelihood of engaging in gun violence. These findings reinforce the notion that the environments and situations where adolescents reside or spend time powerfully affect the likelihood that they engage in gun violence, even among adolescents at high risk for involvement in this behavior in the first place. Most consistently, witnessing nongun violence is significantly related to an increase in the likelihood of shooting a gun and was observed to have the largest relative effect in each of the modeling strategies. The relative magnitude of this relationship further suggests that exposure to neighborhoods where violence is pervasively witnessed by adolescents may lead those adolescents to engage in gun violence out of a need for protection (e.g., Esselmont, 2014; Kleck, 1988), fear of victimization (e.g., Webster, Gainer, & Champion, 1993), as a way to maintain status or respect (i.e., Anderson, 1999; Wright & Rossi, 1986), or even in response to psychological distress (e.g., Hodges & Scalora, 2015).

These results indicate that higher levels of exposure to nongun violence both across individuals and within-persons are both significantly related to a higher likelihood of gun violence. Thus, addressing differences in exposure to violence in high risk groups in general and within the life experience of specific, targeted adolescents could both have an impact on reducing involvement in gun violence. Importantly, this study looked at exposure to nongun violence specifically, suggesting that it is not necessarily the case that adolescent exposure to gun-specific violence begets the commission of further gun violence. Rather, it implies that exposure to violence broadly precipitates involvement in gun violence; a finding consistent with recent research regarding the precursors of gun carrying (Beardslee et al., 2018; Reid et al., 2017).

The emergence of differential effects of exposure to violence reinforces the need to further understand the mechanisms mediating an exposure experience, carrying a gun, and using one. As prior studies have indicated, some individuals report that obtaining access to a weapon or using a weapon is a protective measure (e.g., Cao et al., 1997; Lizotte et al., 2000; Spano & Bolland, 2013) or even as a way to maintain status or respect (i.e., Anderson, 1999; Wright & Rossi, 1986). The current study emphasizes that certain adolescents move beyond just having the gun for defensive assurance, since the chances of engaging in gun violence in a short time period after an increase in witnessing violence is also heightened. Research should continue to explore the motivations for gun-use/carrying and how they translate into harmful behavior.

There are also implications for possible intervention approaches. It is well-known that there is a “cycle of violence” characterized by exposure to violent victimization or behavior followed by an increased likelihood of engaging in violent behavior (e.g., Dodge, Bates, & Pettit, 1990; Fagan, 2005; Widom, 1989). The current study continues to illustrate this cycle; it does so in a particularly extreme—and potentially deadly—pathway toward involvement in gun violence. The potential trauma and psychological realignment possibly accompanying high or increased exposure to nongun violence certainly appears to be worthy of considerable attention in efforts to reduce the prevalence of serious violence within communities (and by default to synergistically reduce opportunities to witness or be a victim of violence; e.g., Molnar et al., 2004) as well as in interventions to reduce the chances of specific adolescents from resorting to gun violence (e.g., Stein et al., 2003). The fact that exposure to violence is related to both gun carrying and gun violence may reflect a continuum in the development of adolescent violence (i.e., Dodge, 2001). As a result, public health and law enforcement efforts to reduce violence in communities might gain considerable traction by addressing the decision-making processes surrounding adolescent involvement with guns.

The findings from this study also echo the fact that involvement with guns is only one part of an adolescent’s overall involvement with violent offending and victimization (e.g., Spano & Bolland, 2011; Spano et al., 2012). There is clearly a need for investigations into how gun violence is both a byproduct and generator of violent behavior. Ultimately, the current study provides an empirical inquiry into the proximal risks for gun violence by examining these dynamics among a high-risk sample, heeding calls from national organizations to invest in empirical research that can inform policy and intervention efforts at both the individual and community level (e.g., American Psychological Association, 2013; Leshner et al., 2013).

Limitations and Future Research

As with any study, there are several notable limitations. First, involvement in gun violence is relatively rare, even in the Pathways sample of high-risk adolescent offenders. This reality led to the inclusion of a highly selective sample and small sample sizes, limiting our capacity to conduct certain analyses (e.g., assessing effects within youth with and without a mental health problem) and investigation of nuances related to the overall findings (e.g., whether limited variation drives the lack of explanatory power among the individual level factors). Additionally, the measure of gun violence is self-reported and there are obvious reasons why individuals may fail to disclose their involvement in such serious criminal behavior. The challenge of balancing the limitations of official records and self-report remains present in most criminological work, however self-reported offending has been shown to be a valuable and reliable tool to capture offending behavior (e.g., Thornberry & Krohn, 2000). Ultimately, this study advances our understanding of gun violence by elucidating the process of gun use in an extremely relevant sample.

The Pathways to Desistance Study does not include event-specific data to explore the situational characteristics associated with incidents involving gun violence. Although both individual and sociocontextual risks are important, more closely related situational factors have also been associated with increases in firearm violence (e.g., Connor, Duberstein, Conwell, & Caine, 2003). For example, gun violence may emerge because of perceived social threats or disrespect that occurs during exchanges between people (e.g., Anderson, 1999; Connor et al., 2003). Involvement in these types of interactions may partially explain the relationship between exposure to nongun violence and the commission of gun violence and hold considerable potential for informing intervention efforts. Data from The Pathways to Desistance Study was collected over a decade ago and may reflect a unique experience of gun violence associated with its own political, cultural, and social climate. Nonetheless, it is unlikely that the nature of offending (and offenders) has changed drastically, such that studying gun violence among this sample does not provide needed empirical and policy relevant findings. Additionally, there may be other risk or protective factors that were not explored in the current study that could be important in explaining between and within-person differences in gun violence that may warrant consideration in future work.

Although political and cultural tensions surrounding gun control and policy remain, there is a need to continue to develop a research agenda that examines characteristics of gun violence, risk and protective factors for involvement in gun violence, and the potential of intervention strategies. For example, future work could evaluate how individual and sociocontextual risk factors also differentiate types of gun violence (e.g., fatal vs. nonfatal, impulsive, targeted or predatory), as the weight of certain risk factors may vary across the type of gun violence examined. Research could also consider tracing how risk and protective factors are related violence in general as well as to the acquisition of, carrying, and use of guns in order to comprehensively understand the process by which individuals integrate firearms into their behavioral repertoire. Lastly, there is need to continue investigating the factors and processes among samples at high risk for involvement in gun violence, like those in the current study. The power and efficiency of future interventions depends on whether they address the issues salient for those individuals most at risk for generating and experiencing harm from using guns.

Public Significance Statement.

An evaluation of individual and sociocontextual risk factors specific to gun violence among a policy relevant high-risk sample provides key insights into understanding gun violence. The most robust finding implicates the cycle of violence by showing that exposure to nongun violence contributes to perpetration of gun violence. Addressing differences in exposure to violence in high risk groups in general and within the life experience of specific, targeted adolescents could both have an impact on reducing involvement in gun violence.

Acknowledgments

The project described was developed under Grant R01 HD086761-01 with funding from the National Institute on Child Health and Development (NICHD). The contents of this article do not necessarily represent the policy of NICHD and you should not assume endorsement by the Federal Government.

Appendix. Correlation Matrix for Study Period

Variable 1 2 3 4 5 6 7 8 9
Wave 1
1. Gun violence —
2. Hostility .12 —
3. Personal rewards .24* .14 —
4. Poor future orientation .23* .20* .38*** —
5. Alcohol use .24* .21* .39*** .41*** —
6. Time in ungainful activity .05 .08 −.06 .04 −.21* —
7. Gang involvement .30** .25* .35** .16 .19 −.15 —
8. Witness NG-violence .36*** .20* .20* .08 .26** .02 .17 —
9. Victim NG-violence .33*** .31** .31** .14 .34*** −.16 .25** .52*** —
10. Neighborhood disadvantage −.06 −.27* −.19 −.35*** −.13 .21 .01 −.07 −.14
Wave 2
1. Gun violence —
2. Hostility .19 —
3. Personal rewards .16 .06 —
4. Poor future orientation .25** .17 .33*** —
5. Alcohol use .20* .19 .35*** .10 —
6. Time in ungainful activity .23* .10 −.08 −.04 −.10 —
7. Gang involvement −.05 .03 .30** .13 .24* −.08 —
8. Witness NG-violence .39*** .47*** .11 .07 .33*** .11 −.01 —
9. Victim NG-violence .25** .14 .03 .09 .17 −.03 .10 .43*** —
10. Neighborhood disadvantage .24* .08 −.24* −.14 −.01 .32** −.11 .17 .00
Wave 3
1. Gun violence —
2. Hostility .06 —
3. Personal rewards .03 .20 —
4. Poor future orientation .15 .01 .23* —
5. Alcohol use −.03 .13 .13 .09 —
6. Time in ungainful activity .19* .10 .01 .05 .04 —
7. Gang involvement .02 .10 .15 .07 .24* −.02 —
8. Witness NG-violence .23* .20 .00 .08 .05 .11 .04 —
9. Victim NG-violence .10 .09 .04 −.10 .06 .12 .10 .21* —
10. Neighborhood disadvantage .11 −.22* −.18 −.14 −.06 .14 −.10 .08 .09
Wave 4
1. Gun violence —
2. Hostility .00 —
3. Personal rewards .15 .22 —
4. Poor future orientation .16 .07 .23* —
5. Alcohol use .14 .12 .30*** .01 —
6. Time in ungainful activity .13 .05 −.06 .12 −.07 —
7. Gang involvement .09 .23* .38*** .22* .29** .00 —
8. Witness NG-violence .39*** .39*** .11 −.07 .14 .11 .10 —
9. Victim NG-violence .10 .25* .18 .15 .19* .20* .32*** .29** —
10. Neighborhood disadvantage .08 −.13 −.16 .03 −.13 .10 −.07 .11 −.04
Wave 5
1. Gun violence —
2. Hostility .16 —
3. Personal rewards .15 .09 —
4. Poor future orientation −.09 −.06 .00 —
5. Alcohol use .05 .13 .22* .03 —
6. Time in ungainful activity .12 .08 −.16 −.03 .00 —
7. Gang involvement .17 .08 .41*** −.18 .19* −.04 —
8. Witness NG-violence .41*** .24* .11 −.26** .26** −.06 .14 —
9. Victim NG-violence .29** .42*** .04 −.22* .15 .01 .20 —
10. Neighborhood disadvantage .04 −.03 −.20 .01 .05 .20 .04 .02 −.05
Wave 6
1. Gun violence —
2. Hostility .10 —
3. Personal rewards .22* .23* —
4. Poor future orientation .12 −.05 .01 —
5. Alcohol use .26** .14 .36*** .09 —
6. Time in ungainful activity .21* −.04 −.16 .27** −.03 —
7. Gang involvement .18 −.01 .20* .02 .36*** .14 —
8. Witness NG-violence .52*** .23* .21* −.06 .23* .05 .24* —
9. Victim NG-violence .46*** .25* .44*** .02 .21* .03 .19 .54*** —
10. Neighborhood disadvantage .04 −.31*** −.15 .11 −.38*** .33** −.11 .03 .21
Open in a new tab
*

p < .05.

**

p < .01.

***

p < .001.

Contributor Information

Zachary R. Rowan, Simon Fraser University

Carol A. Schubert, University of Pittsburgh

Thomas A. Loughran, The Pennsylvania State University

Edward P. Mulvey, University of Pittsburgh

Dustin A. Pardini, Arizona State University

References

  1. Allison PD (2009). Fixed effects regression models. Thousand Oaks, CA: SAGE. 10.4135/9781412993869 [DOI] [Google Scholar]
  2. American Psychological Association. (2013). Gun violence: Prediction, prevention, and policy. Retrieved from http://www.apa.org/pubs/info/reports/gun-violence-prevention.aspx
  3. Anderson E (1999). Code of the Street: Decency, violence, and the moral life of the inner city. New York, NY: Norton. [Google Scholar]
  4. Baskin D, & Sommers I (2014). Exposure to community violence and trajectories of violent offending. Youth Violence and Juvenile Justice, 12, 367–385. 10.1177/1541204013506920 [DOI] [Google Scholar]
  5. Beardslee J, Docherty M, Mulvey E, Schubert C, & Pardini D (2018). Childhood risk factors associated with adolescent gun carrying among Black and White males: An examination of self-protection, social influence, and antisocial propensity explanations. Law and Human Behavior, 42, 110–118. 10.1037/lhb0000270 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bingenheimer JB, Brennan RT, & Earls FJ (2005). Firearm violence exposure and serious violent behavior. Science, 308, 1323–1326. 10.1126/science.1110096 [DOI] [PubMed] [Google Scholar]
  7. Braga AA, Papachristos AV, & Hureau DM (2010). The concentration and stability of gun violence at micro places in Boston, 1980–2008. Journal of Quantitative Criminology, 26, 33–53. 10.1007/s10940-009-9082-x [DOI] [Google Scholar]
  8. Cao L, Cullen FT, & Link BG (1997). The social determinants of gun ownership: Self-protection in an urban environment. Criminology, 35, 629–658. 10.1111/j.1745-9125.1997.tb01233.x [DOI] [Google Scholar]
  9. Cardwell SM, & Piquero AR (2018). Does violence in adolescence differentially predict offending patterns in early adulthood? International Journal of Offender Therapy and Comparative Criminology, 62, 1603–1628. 10.1177/0306624X16688978 [DOI] [PubMed] [Google Scholar]
  10. Cauffman E, & Woolard J (1999). The Future Outlook Inventory. (Unpublished manuscript). MacArthur Network on Adolescent Development and Juvenile Justice. [Google Scholar]
  11. Centers for Disease Control and Prevention (CDC). (2017a). High School Youth Risk Behavior Survey. Retrieved from http://nccd.cdc.gov/youthonline/
  12. Centers for Disease Control and Prevention (CDC). (2017b). Web-based injury statistics query and reporting system. https://www.cdc.gov/injury/wisqars
  13. Chassin L, Knight G, Vargas-Chanes D, Losoya SH, & Naranjo D (2009). Substance use treatment outcomes in a sample of male serious juvenile offenders. Journal of Substance Abuse Treatment, 36, 183–194. 10.1016/j.jsat.2008.06.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Connor KR, Duberstein PR, Conwell Y, & Caine ED (2003). Reactive aggression and suicide—Theory and evidence. Aggression and Violent Behavior, 8, 413–432. 10.1016/S1359-1789(02)00067-8 [DOI] [Google Scholar]
  15. DeLisi M, Piquero AR, & Cardwell SM (2016). The unpredictability of murder: Juvenile homicide in the Pathways to Desistance Study. Youth Violence and Juvenile Justice, 14, 26–42. 10.1177/1541204014551805 [DOI] [Google Scholar]
  16. Derogatis LR, & Melisaratos N (1983). The Brief Symptom Inventory: An introductory report. Psychological Medicine, 13, 595–605. 10.1017/S0033291700048017 [DOI] [PubMed] [Google Scholar]
  17. Dodge KA (2001). The science of youth violence prevention. Progressing from developmental epidemiology to efficacy to effectiveness to public policy. American Journal of Preventive Medicine, 20(Suppl.), 63–70. 10.1016/S0749-3797(00)00275-0 [DOI] [PubMed] [Google Scholar]
  18. Dodge KA, Bates JE, & Pettit GS (1990). Mechanisms in the cycle of violence. Science, 250, 1678–1683. 10.1126/science.2270481 [DOI] [PubMed] [Google Scholar]
  19. Dodge KA, & Frame CL (1982). Social cognitive biases and deficits in aggressive boys. Child Development, 53, 620–635. 10.2307/1129373 [DOI] [PubMed] [Google Scholar]
  20. Dodge KA, Greenberg MT, & Malone PS, & the Conduct Problems Prevention Research Group. (2008). Testing an idealized dynamic cascade model of the development of serious violence in adolescence. Child Development, 79, 1907–1927. 10.1111/j.1467-8624.2008.01233.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Elbogen EB, & Johnson SC (2009). The intricate link between violence and mental disorder: Results from the National Epidemiologic Survey on Alcohol and Related Conditions. Archives of General Psychiatry, 66, 152–161. 10.1001/archgenpsychiatry.2008.537 [DOI] [PubMed] [Google Scholar]
  22. Elliott DS, Wilson WJ, Huizinga D, Sampson RJ, Elliott A, & Rankin B (1996). The effects of neighborhood disadvantage on adolescent development. Journal of Research in Crime and Delinquency, 33, 389–426. 10.1177/0022427896033004002 [DOI] [Google Scholar]
  23. Esselmont C (2014). Carrying a weapon to school: The roles of bullying victimization and perceived safety. Deviant Behavior, 35, 215–232. 10.1080/01639625.2013.834767 [DOI] [Google Scholar]
  24. Fagan J (2005). The relationship between adolescent physical abuse and criminal offending: Support for an enduring and generalized cycle of violence. Journal of Family Violence, 20, 279–290. 10.1007/s10896-005-6604-7 [DOI] [Google Scholar]
  25. Fagan J, & Wilkinson DL (1998). Guns, youth violence, and social identity in inner cities. Crime and Justice, 24, 105–188. 10.1086/449279 [DOI] [Google Scholar]
  26. Farrell AD, & Flannery DJ (2006). Youth violence prevention: Are we there yet? Aggression and Violent Behavior, 11, 138–150. 10.1016/j.avb.2005.07.008 [DOI] [Google Scholar]
  27. Farrington DP, & Loeber R (2000). Epidemiology of juvenile violence. Child and Adolescent Psychiatric Clinics of North America, 9, 733–748. 10.1016/S1056-4993(18)30089-0 [DOI] [PubMed] [Google Scholar]
  28. Guerra NG, Huesmann LR, & Spindler A (2003). Community violence exposure, social cognition, and aggression among urban elementary school children. Child Development, 74, 1561–1576. 10.1111/1467-8624.00623 [DOI] [PubMed] [Google Scholar]
  29. Hawkins JD, Herrenkohl TI, Farrington DP, Brewer D, Catalano RF, Harachi TW, & Cothern L (2000). Predictors of youth violence. Washington, DC: Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention. 10.1037/e524202006-001 [DOI] [Google Scholar]
  30. Hayes DN, & Hemenway D (1999). Age-within-school-class and adolescent gun-carrying. Pediatrics, 103, e64. 10.1542/peds.103.5.e64 [DOI] [PubMed] [Google Scholar]
  31. Health Innovation Program. (2000). Area deprivation index. Retrieved from http://www.hipxchange.org/ADI
  32. Hemenway D, Prothrow-Stith D, Bergstein JM, Ander R, & Kennedy BP (1996). Gun carrying among adolescents. Law and Contemporary Problems, 59, 39–53. 10.2307/1192209 [DOI] [Google Scholar]
  33. Hepburn LM, & Hemenway D (2004). Firearm availability and homicide: A review of the literature. Aggression and Violent Behavior, 9, 417–440. 10.1016/S1359-1789(03)00044-2 [DOI] [Google Scholar]
  34. Hodges HJ, & Scalora MJ (2015). Challenging the political assumption that “Guns don’t kill people, crazy people kill people!”. American Journal of Orthopsychiatry, 85, 211–216. 10.1037/ort0000069 [DOI] [PubMed] [Google Scholar]
  35. Hohl BC, Wiley S, Wiebe DJ, Culyba AJ, Drake R, & Branas CC (2017). Association of drug and alcohol use with adolescent firearm homicide at individual, family, and neighborhood levels. Journal of the American Medical Association Internal Medicine, 177, 317–324. 10.1001/jamainternmed.2016.8180 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Huesmann LR, Eron LD, Lefkowitz MM, & Walder LO (1984). Stability of aggression over time and generations. Developmental Psychology, 20, 1120–1134. 10.1037/0012-1649.20.6.1120 [DOI] [Google Scholar]
  37. Huizinga D, Esbensen F, & Weiher A (1991). Are there multiple paths to delinquency? The Journal of Criminal Law & Criminology, 82, 83–118. 10.2307/1143790 [DOI] [Google Scholar]
  38. Kaplan S (2018, March 12). Congress quashed research into gun violence. Since then, 600,000 people have been shot. The New York Times. Retrieved from https://www.nytimes.com/2018/03/12/health/gunviolence-research-cdc.html?authlogin-email&loginsmartlock&authlogin-smartlock [Google Scholar]
  39. Katz J (1988). Seductions of crime: Moral and sensual attractions in doing evil. New York, NY: Basic Books. [Google Scholar]
  40. Kleck G (1988). Crime control through the private use of armed force. Social Problems, 35, 1–21. 10.2307/800663 [DOI] [Google Scholar]
  41. Laub JH, & Sampson RJ (2003). Shared beginnings, divergent lives: Delinquent boys to age 70. Cambridge, MA: Harvard University Press. [Google Scholar]
  42. Leshner AI, Altevogt BM, Lee AF, McCoy MA (Eds.). (2013). Priorities for research to reduce the threat of firearm-related violence. Washington, DC: The National Academies Press. [Google Scholar]
  43. Lizotte AJ, Howard GJ, Krohn MD, & Thornberry TP (1997). Patterns of illegal gun carrying among young urban males. Valparaiso University Law Review, 31, 375–393. Retrieved from https://scholar.valpo.edu/vulr/vol31/iss2/4 [Google Scholar]
  44. Lizotte AJ, Krohn DM, Howell JC, Tobin K, & Howard GJ (2000). Factors influencing gun carrying among young urban males over the adolescent-young adult life course. Criminology, 38, 811–834. 10.1111/j.1745-9125.2000.tb00907.x [DOI] [Google Scholar]
  45. Loeber R, Pardini D, Homish DL, Wei EH, Crawford AM, Farrington DP, … Rosenfeld R (2005). The prediction of violence and homicide in young men. Journal of Consulting and Clinical Psychology, 73, 1074–1088. 10.1037/0022-006X.73.6.1074 [DOI] [PubMed] [Google Scholar]
  46. Loh K, Walton MA, Harrison SR, Zimmerman M, Stanley R, Chermack ST, & Cunningham RM (2010). Prevalence and correlates of handgun access among adolescents seeking care in an urban emergency department. Accident; Analysis and Prevention, 42, 347–353. 10.1016/j.aap.2009.11.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Loughran TA, Nguyen H, Piquero AR, & Fagan J (2013). The returns to criminal capital. American Sociological Review, 78, 925–948. 10.1177/0003122413505588 [DOI] [Google Scholar]
  48. Loughran TA, Reid JA, Collins ME, & Mulvey EP (2016). Effect of gun carrying on perceptions of risk among adolescent offenders. American Journal of Public Health, 106, 350–352. 10.2105/AJPH.2015.302971 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Maimon D, & Browning CR (2010). Unstructured socializing, collective efficacy, and violent behavior among ubran youth. Crimniology, 48, 443–474. 10.1111/j.1745-9125.2010.00192.x [DOI] [Google Scholar]
  50. Maruna S (2000). Making good: How ex-convicts reform and rebuild their lives. Washington, DC: American Psychological Association. [Google Scholar]
  51. McGinty EE, Webster DW, Jarlenski M, & Barry CL (2014). News media framing of serious mental illness and gun violence in the United States, 1997–2012. American Journal of Public Health, 104, 406–413. 10.2105/AJPH.2013.301557 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Mesters G, van der Geest V, & Bijleveld C (2016). Crime, employment, and social welfare: An individual-level study on disadvantaged males. Journal of Quantitative Criminology, 32, 159–190. 10.1007/s10940-015-9258-5 [DOI] [Google Scholar]
  53. Miller M, Hemenway D, & Wechsler H (2002). Guns and gun threats at college. Journal of American College Health, 51, 57–65. 10.1080/07448480209596331 [DOI] [PubMed] [Google Scholar]
  54. Molnar BE, Miller MJ, Azrael D, & Buka SL (2004). Neighborhood predictors of concealed firearm carrying among children and adolescents: Results from the project on human development in Chicago neighborhoods. Archives of Pediatrics & Adolescent Medicine, 158, 657–664. 10.1001/archpedi.158.7.657 [DOI] [PubMed] [Google Scholar]
  55. Monahan J (1992). Mental disorder and violent behavior. Perceptions and evidence. American Psychologist, 47, 511–521. 10.1037/0003-066X.47.4.511 [DOI] [PubMed] [Google Scholar]
  56. Mulvey EP, Steinberg L, Fagan J, Cauffman E, Piquero AR, Chassin L, … Losoya SH (2004). Theory and research on desistance from antisocial activity among serious adolescent offenders. Youth Violence and Juvenile Justice, 2, 213–236. 10.1177/1541204004265864 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Nagin DS, & Paternoster R (1994). Personal capital and social control: The deterrence implications of individual differences in criminal offending. Criminology, 32, 581–606. 10.1111/j.1745-9125.1994.tb01166.x [DOI] [Google Scholar]
  58. Nagin D, & Tremblay RE (1999). Trajectories of boys’ physical aggression, opposition, and hyperactivity on the path to physically violent and nonviolent juvenile delinquency. Child Development, 70, 1181–1196. 10.1111/1467-8624.00086 [DOI] [PubMed] [Google Scholar]
  59. Nguyen H, & Bouchard M (2013). Need, connections, or competence? Criminal achievement among adolescent offenders. Justice Quarterly, 30, 44–83. 10.1080/07418825.2011.589398 [DOI] [Google Scholar]
  60. Osgood DW, & Anderson AL (2001). Unstructured socializing and rates of delinquency. Criminology, 42, 519–550. 10.1111/j.1745-9125.2004.tb00528.x [DOI] [Google Scholar]
  61. Osgood DW, Wilson JK, O’Malley PM, Bachman JG, & Johnston LD (1996). Routine activities and individual deviant behavior. American Sociological Review, 61, 635–655. 10.2307/2096397 [DOI] [Google Scholar]
  62. Piquero AR, Blumstein A, Brame R, Haapanen R, Mulvey EP, & Nagin DS (2001). Assessing the impact of exposure time and incapacitation on longitudinal trajectories of criminal offending. Journal of Adolescent Research, 16, 54–74. 10.1177/0743558401161005 [DOI] [Google Scholar]
  63. Piquero AR, MacDonald J, Dobrin A, Daigle LE, & Cullen FT (2005). Self-control, violent offending, and homicide victimization: Assessing the general theory of crime. Journal of Quantitative Criminology, 21, 55–71. 10.1007/s10940-004-1787-2 [DOI] [Google Scholar]
  64. Planty M, & Truman JL (2013). Firearm violence, 1993–2011. Washington, DC: Bureau of Justice Statistics, Office of Justice Programs, U. S. Department of Justice. [Google Scholar]
  65. Reid JA, Richards TN, Loughran TA, & Mulvey EP (2017). The relationships among exposure to violence, psychological distress, and gun carrying among male adolescents found guilty of serious legal offenses: A longitudinal cohort study. Annals of Internal Medicine, 166, 412–418. 10.7326/M16-1648 [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Resnick MD, Ireland M, & Borowsky I (2004). Youth violence perpetration: What protects? What predicts? Findings from the National Longitudinal Study of Adolescent Health. Journal of Adolescent Health, 35, 424.e1–424.e10. 10.1016/j.jadohealth.2004.01.011 [DOI] [PubMed] [Google Scholar]
  67. Sampson RJ, & Laub JH (1993). Crime in the making: Pathways and turning points through life. Cambridge, MA: Harvard University Press. 10.1177/0011128793039003010 [DOI] [Google Scholar]
  68. Sampson RJ, Raudenbush SW, & Earls F (1997). Neighborhoods and violent crime: A multilevel study of collective efficacy. Science, 277, 918–924. 10.1126/science.277.5328.918 [DOI] [PubMed] [Google Scholar]
  69. Savolainen J (2009). Work, family, and criminal desistance: Adult social bonds in a Nordic welfare state. British Journal of Criminology, 49, 285–304. 10.1093/bjc/azn084 [DOI] [Google Scholar]
  70. Schubert CA, & Mulvey EP (2014). Aftercare services are key to positive community adjustment. Chicago, IL: MacArthur Foundation. [Google Scholar]
  71. Schubert CA, Mulvey EP, Steinberg L, Cauffman E, Losoya SH, Hecker T, … Knight GP (2004). Operational lessons from the Pathways to Desistance Project. Youth Violence and Juvenile Justice, 2, 237–255. 10.1177/1541204004265875 [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Selner-O’Hagan MB, Kindlon DJ, Buka SL, Raudenbush SW, & Earls FJ (1998). Assessing exposure to violence in urban youth. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 39, 215–224. 10.1017/S002196309700187X [DOI] [PubMed] [Google Scholar]
  73. Sher KJ (1987). Questionnaire for the Missouri Health and Behavior Study (Unpublished instrument). Columbia, MO: University of Missouri-Columbia. [Google Scholar]
  74. Siegel M, Ross CS, & King C III. (2013). The relationship between gun ownership and firearm homicide rates in the United States, 1981–2010. American Journal of Public Health, 103, 2098–2105. 10.2105/AJPH.2013.301409 [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. Spano R, & Bolland J (2011). Is the nexus of gang membership, exposure to violence, and violent behavior a key determinant of first time gun carrying for urban minority youth? Justice Quarterly, 28, 838–862. 10.1080/07418825.2010.547868 [DOI] [Google Scholar]
  76. Spano R, & Bolland J (2013). Disentangling the effects of violent victimization, violent behavior, and gun carrying for minority inner-city youth living in extreme poverty. Crime and Delinquency, 59, 191–213. 10.1177/0011128710372196 [DOI] [Google Scholar]
  77. Spano R, Pridemore WA, & Bolland J (2012). Specifying the role of exposure to violence and violent behavior on initiation of gun carrying: A longitudinal test of three models of youth gun carrying. Journal of Interpersonal Violence, 27, 158–176. 10.1177/0886260511416471 [DOI] [PubMed] [Google Scholar]
  78. Steadman HJ, Monahan J, Pinals DA, Vesselinov R, & Robbins PC (2015). Gun violence and victimization of strangers by persons with a mental illness: Data from the MacArthur Violence Risk Assessment Study. Psychiatric Services, 66, 1238–1241. 10.1176/appi.ps.201400512 [DOI] [PubMed] [Google Scholar]
  79. Stein BD, Jaycox LH, Kataoka SH, Wong M, Tu W, Elliott MN, & Fink A (2003). A mental health intervention for schoolchildren exposed to violence: A randomized controlled trial. Journal of the American Medical Association, 290, 603–611. 10.1001/jama.290.5.603 [DOI] [PubMed] [Google Scholar]
  80. Swanson JW, McGinty EE, Fazel S, & Mays VM (2015). Mental illness and reduction of gun violence and suicide: Bringing epidemiologic research to policy. Annals of Epidemiology, 25, 366–376. 10.1016/j.annepidem.2014.03.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  81. Thornberry TP, Huizinga D, & Loeber R (1995). Prevention of serious delinquency and violence: Implications from the Program of Research on the Causes and Correlates of Delinquency. In Howell JC, Krisberg B, Hawkins JD, & Wilson JJ (Eds.), Sourcebook on serious, violent, and chronic juvenile offenders (pp. 213–237). Thousand Oaks, CA: SAGE. [Google Scholar]
  82. Thornberry TP, & Krohn MD (2000). The self-report method for measuring delinquency and crime. In Duffee D, Crutchfield RD, Mastrofski S, Mazerolle L, & McDowall D (Eds.), Criminal justice 2000, Vol. 4: Innovations in measurement and analysis (pp. 33–83). Washington, DC: National Institute of Justice, Office of Justice Programs, U.S. Department of Justice. [Google Scholar]
  83. Thornberry TP, Lizotte AJ, Krohn MD, Farnworth M, & Jang SJ (1994). Delinquent peers, beliefs, and delinquent behavior: A longitudinal test of interactional theory. Criminology, 32, 47–83. 10.1111/j.1745-9125.1994.tb01146.x [DOI] [Google Scholar]
  84. Uggen C (2000). Work as a turning point in the life course of criminals: A duration model of age, employment, and recidivism. American Sociological Review, 65, 529–546. 10.2307/2657381 [DOI] [Google Scholar]
  85. Van Dorn R, Volavka J, & Johnson N (2012). Mental disorder and violence: Is there a relationship beyond substance use? Social Psychiatry and Psychiatric Epidemiology, 47, 487–503. 10.1007/s00127-011-0356-x [DOI] [PubMed] [Google Scholar]
  86. Webster DW, Gainer PS, & Champion HR (1993). Weapon carrying among inner-city junior high school students: Defensive behavior vs aggressive delinquency. American Journal of Public Health, 83, 1604–1608. 10.2105/AJPH.83.11.1604 [DOI] [PMC free article] [PubMed] [Google Scholar]
  87. Weinberger DA, & Schwartz GE (1990). Distress and restraint as superordinate dimensions of self-reported adjustment: A typological perspective. Journal of Personality, 58, 381–417. 10.1111/j.1467-6494.1990.tb00235.x [DOI] [PubMed] [Google Scholar]
  88. Weisburd DL (2015). The law of crime concentration and the criminology of place. Criminology, 53, 133–157. 10.1111/1745-9125.12070 [DOI] [Google Scholar]
  89. Widom CS (1989). The cycle of violence. Science, 244, 160–166. 10.1126/science.2704995 [DOI] [PubMed] [Google Scholar]
  90. Wintemute GJ (2015). The epidemiology of firearm violence in the twenty-first century United States. Annual Review of Public Health, 36, 5–19. 10.1146/annurev-publhealth-031914-122535 [DOI] [PubMed] [Google Scholar]
  91. Wright JD, & Rossi PH (1986). The armed criminal in America. Washington, DC: U. S. Department of Justice. [Google Scholar]

RESOURCES