Direct and Indirect Exposure to Violence as Predictors of Vaping: Examining Social and Cognitive Mediators

Abstract

Exposure to violence has been identified as a risk factor for vaping. However, there is a dearth of research which has delineated direct and indirect exposures to violence as predictors. Similarly, there is limited work which has examined mediators of these potential relationships. The present study sought to address these gaps in the literature by examining deviant peer association and impulse control as mediators of the relationships between the various forms of exposure to violence and vaping. The Adolescent Brain Cognitive Development study data were analyzed. Generalized structural equation modeling was used to test direct and indirect relationships of interest. Findings indicated that only indirect exposure to violence significantly predicted increased risk for vaping. Deviant peer association significantly mediated this relationship, but impulse control did not. Implications are discussed.

Keywords

vaping exposure to violence impulse control deviant peer association mediation

Introduction

Exposure to violence has been identified as a robust risk factor for a range of substance use behaviors (Henry, 2020; Lawrence et al., 2023; Saadatmand et al., 2022), including vaping (Lee & Kim, 2021; Wu et al., 2023). Vaping is a relatively new substance use behavior wherein tobacco, nicotine, or other substances are vaporized and inhaled instead of going through the traditional process of combustion. While vaping has been advertised as a safer alternative to combustible tobacco products, research has indicated that there still exists a myriad of health risks associated with the behavior; like premature death, lung issues, and poor ocular health, among others (Levy et al., 2021; Martheswaran et al., 2021; Nicholson et al., 2021). Increased prevalence of vaping over the last decade has led to concerns about long-term population health impacts in this regard. The increase in popularity paired with these health risks highlight the need to identify predictors of vaping, particularly among youth given that early onset may trigger elevated risk for continued and chronic use issues (Bold et al., 2017; Sun & Yi, 2024). One risk factor for vaping that has been identified by prior research is exposure to violence (Lee & Kim, 2021; Wu et al., 2023). That said, there remain several important limitations of prior research in this area.

First, there is a dearth of research which has examined differences in predictive effects of direct and indirect exposures to violence for understanding risk for vaping. This is problematic, as past research has indicated that both forms of this stressor may be differentially related to substance use outcomes (Beharie et al., 2019; Pinchevsky et al., 2014). Identification of differences in predictive power of these forms of exposure to violence may entail more effective design and implementation of treatment programming for survivors of these various experiences. Second, there is similarly a lack of research that has sought to identify mechanisms underlying the relationship between exposure to violence and vaping. This, again, presents issues because understanding why exposure to violence increases vaping risk would allow for the tailoring of interventions that address the specific mechanism through which exposure to violence increases vaping risk. This is particularly relevant here because such mechanisms may only explain the relationship between direct or indirect exposure to violence and vaping, thus, creating the potential for more targeted treatment that may have increased effectiveness and may be more cost efficient. Deviant peer association and impulse control may be one set of mechanisms that explains the relationship between exposure to violence and vaping, as prior research has indicated that both affiliation with deviant peers and low impulse control are risk factors for vaping and experiencing exposure to violence increases risk for deviant peer association and diminishes impulse control (Davis et al., 2022; Dixon et al., 2024; Lanza et al., 2020; Rudolph et al., 2014). The present study sought to address these gaps in the extant literature by examining direct and indirect exposures to violence as predictors of vaping and examined deviant peer association and impulse control as mediators of these relationships among a sample of adolescents. This study focused specifically on the use of tobacco/nicotine devices that function to vaporize substance without combustion as the outcome of interest.

Direct versus Indirect Exposure to Violence as Predictors of Vaping

Exposure to violence may be delineated into different categories based on whether it was directly experienced or indirectly experienced. Direct exposure to violence is defined as experiences where an individual directly experiences the exposure themselves through victimization (e.g., beaten up by someone else or a group, attacked with a weapon, etc.), whereas indirect exposure to violence is defined as an experience where an individual witnesses violence occur to someone else or becomes knowledgeable of a violent act happening to a loved one (e.g., seeing parents physically fight, watching someone get beaten, etc). Both forms of exposure to violence may trigger a stress response in adolescents and have both been found to be related to a variety of substance use outcomes (Gilreath et al., 2014; Wright et al., 2013; Zinzow et al., 2009). For example, Pinchevsky et al. (2014) observed that indirect exposure to violence produced a stronger magnitude effect on substance use outcomes in certain cases, but that experiencing both forms of exposure to violence generally produced the greater impacted on substance use risk among a sample of adolescents. Again, research tends to treat exposure to violence like this as stress regardless of whether that exposure is directly experiences as victimization or indirectly experienced through vicarious exposures like witnessing a loved one be physically harmed. General strain theory provides perhaps the most prominent criminological theory focused on stress stimuli as predictors of criminal behavior (Agnew, 1992). This theory posits that youth experience high magnitude strain, like exposure to violence, and use delinquent behavior, like vaping, as a coping response to mitigate the effects of strain. Following from this perspective, one may assume that directly experiencing exposure to violence should produce the greatest impact on outcomes of interest because it would make sense that directly experiencing victimization would trigger a stronger stress response and necessitate coping to a greater extent. However, the results of the highlighted research indicate the need for further interrogation of these relationships. It may indeed be that stress responses are triggered to a similar degree whether or not those experiences are directly or indirectly experienced. That said, this also raises the question of whether indirect victimization may operate through alternative mechanisms to predict vaping. While general strain theory focuses on negative affect as a mediator of this relationship (Agnew, 1992), there may be a myriad of other impacts that exposure to violence may have on individuals that may provide additional mechanisms liking these distinct forms of strain to vaping risk. Not only does there exist a dearth of research which has examined vaping as an outcome in this regard, but there similarly has been a lack of research that has examined alternative mediating pathways explaining the relationships between the various forms of exposure to violence and vaping. This study sought to explore the potential that peer selection processes may play a role in this regard.

Mediation of the Relationship Between Exposure to Violence Subtypes and Vaping

As noted above, the traditional theoretical perspective explaining the relationship between exposure to violence and substance use has been via stress processes, with general strain theory playing a prominent role in this regard (Agnew, 1992). That said, other research on stress has examined the impact of exposure to violence on non-affective mechanisms linking stressors to behavioral outcomes. This is understandable given that traumatic stressors like exposure to violence has been found to impact a range of neurological outcomes that may exhibit chronic effects in this regard (McEwen, 1998; Stam, 2007). Impulse control refers to the capacity to stop and consider potential consequences of an action prior to engagement. Prior research has established that this facet of cognition may be damaged following exposures to violence like direct victimization, resulting in diminished impulse control for survivors of trauma in adolescence and also young adulthood (Davis et al., 2017; Mattheiss et al., 2022; Monahan et al., 2015). The prefrontal cortex governs communication across brain regions that facilitates impulse control and research has indicated that exposure to violence may damage development of these systems among adolescents (Cará, 2018; Cará et al., 2019; Mattheiss et al., 2022). This is particularly relevant of this population considering that their brains are developing at a rapid rate during this period of the life-course. Prior research has established that such dysfunctional development caused by exposure to violence, specifically increased variety of exposures to both direct and indirect violence, and may result in lower impulse control and that this may manifest in greater substance use risk in adolescent samples (Davis et al., 2017; Wojciechowski, 2022a). These processes may be particularly salient for adolescents, as the prefrontal cortex undergoes steady development during this period of the life-course, so interruption of this development may inhibit healthy development of impulse control (Steinberg, 2010; Steinberg et al., 2008). In this way, exposure to violence may operate via typical stress processes to increase risk for vaping among adolescents. However, stress processes triggered by exposure to violence may also operate through alternative mechanisms to influence substance use risk also.

Exposure to violence, whether directly or indirectly experienced, has generally been conceptualized as a stressor regardless of the nature of the exposure. While this may lead to dysfunctional cognitive development, prior research has also identified increased risk for deviant peer association as another outcome associated with exposure to violence among adolescents (Farrell et al., 2022; Rudolph et al., 2014). For example, Faus et al. (2019) observed that youth who have experienced community violence or abuse during childhood later reported a greater risk for affiliation with peers who engage in violent behavior themselves. Further, deviant peer association has been identified by prior research as a robust risk factor for tobacco/nicotine vaping outcomes, such as onset risk and frequency of use (Hesse & Fite, 2023; Lanza et al., 2020; Rocheleau et al., 2020). Prior research has also identified nicotine vaping as a coping response that individuals may use following exposure to violence, like directly being victimized or experiencing bullying (Boccio et al., 2022; Hansen et al., 2021; Wu et al., 2023). In sum, these studies highlight the potential that deviant peer association may also provide a mediating pathway linking exposure to violence to increased risk for vaping. Deviant peer association may offer a particularly relevant pathway to vaping following exposure to violence among adolescents specifically. While vaping is legal in the United States, youth are prohibited from buying or possessing vapes regardless of the substances that they disperse. This then necessitates illicit channels for adolescents to obtain vaping devices and cartridges. Deviant peers may offer one such channel. Prior research has indicated that substance using youth may act as social supply dealers that facilitate the transmission of illicit substances through their broader social networks (Chatwin & Potter, 2014; McLean, 2019). Youth who have been exposed to violence then may select into peer groups characterized by substance use behaviors among members of the group and vaping may be one potential option for coping with the effects of exposure to violence that the group can offer. In this way, adolescents may be able to procure vapes that may be used as a means of stress relief.

Prior research has identified deviant peer association as a significant mediator of relationships between exposure to violence and outcomes of interest among adolescents, like aggression and subsequent exposures to violence (Farrell et al., 2022; Lin et al., 2020; Wojciechowski, 2022b). For example, Wojciechowski (2022b) observed that deviant peer association mediated the relationship between exposure to violence and future exposure to violence. While this relationship was observed in the highlighted study, this is obviously different than the pathway to vaping as a behavioral outcome that is posited in the current study. This is problematic, as the theory of exposure to violence pushing youth into peer networks with social supply dealers that facilitate access to illicit substances has yet to be tested. Considering the general increased prevalence of vaping among adolescents over the past decade and the relevance of peer influence on behavior during adolescence (Farzal et al., 2019; Miech et al., 2021), this lack of attention here in particular is a major omission.

Prior research has identified exposure to violence, low impulse control, and deviant peer association as risk factors for vaping (Boccio et al., 2022; Davis et al., 2022; Hansen et al., 2021; Hesse & Fite, 2023; Lanza et al., 2020; Rocheleau et al., 2020; Wu et al., 2023). That said, there is a dearth of research which has examined how these risk factors may form a set of mediating pathways explaining youth vaping behaviors. Exposure to violence may trigger neurodevelopmental changes that result in diminished impulse control and prior research has established that such dysfunctional developmental may explain the relationship between exposure to violence and behavioral outcomes (Farrell et al., 2022; Lin et al., 2020). However, this mediating relationship has yet to be examined for vaping specifically. Relatedly, there has yet to be any study that has examined deviant peer association as a mediator in the relationship between exposure to violence and vaping in a manner consistent with the social supply dealer theory of peer facilitation of substance us behaviors. Further complicating these issues is that fact that exposure to violence may be either directly or indirectly experienced and that this distinction may drive variation in the salience of mediating relationships of interest. The differential salience of the hypothesized mediators in relation to direct versus indirect exposures to trauma remains a key issue as well. Identification of different mediating pathways to vaping dependent on whether exposure to violence was directly or indirectly experienced would provide important information that may facilitate more effective design and implementation of prevention programming to reduce vaping behaviors among adolescents. The present study sought to address these gaps in the literature by testing the following hypotheses:

Hypothesis 1

Direct and indirect exposure to violence will be associated with increased risk for vaping.

Hypothesis 2

Deviant peer association and impulse control will significantly mediate the relationships between the various forms of exposure to violence and vaping.

Methods

The present study utilized data from the Waves 1–4 of the Adolescent Brain Cognitive Development study. This is an ongoing panel study following 11,880 youth participants aged 9–10 at baseline, with the intent of continuing to collect data from these participants up through emerging adulthood. These waves were used because of the necessity to use at least three waves of data for the mediation analyses, but used the Wave 4 measure of vaping as the dependent variable because this was the first wave where there was the requisite variation on this measure for analyses. These data were collected at 24 study sites across the United States of America. Data collection for the study began in 2017. All minor participants provided assent to participate in the study and parents/guardians of the participants provided informed consent.

Measures

Vaping

The main dependent variable examined in these analyses was vaping at Wave 4. This binary measure delineated participants who reported any nicotine/tobacco vaping during the Wave 4 observation period from those who did not (0 = No; 1 = Yes). A total of 1.68% of participants (N = 105) reported vaping during the Wave 4 observation period.

Exposure to Violence

Two main independent variables examined in this study assessed direct and indirect exposures to violence experienced prior to the Wave 2 observation period. This counted any exposures that occurred prior to baseline and any that occurred between baseline measurements and the Wave 2 observation period. This was done in order to ensure that all exposures that occurred prior to Wave 2 were included in the measure to provide the necessary three waves to test for mediation and establish temporal ordering in the causal chain. The post-traumatic stress disorder module of the Kiddie Schedule of Affective Disorders and Schizophrenia was used to measure these constructs (Puig-Antich & Ryan, 1986). Participants were asked whether they had experienced various forms of direct and indirect exposure to violence measured using this instrument (e.g., Beaten to the point of having bruises; witnessed someone shot or stabbed). Binary variables were used to delineate participants who reported experiencing direct exposure to violence and indirect exposure to violence from those who did not report these exposures (0 = No; 1 = Yes). A total of 6.91% of participants (N = 776) reported experiencing direct exposure to violence prior to Wave 2 measurements and 28.18% of participants (N = 3163) reported experiencing indirect exposure to violence prior to this time.

Deviant Peer Association

One of the mediating variables examined in this study was deviant peer association at Wave 3. This construct was measured using The Peer Behavior Inventory (Bingham et al., 1995). This instrument asked participants to indicate the degree to which their peers were involved in various rule-breaking behaviors (skipping school, suspended from school, shoplifting) using an ordinal scale, with higher scores corresponding to more peers who were reported to be involved in these behaviors. Factor analysis was used to generate a single deviant peer association score for all participants using a varimax rotation. Factor loadings were adequate for all variables on this general factor (Shoplifting eigenvalue = .493, School suspension eigenvalue = .574, Skipping school eigenvalue = .551).

Impulse Control

The other mediator examined in these analyses was impulse control at Wave 3. This construct was measured using the UPPS Impulsive Behavior Scale (Whiteside et al., 2005). The lack of premeditation subscale was used to measure impulse control because of the shared focus of this scale and the impulse control construct on the capacity to consider consequences before acting (e.g., I like to stop and think about things before I do it). An ordinal scale was used to measure the individual items comprising this subscale, with higher scores corresponding to lower impulse control. Factor analysis was used to generate a single latent impulse control factor using a varimax rotation. Factor scores were adequate for all items included in the subscale (Item 1 = .611; Item 2 = .326; Item 3 = .331; Item 4 = .619).

Control Variables

Several control variables were also included in analyses to ensure that estimates were free of bias. Gender was the first of these control variable because past research has indicated that boys present greater risk for vaping than do girls (Tehrani et al., 2022). A binary variable was used to measure gender at baseline (0 = Boys; 1 = Girls).¹

It was also important to control for race/ethnicity in these analyses, as past research has identified racial/ethnic differences in vaping risk (Cambron, 2023). Race was measured at baseline using a three-category nominal variable with the following response options: White, Black, Other Race. Because of limited variation for the Other Race category for vaping, participants were delineated to compare White and Racially Minoritized participants using a binary variable (0 = Racially Minoritized; 1 = White). Latinx ethnicity was also controlled for using a binary variable that delineated participants who were Latinx from all other participants (0 = Non-Latinx; 1 = Latinx).

Parent annual income at baseline was also controlled for here, as prior research has provided evidence of the impact of socioeconomic status on vaping risk among youth (Adebisi et al., 2024). Parent annual income at baseline was measured using an ordinal variable, with higher scores corresponding to greater annual income.

It was also important to control for age, as it would be expected that risk for vaping would be greater among older adolescents based on prevalence rates observed in prior research (Chapman & Wu, 2014). Age was measured at Wave 4 in single-year intervals. Participants reported a mean age of 11.579 years of age at Wave 3 (Standard deviation = .704, Range = 9–14) and a mean age of 12.427 at Wave 4 (Standard deviation = .683, Range = 10–14).

The final control variable included in these analyses was parental monitoring, as past research has indicated that stronger parental monitoring of youth behaviors may reduce the risk for substance use behaviors (Szoko et al., 2021). Parental monitoring was measured at Wave 3 using the Parental Monitoring Survey (Chilcoat & Anthony, 1996). This instrument asked participants about their parents’ degree of knowledge of their behaviors and monitoring of their behaviors using a series of ordinal items, with greater scores indicating greater parental monitoring (e.g., How often do your parents/guardians know where you are?). A mean score was computed from these individual ordinal scores so that every participant had a single parental monitoring score at Wave 3.

Analytic Strategy

The present study utilized generalized structural equation modeling (GSEM) to examine direct and indirect relationships of interest. This method was chosen because of its capacity breakdown the individual components of causal pathways linking independent, mediating, and dependent variables and to be extended to test for the statistical significance of mediated relationships. GSEM was used over linear structural equation modeling because of the binary coding of the vaping dependent variable. Coefficients were interpreted as the predicted change in the log-odds of vaping given a one-unit increase in an independent variable of interest, net of all other variables in the model. Two models were estimated. Model 1 examined the direct effects of indirect and direct exposure to violence on vaping net of all control covariates. Model 2 examined these same relationships, but also included the hypothesized mediating pathways running through impulse control and deviant peer association to determine the magnitude of attenuation of the direct relationships of exposure to violence variables from Model 1 upon inclusion of the mediating pathways in Model 2. Listwise deletion was used to manage missing data because more advanced forms of missing data were not compatible with the gsem function in Stata. Sensitivity analyses were estimated using linear structural equation modeling with the binary dependent variable using full-information maximum likelihood estimation to manage missing data to test the robustness of these results.

The second phase of analyses entailed extension of the GSEM analyses to formally test whether any of the causal pathways tested in analyses constituted statistically significant mediation. Standard errors for each of the causal pathways were needed to conduct these formal tests of significance and compute p-values. While the delta method can be used to compute these standard errors, this can result in non-normally distributed standard errors that leads to bias in estimation of statistical significance. The Preacher and Hayes (2008) method of bootstrap resampling addresses this concern and this method was used in these analyses. A total of 500 bootstrap iterations were carried out and pooled in these analyses to compute unbiased estimates of statistical significance of mediation effects.

Results

Table 1 provides descriptive statistics for all variables examined in these analyses. Table 2 then provides Model 1 and Model 2 GSEM results. Model 1 results indicated that experiencing lifetime indirect exposure to violence prior to Wave 2 measurements was associated with increased risk of vaping (Coefficient = .539, p < .017). Direct exposure to violence history was not a significant predictor of vaping risk in this model. Lower income, being older, and weaker parental monitoring were also all significant predictors of increased vaping risk in this model also. Model 2 results indicated that inclusion of the hypothesized mediating pathways attenuated the direct effect of indirect exposure to violence on vaping risk to non-significance (Coefficient = .126, p < .653), with about 80% of the effect observed in Model 1 accounted for. The direct effect of direct exposure to violence on vaping remained non-significant in this model. Greater deviant peer association significantly predicted increased risk of vaping in this model (Coefficient = .965, p < .001). Impulse control was not a significant predictor of vaping in this model. Being older and weaker parental monitoring were also significant predictors of vaping. Indirect exposure to violence was a significant predictor of greater deviant peer association also (Coefficient = .123, p < .001), but direct exposure to violence was not. Both indirect exposure to violence (Coefficient = .040, p < .007) and direct exposure to violence (Coefficient = .088, p < .001) were significant predictors of lower impulse control in this model as well.

Table 1.

Descriptive Statistics.

	Mean/Proportion	Standard Deviation	Minimum	Maximum
Vaping at wave 4No = 98.32% (N = 6146)Yes = 1.68% (N = 105)
Direct exposure to violence prior to wave 2No = 93.09% (N = 10,449)Yes = 6.91% (N = 776)
Indirect exposure to violence prior to wave 2No = 71.82% (N = 8.062)Yes = 28.18% (N = 3163)
Deviant peer association at wave 3	<.001	.720	−.367	6.587
Impulse control at wave 3	<.001	.682	−1.023	2.136
Gender
Female = 47.83% (N = 5680)
Male = 52.17% (N = 6196)
Race^a
Racially Minoritized = 25.87% (N = 3072)
White = 74.13% (N = 8804)
Latinx ethnicity
No = 79.43% (N = 9312)
Yes = 20.57% (N = 2411)
Annual family income at baseline	6.840	2.678	1	10
Age at wave 4	12.428	.683	10	14
Parental monitoring at wave 3	4.484	.466	1	5

^aPercentages do not add up to 100% because of some overlap between categories due to multiracial participants.

Table 2.

Generalized Structural Equation Logistic Regression Modeling of Exposure to Violence on Log-Odds of Vaping at Wave 4 and Proposed Mediating Pathways (N = 5489).

	Model 1			Model 2
	Coefficient	p-Value	95% Confidence Interval	Coefficient	p-Value	95% Confidence Interval
Direct exposure to violence prior to wave 2 (0 = No; 1 = Yes)	−.150	.714	−.954 .654	.195	.665	−.689 1.080
Indirect exposure to violence prior to wave 2 (0 = No; 1 = Yes)	.539	.017	.095 .983	.126	.653	−.424 .676
Deviant peer association wave 3				.965	<.001	.783 1.147
Impulse control wave 3				.101	.589	−.265 .467

Gender (0 = Male; 1 = Female)	.332	.135	−.103 .768	.519	.052	−.004 1.043
Race (0 = Racially minoritized; 1 = White)	.350	.242	−.236 .936	.654	.091	−.105 1.412
Latinx ethnicity (0 = No; 1 = Yes)	−.202	.447	−.723 .319	−.224	.486	−.853 .406
Parental annual income at baseline	−.108	.032	−.206 −.009	−.016	.811	−.146 .114
Age wave 4	.903	<.001	.559 1.247	1.124	<.001	.677 1.572
Parental monitoring wave 3	−.943	<.001	−1.334 −.551	−.647	.014	−1.163 −.131
Constant	−10.855	<.001	−15.478–6.232	−16.285	<.001	−22.576–10.194
Direct exposure to Violence→Impulse control				.088	.001	.036 .139
Direct exposure to Violence→Deviant peer association				.034	.278	−.027 .095
Indirect exposure to Violence→Impulse control				.040	.007	.011 .069
Indirect exposure to Violence→Deviant peer association				.123	<.001	.088 .157

The second phase of analyses entailed determination of whether any of the causal pathways tested here constituted statistically significant mediation. The Preacher and Hayes (2008) method of bootstrap resampling was used to compute these estimates of statistical significance. These findings indicated that there was only one statistically significant mediation effect observed in the main analyses. The relationship between indirect exposure to violence and vaping was significantly mediated by deviant peer association (Coefficient = .119; Standard error = .023; p < .001; 95% confidence interval = .075---.164). None of the other mediation effects pertaining to indirect nor direct exposure to violence were statistically significant.

Sensitivity analyses entailed re-examination of these mediation effects using linear structural equation modeling with the binary vaping dependent variable and full-information maximum likelihood estimation to manage missing data. Findings from these analyses were analogous to those of the main analyses, with only indirect exposure to violence exerting a statistically significant effect on vaping risk, deviant peer association significantly mediating this relationship, and this being the only statistically significant mediation effect.

There were concerns about small cell sizes for the overlap of participants who reported either form of exposure to violence and vaping considering that only 105 participants reported vaping during the observation period of interest. For indirect victimization, there were 40 participants who reported both experiencing indirect exposure to violence and vaping, assuaging some concerns for this specific relationship. However, there were only 7 participants who reported both experiencing both direct victimization and vaping, indicating that the nonsignificant effect here is likely driven by a lack of statistical power.

Discussion

This study provided an important examination of the relationship between exposure to violence to vaping and the stress mechanisms linking the two. Only indirect exposure to violence was observed to be a significant predictor of increased vaping risk, but not direct exposure to violence. Consistent with hypotheses, deviant peer association significantly mediated the relationship between indirect exposure to violence and vaping risk. However, deviant peer association was not a significant mediator of the nonsignificant relationship between direct exposure to violence and vaping. Impulse control was not a significant mediator of either hypothesized relationship. There are a number of important implications of these findings for the prevention of vaping and treatment design and implementation for adolescents.

As mentioned above, only indirect exposure to violence was observed as a significant predictor of vaping, but not direct exposure to violence. Sensitivity analyses indicated that these results were robust even when only one or the other form of exposure to violence was included as a predictor in the GSEM models. These findings suggest that youth who have witnessed violence in their childhoods, but not directly experienced victimization themselves, should be prioritized for treatment to prevent vaping specifically. That said, this finding is inconsistent with prior research which has indeed identified directly experienced exposure to violence as a risk factor for vaping (Boccio et al., 2022; Wu et al., 2023). While this should temper the interpretation of this specific finding as potentially being an outlier, there should still be consideration that programming specifically targeting the population of youth who have indirectly experienced exposure to violence may be beneficial for preventing vaping. Trauma-informed care (TIC) is a treatment philosophy focused on avoiding re-traumatization during the treatment process. TIC prioritizes three pillars during treatment: providing a safe treatment environment, fostering trusting connections in interpersonal relationships, and providing healthy coping responses for managing emotions (Bath, 2008). In focusing program staff’s efforts in these areas, it is believed that overall self-regulation can be strengthened and that clients may seek out supportive relationships and/or prioritize utilizing healthier coping responses when future stressors arise; including lingering stress and related symptoms that may stem from the antecedent stressor that led them to seek treatment in the first place. Considering that high magnitude stress can lead to lasting dysfunction in the stress processes systems of the body that leads to a diminished capacity to manage future stress (McEwen, 1998; Stam, 2007), such an approach can potentially reduce the use of vaping as a coping response for these youth. This, of course, remains speculative. While TIC has demonstrated some promise for addressing some outcomes (Burge et al., 2021; Sullivan et al., 2016), there is still limited evidence-based work indicating its effectiveness. Part of this has to do with the fact that TIC is more of an overarching treatment philosophy than it is a discrete program. However, this also lends a high degree of flexibility to TIC that may be amenable to addressing the mediating mechanism of interest observed in the findings of this study, that is, deviant peer association.

Deviant peer association was observed to significantly mediate the relationship between indirect exposure to violence and vaping risk, but did not mediate the relationship between direct exposure to violence and vaping risk. Both direct and indirect exposure to violence were indeed significant predictors of greater deviant peer association though, a finding consistent with prior research on the topic (Farrell et al., 2022; Rudolph et al., 2014). The significant mediation observed here indicates that vaping risk could be reduced among youth who have witnessed violence by designing and implementing prevention programming oriented around reducing deviant peer association. Mentoring programs may have some utility in this regard, as these programs have demonstrated some utility for impacting behavioral outcomes (Grossman & Tierney, 1998; Sullivan & Jolliffe, 2012; Tolan et al., 2014). These programs focus on pairing at-risk youth with a prosocial model with the intent of reducing contact with antisocial influences and attenuating deviant socialization processes. Such programming may then have some capacity to disrupt deviant peer relationships that may provide channels through which youth obtain vaping devices and through which deviance contagion may occur. However, there are concerns about the validity of findings from some evaluations of such programs (Abrams et al., 2014; Harp, 2020; Pitzel et al., 2021). Mentoring programs have demonstrated some promise for reducing vaping outcomes specifically also (Lyu et al., 2022; Wyman et al., 2021). For example, a peer-led program called Above the Influence of Vaping was demonstrated to have utility in this regard. Peer leaders were trained to deliver programming to other adolescents and results indicated that friends of peer leaders reported lower risk of vaping with data analyzed using social network analysis (Wyman et al., 2021). Findings from studies like this suggest that mentoring programs can have some capacity for reducing vaping risk, though the specific utility for addressing this behavior among youth who have experienced indirect exposure to violence remains understudied at this point and in need of additional attention.

One potential implication for design and implementation of mentoring programming that may be particularly effective for reducing vaping risk rooted in these findings is potentially orienting these mentoring programs under a TIC philosophy. Because TIC recognizes the realities of trauma exposure that may make recovery difficult concerns forming trusting and safe relationships, training mentors in the TIC philosophy may make them particularly skilled in working with the specific population identified as being at-risk for vaping here: youth who have witnessed violence. Prior research has indicated that youth who have been exposed to traumatic stress are at increased risk for facing social rejection by peers (Demol et al., 2020; Hodges & Perry, 1999), which may contribute to such youth coalescing into antisocial peer groups. Providing TIC training for mentors may facilitate a better understanding or how to work with such youth to develop skills that would allow for them to form more connections with prosocial peers and reduce the likelihood that they will seek antisocial peer ties. Doing so would potentially address both risk factors of indirect exposure to violence and deviant peer association to reduce risk for vaping among adolescents. That said, this remains speculative, as such programming that seeks to address both of these risk factors in this manner has yet to be tested. This indicates the need for pilot studies to be conducted on such programming to identify the potential efficacy of this dual pronged approach that may inform scaling up of any interventions that can reduce vaping risk among this subpopulation of youth.

The present study provided an important examination of the pathways linking exposure to violence to vaping risk, but there remain a number of important limitations. The first limitation pertains to the coarse measures of exposure to violence examined in this study as predictors of vaping. While the present study did delineated indirect and direct forms of exposure to violence as a novel contribution to the literature in this regard, there are also other important ways in which exposure to violence could be examined in different domains. For example, examining physical versus sexual violence exposure could indicate that only one or the other is relevant for predicting vaping in a manner consistent with what was observed in the present study for direct and indirect exposure to violence. Examining violence exposures occurring in different contexts could matter in this regard as well. Experiencing exposure to violence in the community, the home, school, etc. could all have differential salience for predicting vaping behaviors as well. Such examinations, however, were beyond the scope of this study. This indicates the need for future research to re-examine these processes with exposure to violence delineated in these additional ways to best understand the most salient predictors of vaping. Another limitation of this study is the use of a binary dependent variable using the GSEM method. While this is the appropriate modeling strategy given the coding of this variable, advanced forms of missing data analysis were not compatible with this function in Stata, leaving listwise deletion of missing data as the only option. This is problematic because data not missing at random may have led to bias in estimation of relationships of interest. While the sensitivity analyses attempted to address this issue using linear structural equation modeling with full-information maximum likelihood estimation to manage missing data, the use of this method with a binary dependent variable again leaves open the possibility of biased estimates of significance. Even though the results observed in the sensitivity analyses were consistent with the results of the main analyses, these methodological limitations indicate the need for re-examination of these processes using data and methods that can address these shortcomings as a whole. Another limitation pertains to the small cell size for participants who reported both experiencing direct exposure to violence and vaping (N = 7). This undoubtedly led to a power issue that drove the nonsignificant relationship observed in analyses. This would also help explain this finding that was inconsistent with prior research on this topic. This indicated the need to re-examine these relationships using data that is not hindered by these data limitations. Future research should seek to explore these relationships of interest using such data to determine whether findings are robust absent of issues with statistical power.

A final concerning limitation pertains to current data limitations at the time of writing this manuscript. Reviewers suggested several revisions to analyses pertaining specifically to inclusion of other variables that are available in the ABCD dataset. Unfortunately, these revisions were unable to be carried out. The ABCD dataset is stored within hundred of subscale data files with uninstructive variable names and labels. The only way for researchers not affiliated with data collection to identify what specific variables measure in each of these subscale datasets is through referencing the online ABCD codebook. Unfortunately, at the time of the decision for revise and resubmit, the ABCD team has taken the codebook offline in anticipation of the new 6.0 data release and have provided indication that the codebook will be back online once it has been updated for the new data release. However, this data release has already been delayed for over a year and there is no indication of specifically when the release will occur nor when the updated codebook will be back online. This unique circumstance has led to an inability to revise the manuscript in several potentially important ways through inclusion of additional variables in analyses. These variables include: earlier vaping experiences prior to Wave 3, motivations for vaping, and peer engagement in vaping. It is believed that inclusion of these variables in analyses would provide a more valid examination of relationships of interest, but the current issue precluded their inclusion. Future research should seek to re-examine these relationships while able to account for these additional constructs.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Thomas Wojciechowski

Note

Author Biography

Thomas Wojciechowski is an assistant professor with the Michigan State University School of Criminal Justice. His research mainly focuses on substance use, mental health, and human development.

References

Abrams

L. S.

Mizel

M. L.

Nguyen

Shlonsky

(2014). Juvenile reentry and aftercare interventions: Is mentoring a promising direction? Journal of Evidence-Based Social Work, 11(4), 404–422. https://doi.org/10.1080/10911359.2014.897115

Adebisi

Y. A.

Bafail

D. A.

Oni

O. E.

(2024). Prevalence, demographic, socio-economic, and lifestyle factors associated with cigarette, e-cigarette, and dual use: Evidence from the 2017–2021 scottish health Survey. Internal and Emergency Medicine, 19(8), 1–15. https://doi.org/10.1007/s11739-024-03716-2

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30(1), 47–88. https://doi.org/10.1111/j.1745-9125.1992.tb01093.x

Bath

(2008). The three pillars of trauma-informed care. Reclaiming children and youth, 17(3), 17–21.

Beharie

Scheidell

J. D.

Quinn

McGorray

Vaddiparti

Kumar

P. C.

Frueh

B. C.

Boone

Khan

M. R.

(2019). Associations of adolescent exposure to severe violence with substance use from adolescence into adulthood: Direct versus indirect exposures. Substance Use & Misuse, 54(2), 191–202. https://doi.org/10.1080/10826084.2018.1495737

Bingham

C. R.

Fitzgerald

H. E.

Zucker

R. A.

(1995). Peer behavior profile/peer activities questionnaire. Unpublished questionnaire. Department of Psychology, Michigan State University.

Boccio

C. M.

Leal

W. E.

Jackson

D. B.

(2022). Bullying victimization and nicotine and marijuana vaping among Florida adolescents. Drug and Alcohol Dependence, 237, 109536. https://doi.org/10.1016/j.drugalcdep.2022.109536

Bold

K. W.

Morean

M. E.

Kong

Simon

Camenga

D. R.

Cavallo

D. A.

Krishnan-Sarin

(2017). Early age of e-cigarette use onset mediates the association between impulsevity and e-cigarette use frequency in youth. Drug and Alcohol Dependence, 181, 146–151. https://doi.org/10.1016/j.drugalcdep.2017.09.025

Burge

Tickle

Moghaddam

(2021). Evaluating trauma informed care training for services supporting individuals experiencing homelessness and multiple disadvantage. Housing, Care and Support, 24(1), 14–25. https://doi.org/10.1108/hcs-01-2021-0002

10.

Cambron

(2023). Racial/ethnic differences in vaping product use among youth: A state-level analysis. International Journal of Environmental Research and Public Health, 20(9), 5729. https://doi.org/10.3390/ijerph20095729

11.

Cará

V. M.

(2018). Youth violence and the brain: Exposure to violence and deactivation of an insular-frontal network of cortical areas associated with inhibitory control.

12.

Cará

V. M.

Esper

N. B.

De Azeredo

L. A.

Iochpe

Dalfovo

N. P.

Santos

R. C.

Sanvicente-Vieira

Grassi-Oliveira

Franco

A. R.

Buchweitz

(2019). An fMRI study of inhibitory control and the effects of exposure to violence in Latin-American early adolescents: Alterations in frontoparietal activation and performance. Social Cognitive and Affective Neuroscience, 14(10), 1097–1107. https://doi.org/10.1093/scan/nsz092

13.

Chapman

S. L. C.

L. T.

(2014). E-Cigarette prevalence and correlates of use among adolescents versus adults: A review and comparison. Journal of Psychiatric Research, 54, 43–54. https://doi.org/10.1016/j.jpsychires.2014.03.005

14.

Chatwin

Potter

(2014). Blurred boundaries: The artificial distinction between “use” and “supply” in the UK cannabis market. Contemporary Drug Problems, 41(4), 536–550.

15.

Chilcoat

H. D.

Anthony

J. C.

(1996). Impact of parent monitoring on initiation of drug use through late childhood. Journal of the American Academy of Child & Adolescent Psychiatry, 35(1), 91–100. https://doi.org/10.1097/00004583-199601000-00017

16.

Davis

D. R.

Bold

K. W.

Morean

M. E.

Kong

Jackson

Simon

Rajesh-Kumar

Krishnan-Sarin

(2022). Association of youth impulsivity and use of e-cigarette devices, flavors, and frequency of use. Addictive Behaviors, 134, 107386. https://doi.org/10.1016/j.addbeh.2022.107386

17.

Davis

J. P.

Dumas

T. M.

Berey

B. L.

Merrin

G. J.

Cimpian

J. R.

Roberts

B. W.

(2017). Effect of victimization on impulse control and binge drinking among serious juvenile offenders from adolescence to young adulthood. Journal of Youth and Adolescence, 46(7), 1515–1532. https://doi.org/10.1007/s10964-017-0676-6

18.

Demol

Leflot

Verschueren

Colpin

(2020). Revealing the transactional associations among teacher-child relationships, peer rejection and peer victimization in early adolescence. Journal of Youth and Adolescence, 49(11), 2311–2326. https://doi.org/10.1007/s10964-020-01269-z

19.

Dixon

K. E.

Bindbeutel

K. M.

Daugherty

Y. T.

Robertson

A. C.

Lee

Galano

M. M.

Mahoney

C. T.

(2024). The differential impact of childhood trauma on adult impulsivity and impulse control. Traumatology, 30(3), 468–468. https://doi.org/10.1037/trm000044

20.

Farrell

A. D.

Pittman

S. K.

O'Connor

K. E.

Sullivan

T. N.

(2022). Peer factors as mediators of relations between exposure to violence and physical aggression in middle school students in a low-income urban community. Psychology of Violence, 12(3), 170–182. https://doi.org/10.1037/vio0000405

21.

Farzal

Perry

M. F.

Yarbrough

W. G.

Kimple

A. J.

(2019). The adolescent vaping epidemic in the United States—how it happened and where we go from here. JAMA Otolaryngology – Head & Neck Surgery, 145(10), 885–886. https://doi.org/10.1001/jamaoto.2019.2410

22.

Faus

D. P.

de Moraes

C. L.

Reichenheim

M. E.

da Matta Souza

L. M. B.

Taquette

S. R.

(2019). Childhood abuse and community violence: Risk factors for youth violence. Child Abuse & Neglect, 98, 104182. https://doi.org/10.1016/j.chiabu.2019.104182

23.

Gilreath

T. D.

Astor

R. A.

Estrada

J. N.

Benbenishty

Unger

J. B.

(2014). School victimization and substance use among adolescents in California. Prevention Science: The Official Journal of the Society for Prevention Research, 15(6), 897–906. https://doi.org/10.1007/s11121-013-0449-8

24.

Grossman

J. B.

Tierney

J. P.

(1998). Does mentoring work? An impact study of the big brothers big sisters program. Evaluation Review, 22(3), 403–426.

25.

Hansen

Morgenstern

Isensee

Galimov

Hanewinkel

(2021). Association between bullying victimization and e‐cigarette use among German students. Aggressive Behavior, 47(4), 421–429. https://doi.org/10.1002/ab.21951

26.

Harp

(2020). In focus: Mentoring youth (OJJDP in Focus, NCJ 253045). Office of Juvenile Justice and Delinquency Prevention. https://ojjdp.ojp.gov/library/publications/focus-mentoring-youth

27.

Henry

B. F.

(2020). Typologies of adversity in childhood & adulthood as determinants of mental health & substance use disorders of adults incarcerated in US prisons. Child Abuse & Neglect, 99, 104251. https://doi.org/10.1016/j.chiabu.2019.104251

28.

Hesse

D. R.

Fite

P. J.

(2023). Sources of socialization in E-cigarette initiation among rural middle school-age youth. Journal of Child and Family Studies, 33(9), 1–12. https://doi.org/10.1007/s10826-023-02705-x

29.

Hodges

E. V.

Perry

D. G.

(1999). Personal and interpersonal antecedents and consequences of victimization by peers. Journal of Personality and Social Psychology, 76(4), 677–685. https://doi.org/10.1037//0022-3514.76.4.677

30.

Lanza

H. I.

Motlagh

Orozco

(2020). E-Cigarette use among young adults: A latent class analysis examining co-use and correlates of nicotine vaping. Addictive Behaviors, 110, 106528. https://doi.org/10.1016/j.addbeh.2020.106528

31.

Lawrence

T. I.

Mcfield

A. A.

Byrne

M. M.

Tarver

Stewart

T. K.

(2023). Depression and substance use as consequences of exposure to family violence: A moderation mediation and self-medication hypothesis study. Journal of Child & Adolescent Trauma, 16(1), 69–79. https://doi.org/10.1007/s40653-022-00464-3

32.

Lee

T. H.

Kim

(2021). Differences in electronic cigarette use among adolescents in Korea: A nationwide analysis. Journal of Substance Abuse Treatment, 131, 108554. https://doi.org/10.1016/j.jsat.2021.108554

33.

Levy

D. T.

Tam

Sanchez-Romero

L. M.

Yuan

Jeon

Meza

(2021). Public health implications of vaping in the USA: The smoking and vaping simulation model. Population Health Metrics, 19, 1–18. https://doi.org/10.1186/s12963-021-00250-7

34.

Lin

Chen

Zhang

Cao

Liu

(2020). Predicting adolescent aggressive behavior from community violence exposure, deviant peer affiliation and school engagement: A one-year longitudinal study. Children and Youth Services Review, 111, 104840. https://doi.org/10.1016/j.childyouth.2020.104840

35.

Lyu

J. C.

Afolabi

White

J. S.

Ling

P. M.

(2022). Perceptions and aspirations toward peer mentoring in social media–based electronic cigarette cessation interventions for adolescents and young adults: Focus group study. JMIR Formative Research, 6(12), Article e42538. https://doi.org/10.2196/42538

36.

Martheswaran

Shmunes

M. H.

Ronquillo

Y. C.

Moshirfar

(2021). The impact of vaping on ocular health: A literature review. International Ophthalmology, 41(8), 2925–2932. https://doi.org/10.1007/s10792-021-01842-w

37.

Mattheiss

S. R.

Levinson

Rosenberg-Lee

Graves

W. W.

(2022). Exposure to violence is associated with decreased neural connectivity in emotion regulation and cognitive control, but not working memory, networks after accounting for socioeconomic status: A preliminary study. Behavioral and Brain Functions: BBF, 18(1), 15. https://doi.org/10.1186/s12993-022-00201-8

38.

McEwen

B. S.

(1998). Stress, adaptation, and disease: Allostasis and allostatic load. Annals of the New York Academy of Sciences, 840(1), 33–44.

39.

McLean

(2019). Gangs and drug supply. In Gangs, drugs and (Dis) organised crime (pp. 143–174). Bristol University Press.

40.

Miech

Leventhal

Johnston

O’Malley

P. M.

Patrick

M. E.

Barrington-Trimis

(2021). Trends in use and perceptions of nicotine vaping among US youth from 2017 to 2020. JAMA Pediatrics, 175(2), 185–190. https://doi.org/10.1001/jamapediatrics.2020.5667

41.

Monahan

K. C.

King

K. M.

Shulman

E. P.

Cauffman

Chassin

(2015). The effects of violence exposure on the development of impulse control and future orientation across adolescence and early adulthood: Time-specific and generalized effects in a sample of juvenile offenders. Development and Psychopathology, 27(4 pt 1), 1267–1283. https://doi.org/10.1017/S0954579414001394

42.

Nicholson

Scott

Newton Ede

Jones

S. W.

(2021). The impact of E-cigarette vaping and vapour constituents on bone health. Journal of Inflammation, 18(1), 16. https://doi.org/10.1186/s12950-021-00283-7

43.

Pinchevsky

G. M.

Fagan

A. A.

Wright

E. M.

(2014). Victimization experiences and adolescent substance use: Does the type and degree of victimization matter? Journal of Interpersonal Violence, 29(2), 299–319. https://doi.org/10.1177/0886260513505150

44.

Pitzel

Kearley

Jolivette

Sanders

(2021). Contextualizing mentoring programs into juvenile justice facilities. Journal of Correctional Education, 72(2), 5–23.

45.

Preacher

K. J.

Hayes

A. F.

(2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40(3), 879–891. https://doi.org/10.3758/brm.40.3.879

46.

Puig-Antich

Ryan

(1986). Kiddie schedule for affective disorders and schizophrenia. Western Psychiatric Institute.

47.

Rocheleau

G. C.

Vito

A. G.

Intravia

(2020). Peers, perceptions, and e-cigarettes: A social learning approach to explaining e-cigarette use among youth. Journal of Drug Issues, 50(4), 472–489. https://doi.org/10.1177/0022042620921351

48.

Rudolph

K. D.

Lansford

J. E.

Agoston

A. M.

Sugimura

Schwartz

Dodge

K. A.

Pettit

G. S.

Bates

J. E.

(2014). Peer victimization and social alienation: Predicting deviant peer affiliation in middle school. Child Development, 85(1), 124–139. https://doi.org/10.1111/cdev.12112

49.

Saadatmand

Dearfield

Bronson

Harrison

(2022). Exposure to personal and community violence and associated drug use outcomes in African American young adults. Journal of Ethnicity in Substance Abuse, 21(2), 708–729. https://doi.org/10.1080/15332640.2020.1795040

50.

Stam

(2007). PTSD and stress sensitisation: A tale of brain and body: Part 1: Human studies. Neuroscience & Biobehavioral Reviews, 31(4), 530–557. https://doi.org/10.1016/j.neubiorev.2006.11.010

51.

Steinberg

(2010). A dual systems model of adolescent risk‐taking. Developmental Psychobiology, 52(3), 216–224. https://doi.org/10.1002/dev.20445

52.

Steinberg

Albert

Cauffman

Banich

Graham

Woolard

(2008). Age differences in sensation seeking and impulsivity as indexed by behavior and self-report: Evidence for a dual systems model. Developmental Psychology, 44(6), 1764–1778. https://doi.org/10.1037/a0012955

53.

Sullivan

C. J.

Jolliffe

(2012). Peer influence, mentoring, and the prevention of crime. The Oxford Handbook of Crime Prevention (pp. 207–225).

54.

Sullivan

K. M.

Murray

K. J.

Ake III

G. S.

(2016). Trauma-informed care for children in the child welfare system: An initial evaluation of a trauma-informed parenting workshop. Child Maltreatment, 21(2), 147–155. https://doi.org/10.1177/1077559515615961

55.

Sun

(2024). Early onset of e-cigarette use and subsequent use frequency among US high school students. Preventive Medicine Reports, 48, 102935. https://doi.org/10.1016/j.pmedr.2024.102935

56.

Szoko

Ragavan

M. I.

Khetarpal

S. K.

Chu

K. H.

Culyba

A. J.

(2021). Protective factors against vaping and other tobacco use. Pediatrics, 148(2), Article e2020048066. https://doi.org/10.1542/peds.2020-048066

57.

Tehrani

Rajabi

Ghelichi-Ghojogh

Nejatian

Jafari

(2022). The prevalence of electronic cigarettes vaping globally: A systematic review and meta-analysis. Archives of Public Health, 80(1), 240. https://doi.org/10.1186/s13690-022-00998-w

58.

Tolan

P. H.

Henry

D. B.

Schoeny

M. S.

Lovegrove

Nichols

(2014). Mentoring programs to affect delinquency and associated outcomes of youth at risk: A comprehensive meta-analytic review. Journal of Experimental Criminology, 10(2), 179–206. https://doi.org/10.1007/s11292-013-9181-4

59.

Whiteside

S. P.

Lynam

D. R.

Miller

J. D.

Reynolds

S. K.

(2005). Validation of the UPPS impulsive behaviour scale: A four‐factor model of impulsivity. European Journal of Personality, 19(7), 559–574. https://doi.org/10.1002/per.556

60.

Wojciechowski

(2022a). Exposure to violence variety as a risk factor for cigarette smoking: Relevance of sensation-seeking and impulsivity as mediators. Substance Use & Misuse, 57(2), 239–248. https://doi.org/10.1080/10826084.2021.2002901

61.

Wojciechowski

(2022b). Understanding how cumulative victimization predicts future victimization risk: Relevance of cognitive versus social mediation. Criminal Justice Review, 47(4), 417–433. https://doi.org/10.1177/07340168211015725

62.

Wright

E. M.

Fagan

A. A.

Pinchevsky

G. M.

(2013). The effects of exposure to violence and victimization across life domains on adolescent substance use. Child Abuse & Neglect, 37(11), 899–909. https://doi.org/10.1016/j.chiabu.2013.04.010

63.

Conklin

(2023). The relationship between victimization and vaping results from monitoring the future. International Journal of Law, Crime and Justice, 74, 100613. https://doi.org/10.1016/j.ijlcj.2023.100613

64.

Wyman

P. A.

Rulison

Pisani

A. R.

Alvaro

E. M.

Crano

W. D.

Schmeelk-Cone

… Espelage

D. L.

(2021). Above the influence of vaping: Peer leader influence and diffusion of a network-informed preventive intervention. Addictive Behaviors, 113, 106693.

65.

Zinzow

H. M.

Ruggiero

K. J.

Hanson

R. F.

Smith

D. W.

Saunders

B. E.

Kilpatrick

D. G.

(2009). Witnessed community and parental violence in relation to substance use and delinquency in a national sample of adolescents. Journal of Traumatic Stress, 22(6), 525–533. https://doi.org/10.1002/jts.20469