Understanding How Cumulative Victimization Predicts Future Victimization Risk: Relevance of Cognitive Versus Social Mediation

Abstract

Cumulative victimization represents the summation of victimization experiences across multiple contexts, with greater accumulation generally predicting greater dysfunction than less accumulation of exposures. Past research has indicated that cumulative victimization predicts increased risk for future revictimization also. The dual systems model may help to understand this relationship. This framework comprises constructs of sensation-seeking and impulse control in developmental context. Deviant peer association may provide a social factor that helps to understand this relationship. Victimization has been found to influence all of these constructs identified here. It is predicted that increased accumulation of victimization experiences may drive variation in these constructs that results in elevated risk for revictimization. This study sought to test the theory that each of these three constructs independently mediated the cumulative victimization–revictimization relationship. The Pathways to Desistance data were used in analyses. This sample was comprised of 1,354 juvenile offenders followed for 7 years after a recent adjudication prior to baseline measurements. The first three waves of data were used in analyses. Generalized structural equation modeling was used to test for the relationships of interest. A bootstrapping process of computing standard errors was carried out to determine significance of mediation effects. Results indicated that increased cumulative victimization scores at baseline predicted increased probability of experiencing victimization at Wave 3. This relationship was attenuated by about 15% when all mediators were added to the model and the relationship remained significant. Further analyses indicated that the specific indirect effect running through deviant peer association was significant, as was the total indirect effect. Findings indicate that increases in cumulative victimization may result in increased affiliation with deviant peers that further increases their future victimization risk. Service providers for survivors of violence should focus on screening of social relationships of those they provide care for in order to assess safety concerns.

Keywords

criminal victimization quantitative methods other crime over the lifecourse crime/delinquency theory

Past research has indicated that individuals who have experienced victimization have high rates of being revictimized later in life (Pantalone et al., 2015; Walker et al., 2019). Cumulative victimization, or the number of different types of domains within which one has experienced victimization, has been found to increase this risk for future victimization as well (Cole et al., 2008; Edalati et al., 2016). While this is a well-established relationship, there is less consensus as to why this phenomenon exists. There exist many competing theories as to why this may occur, ranging from psychological to sociological explanations. The present study sought to understand why increased cumulative victimization scores in childhood/early adolescence predicts increased risk of experiencing victimization later in life. Psychological (impulse control; sensation-seeking) and sociological (deviant peer association) variables were tested to determine whether they mediated this relationship. It was proposed that increased cumulative victimization early in the life course led to variation in these constructs of interest and that this variation was associated with increased risk for experiencing victimization. These relationships were tested using longitudinal data in order to establish temporal ordering in causal relationships of interest. The data utilized in this study were from juvenile offenders, as population established as having elevated risk for experiencing victimization (Baglivio et al., 2014).

Cumulative Victimization as a Predictor of Revictimization Risk

While past research has indicated that individuals who are victimized early in life may be at elevated risk to experience victimization again later in life (Barnes et al., 2009; Widom et al., 2008), this risk may be compounded by the degree of severity of this victimization. Past research has noted difficulties in the measurement of victimization in a general sense that may lead to underreporting of experiences (Smith, 1994). Additional issues arise when attempting to operationalize what is meant by victimization severity. Understanding cumulative victimization in terms of the number of types of trauma that a person has experienced is one means of understanding severity of trauma exposure that has been used in prior research (Kennedy et al., 2014; Mounier & Andujo, 2003). Cumulative victimization describes patterns of victimization that may occur across multiple domains (Grasso et al., 2013; Kennedy et al., 2014; Mustanski et al., 2016); thus, providing a victimization variety score. This measurement is often used because of issues with recall bias related to trauma exposure frequency, as individuals suffering from long-term patterns of trauma may lack the capacity to be precise in recalling this aspect of trauma exposure (Skogan, 1981; 1986; Wolfer, 1999). As such, variety of victimization exposure can act as a proxy measure of severity of victimization exposure and has been used by past research in various ways to assess victimization severity (Finkelhor et al., 2005; Wojciechowski, 2020). While victimization in one or multiple domains may both lead to iatrogenic outcomes, greater mental health problems are observed among individuals who score higher in variety and have thus accumulated more victimization (Dierkhising et al., 2019; Evans et al., 2014; Ford et al., 2010; Golder & Logan, 2011; Turner et al., 2006). One example of this was observed by Yoder et al. (2019), as they found that the cumulative victimization-revictimization risk was partially mediated by increased trauma symptomatology.¹ This indicates validity for this operationalization of victimization severity, as individuals reporting greater cumulative victimization scores generally report greater dysfunction; thus, indicating greater severity of victimization. Considering the impacts that greater cumulative victimization appears to have on psychological and behavioral outcomes, this dysfunction may help to understand revictimization risk later in life. It may be that such dysfunction offers mediating pathways through which previous victimization experiences result in increased risk for revictimization later in life. It is then expected that greater cumulative victimization would be associated with greater dysfunction in potential cognitive and social domains which exacerbate risk for later revictimization.

Psychological and Sociological Constructs as Mediators

There exist both sociological and psychological explanations for why individuals who have experienced high levels of victimization early in life demonstrate high risk for experiencing revictimization. The dual systems model provides one psychological framework for understanding why this high risk of revictimization is observed. While there is a great deal of research indicating that low self-control is associated with victimization risk (Higgins et al., 2009; Pratt et al., 2014; Schreck et al., 2006), little of this research has sought to understand the concept of self-control as being more nuanced than a simple unidimensional construct. One self-control framework that does just this and has risen to prominence in recent years is the dual systems model. The dual systems model is comprised of two psychological constructs related to self-control: sensation-seeking and impulse control. Sensation-seeking refers to the degree to which an individual has the desire to seek out novel and thrilling experiences, whereas impulse control refers to one’s capacity to delay gratification and halt engagement in behaviors and consider the consequences before further action (Steinberg et al., 2008; Steinberg, 2010). The dual systems model has been found to be useful for predicting engagement in a range of risky behaviors (Ellingson et al., 2019; Rhyner et al., 2018; Shulman et al., 2016).

Much of the research on this topic focuses on differential development of these constructs as they relate to peaks in adolescent engagement in risky behavior (Burt et al., 2014; Forrest et al., 2019). However, some research has also indicated the relevance of this model for understanding victimization risk as well. Connolly et al. (2020) found that developmental changes in impulsivity during adolescence and early adulthood resulted in changes in victimization risk during this time, but this relationship was not observed for changes in sensation-seeking. While the highlighted study provided a first step in understanding how these constructs may predict victimization risk overall, there remain unanswered questions in this domain. This study did not examine ways in which victimization may be a driver of variation in these constructs that results in increased later victimization risk.

So, while general development of sensation-seeking did not predict changes in victimization risk, examining heterogeneity in the variation of this construct that is driven by trauma may result in survivors of victimization to have high risk of victimization later in life. Such a relationship would make sense, as individuals with a stronger predilection for situations that may place them at risk for victimization would almost by definition have greater risk for being victimized. Indeed, past research has indicated that individuals high in sensation-seeking may be at greater risk for victimization (Averdijk et al., 2019; Monks et al., 2010). Research has also indicated that survivors of trauma may demonstrate elevated sensation-seeking (Bornovalova et al., 2008; Brady & Donenberg, 2006). However, the alternative must also be considered. Research on traumatic stress has indicated that survivors may also exhibit avoidance behaviors (American Psychiatric Association, 2013; Dulin & Passmore, 2010). Such behaviors may lead individuals to actively attempt to avoid retraumatization through mitigating exposure to risky situations. For example, Dulin and Passmore (2010) found that the desire to avoid traumatic situations mediated the relationship between lifetime traumatic stress exposure and mental health symptoms. These competing strands of research indicate the need for greater examination of these relationships through mediation analysis suggested here.

A similar relationship may exist for diminished impulse control that is driven by trauma exposure. Individuals with low impulse control may be more likely to respond to provocations with physical violence which may result in them being the victim of violent responses themselves (Berg & Felson, 2016; Felson et al., 2018). Further research indicates that low impulse control may place individuals at greater risk for victimization in both online and off-line contexts also (Nedelec, 2018). This may be due to the fact that individuals lacking in impulse control may not stop to consider consequences of their actions before placing themselves in risky situations that may leave vulnerable to experiencing victimization (e.g., doing drugs in an unknown venue with strangers may make individuals vulnerable for victimization). (Steinberg, 2010). Further, past research has also indicated that survivors of trauma demonstrate diminished impulse control (Davis et al., 2017; Monahan et al., 2015). It is here that a potential mediating relationships between past accumulated victimization, impulse control, and future victimization may be present. Individuals who experience victimization earlier in life may demonstrate diminished impulse control and this diminished impulse control may lead them to lack the capacity to stop and consider the consequences of their actions that may put them in situations that place them at-risk for future victimization. If previous victimization is a driver of variation in the dual systems constructs as has been indicated above, then it seems logical that greater accumulation of victimization experiences may cause greater dysfunction. If this is the case, then those individuals who have experienced greater cumulative victimization in childhood or adolescence should have the lowest levels of impulse control and/or highest levels of sensation-seeking and thus should have the greater risk for experiencing victimization later in life. In this way, these constructs may mediate the cumulative victimization–later victimization relationship. While these psychological constructs present one means of understanding this relationship, social factors may also play a role here as well.

While the dual systems constructs provide psychological mechanisms for understanding why individuals who suffer from increasing levels of cumulative victimization may be at greater risk for revictimization, the social construct of deviant peer association may also help to explain this relationship. Deviant peer association refers the general degree to which an individual’s social relationships are comprised of individuals who engage in delinquent, criminal, or otherwise deviant behavior. Past research has indicated that greater deviant peer association is associated with increased victimization risk (Hong et al., 2017; Rudolph et al., 2020). This is likely due to a number of reasons. First, spending time with individuals engaging in criminal activity may place one in situations where they are more likely to encounter violence. Indeed, past research indicates that affiliation with deviant peers is associated with increased involvement in crime and the victim–offender overlap is well-documented (Brezina & Azimi, 2018; Reingle, 2014; Wojciechowski, 2018). Second, and relatedly, criminal peers may be more likely to be violent themselves, and disagreements between friends could be more likely to escalate into violent situations. Alternatively, peer victimization has been found to predicted subsequent increases in deviant peer association (Rudolph et al., 2014). Past research has also indicated that such rejection by peers may also result in subsequent increases in association with deviant peers (Chen et al., 2015; Laird et al., 2001). Relatedly, there is also a great deal of conceptual overlap and high correlation between peer rejection and peer victimization (Hanish & Guerra, 2000). This is consistent with Moffit’s (1993) dual taxonomy of offending which predicts that youth may be rejected due to aggressive and antisocial behaviors, and these individuals may then coalesce into peer groups. In this way, suffering victimization early in life may lead individuals to sort into peer groups which place them at increased risk for later victimization, thus, acting as a social mediator. Much like psychological constructs, it would make sense that greater accumulation of victimization would then impact these constructs to a greater extent also.

Figure 1 describes the multiple mediation model hypothesized. Despite past research indicating that the identified social and psychological constructs may mediate the cumulative victimization–later victimization relationship, these processes have remained understudied. Identifying if one or more of these constructs mediates this relationship would allow for targeting of treatment. Alternatively, if certain constructs do not significantly mediate this relationship, then this may guide diversion of program funding to efforts associated with significant mediators. This study sought to address these gaps in the extant literature by testing the following hypotheses:

Hypothesis 1: Increased cumulative victimization prior to baseline will be associated with increased risk for experiencing victimization at follow-up.

Hypothesis 2: Sensation-seeking, impulse control, and deviant peer association will independently and significantly mediate the cumulative victimization-later victimization relationship.

Figure 1.
Multiple mediation model.

Method

The present study utilized data from the first three waves of the Pathways to Desistance study. In its entirety, this data set was comprised of the responses of 1,354 juvenile offenders followed for 84 months after an adjudication for a serious offenses. Qualifying adjudications consisted of all felony charges, as well as misdemeanor weapons-related charges and sexual assault. These adjudications had to be for offenses committed when participants were between the ages 14 and 17 years, with all participants being between the ages 14 and 19 years at baseline measurements. Participants were recruited from study sites in Maricopa County, Arizona and Philadelphia, Pennsylvania, with recruitment occurring from 2000 to 2003. The entire study period spanned 2000–2010. Of all of the qualified juvenile offenders approached regarding their interest in participation, 20% declined the opportunity. A cap on the overall proportion of male drug offenders included in the sample was applied at 15% in order to ensure heterogeneity in baseline characteristics. The peak rate of attrition was observed at the final wave of data at 16.2%, with the peak attrition rate of data used in the present study observed at the third wave of measurement at 7.5%. These waves of data were specifically chosen for analyses because they had the lowest amount of missing data in the entire data set.

Data for the present study were collected via participant self-report. The research team provided participants with laptop computers during interview sessions. Members of the research team administered verbal prompts to participants and participants then manually entered responses into the laptop computer. This was done to maximize confidentiality and honesty in reporting. Interviews were conducted in locations that were convenient for participants (e.g., libraries, participants’ homes, criminal justice facilities).

Measures

Victimization

Victimization experiences were examined as both the dependent and independent variables. The dependent variable examined was a binary measure, delineating participants who reported being victimized during this same time period from those who did not (0 = no; 1 = yes).²

Cumulative victimization

The independent cumulative victimization variable was measured as a count of the number of different types of direct victimization that participants reported experiencing during their lifetimes prior to baseline. The following items were used to measure victimization experiences: Have you been chased where you thought you might be seriously hurt? Have you been beaten up, mugged, or seriously threatened by another person? Have you been raped, had someone attempt to rape you or been sexually attacked in some other way? Have you been attacked with a weapon, like a knife, box cutter, or bat? Have you been shot at? Have you been shot?

Impulse control

One of the other independent variables utilized in the present study was impulse control at Wave 2. This construct was measured using the Weinberger Adjustment Inventory (Weinberger & Schwartz, 1990). This scale utilized a series of statements which participants responded to indicating the degree to which the statement was true of their own behavior or attitudes (e.g., I say the first thing that comes into my mind without thinking enough about it). Participants responded using a five-point ordinal scale (1 = false; 5 = true). Seven of the eight items were reverse coded so that higher scores corresponded with higher impulse control. A mean score of these individual items was computed so that each participant had a single impulse control score. The following items were used to measure impulse control: I’m the kind of person who will try anything once, even if it’s not that safe; I should try harder to control myself when I’m having fun; I do things without giving them enough thought; I become “wild and crazy” and do things other people might not like; When I’m doing something for fun (e.g., partying, acting silly), I tend to get carried away and go too far; I like to do new and different things that many people would consider weird or not that safe; I say the first thing that comes to my mind without thinking enough about it; I stop and think things through before I act

Sensation-seeking

Another independent variable utilized in analyses was sensation-seeking at Wave 2. This construct was measured using the Youth Psychopathic Traits Inventory (Andershed et al., 2002). This scale consisted of a series of five statements which participants rated the degree to which each applied to their own behaviors/attitudes using a 4-point ordinal scale (1 = does not apply at all; 4 = applies very well). Scores on each individual item were added together to form an index ranging from 5 to 20. The following items were used to measure sensation-seeking: I like to be where exciting things happen, I get bored quickly when there is too little change, I like to do things just for the thrill of it, and I get bored quickly be doing the same thing over and over, I like to do exciting and dangerous things, even if it is forbidden or illegal

Deviant peer association

The final independent variable included in analyses was deviant peer association at Wave 2. This construct was measured using a series of individual ordinal items assessing the general number of friends who participants reported attempted to influence them to engage in seven different antisocial acts, with higher scores indicating more peers (1 = none of them; 5 = all of them). A mean score was computed from these individual items so that each participant had a single deviant peer association score at Wave 2. The following items were used to measure deviant peer association: How many of your friends have suggested that you should go out drinking with them? How many of your friends have suggested or claimed that you have to get drunk to have a good time? How many of your friends have suggested or claimed that you have to be high on drugs to have a good time? How many of your friends have suggested that you should sell drugs? How many of your friends have suggested that you should steal something? How many of your friends have suggested that you should hit or beat someone up? How many of your friends have suggested that you should carry a weapon?

Control variables

Several additional variables were included in analyses in order to control for bias in estimation. The first of these variables was gender. This is because past research has indicated that risk for victimization among juvenile offenders is stratified by gender and may also differ by type of victimization (Baglivio et al., 2014). Gender was measured at baseline as a binary variable which delineated male and female participants (0 = male; 1 = female).

Another control variable included in analyses was race. This is because past research has indicated that the impact of trauma on outcomes of interest among juvenile offenders may differ by race, with Black juvenile offenders demonstrating exacerbated effects (Johnson, 2018). Race was measured at baseline as a nominal variable which delineated participants into one of four race categories: Black, Hispanic, White, and Other race. A series of dummy variables was computed from this original variable, with each delineating participants in one race category from all other participants (e.g., 1 = Black; 0 = all other participants). The dummy variable corresponding to White participants was omitted from analyses in order to provide a reference category.

Lower socioeconomic status (SES) has also been identified as a risk factor related to victimization among juvenile offenders (Loeber et al., 2001), necessitating controlling for this concept in analyses. SES was measured at baseline as a weighted score comprised of a combination of the occupational prestige and educational attainment of parents of participants. If both parents were available to provide data, then a mean score was computed so that each participant was provided with a single SES score.

Another variable included in analyses was age, as research indicates that risk for victimization tends to peak in early adolescence and decline thereafter among juvenile offenders (Pereda et al., 2017). Age was measured as an interval variable, and the Wave 3 measurement of this construct was included in analyses.

While all observation periods for variables included in analyses generally spanned 6 months, there remained some variance in the exact length for each participant. For this reason, the number of days in the Wave 3 observation period was controlled for each participant, as longer observation periods would necessarily mean greater risk for victimization simply because of additional exposure time.

The final control variable included in analyses was the proportion of the Wave 3 observation period that participants spent in a secure facility with no community access. This mainly referred to criminal justice settings, like jail, but could also refer to places like inpatient psychiatric facilities. This was controlled for because having less community access could impact the risk that individuals would be exposed to victimization. This variable was operationalized as a proportion ranging from “0” to “1,” with scores of “0” indicating no time spent in secured facilities and scores of “1” indicating 100% of the wave three observation period spent in a secured facility.

Analytic Strategy

The analyses for the present study proceeded in two phases. The first phase of analyses utilized generalized structural equation modeling to estimate direct effects of victimization experienced prior to baseline on revictimization risk and test for mediating effects of cognitive and social variables. This method was chosen because of its capacity to test for mediating effects and conduct later decomposition analysis allowing for understanding of whether specific indirect effects and the total indirect effect were significant in explaining how earlier victimization predicted later revictimization risk. Generalized structural equation modeling was chosen here because of the binary nature of the dependent variable which necessitated the use of logistic regression to estimate effects. Coefficients described the predicted change in the probability of experiencing victimization at wave three based on a one-unit change in a given independent variable, net of all covariates.

The second phase of analyses entailed testing to determine whether or not either, both or the total mediation effects were significant. While this can be completed in a straightforward manner, oftentimes standard errors for obtained coefficients are not normally distributed. To avoid bias in estimation, the Preacher and Hayes (2008) method for computing bootstrapped standard errors was carried out with 500 repetitions. This provided indication of whether the specific indirect effects and/or the total indirect effect significantly mediated the direct relationship of interest.

Results

Table 1 provides descriptive statistics for variables included in analyses. Table 2 provides a correlation matrix comprising all variables included in analyses. Table 3 provides Model 1 estimates of direct effects of prebaseline victimization accumulation scores on odds of victimization, whereas Table 4 provides Model 2 estimates with inclusion of mediating variables. Table 5 provides specific and total indirect effects for these models with bootstrapped standard errors.

Table 1.
Descriptive Statistics.

Variables Mean/Proportion Standard Deviation Minimum Maximum

Victimization at Wave 3 (0 = no; 1 = yes) 0.140 NA 0 1

Cumulative victimization score at baseline 1.575 1.457 0 6

Sensation-seeking at Wave 2 13.393 3.350 4 20

Impulse control at Wave 2 3.071 0.941 1 5

Deviant peer association at Wave 2 0.963 .875 0 4

Gender (0 = male; 1 = female) 0.136 NA 0 1

Race

Black 0.414 NA 0 1

Hispanic 0.335 NA 0 1

White 0.202 NA 0 1

Other race 0.048 NA 0 1

Socioeconomic status 51.409 12.299 11 77

Age at Wave 3 17.048 1.154 15 20

Proportion of Wave 3 spent in secure facilities 0.426 0.432 0 1

Wave 3 observation period length 175.192 30.277 61 245

Table 2.
Correlation Matrix.

Variables 1 2 3 4 5 6 7 8 9 10 11 12 13

1. Victimization (Wave 3) — .181* −.062* .070* .197* −.069* −.070* .039 .008 .017 −.020 .029 −.060*

2. Baseline Cumulative Victimization — — −.231* .235* .287* .118* −.119* .080* .030 −.027 .156* −.025 .164*

3. Impulse control — — — −.481* −.275* .048 .173* −063* −.035 −.025 .054 −.014 −.102*

4. Sensation-seeking — — — — .270* −.025 −.071* −.028 .041 −.059 .023 −.007 .053

5. Deviant peer association — — — — — −.147* −.056 .078* .012 .074* .044 −.035 .203*

6. Gender (0 = male; 1 = female) — — — — — — −.036 −.026 .022 −.051 −.017 −.016 −.196*

7. Black ^a — — — — — — — — — −.114* .022 −.025 .148*

8. Hispanic — — — — — — — — — .321* −.012 .010 −.063*

9. Other Race — — — — — — — — — −.037 .046 .025 .015

10. SES — — — — — — — — — — −.058* .036 .036

11. Age — — — — — — — — — — — −.010 .030

12. Observation Period Length — — — — — — — — — — — — −.035

13. Time Spent in Secured Facilities — — — — — — — — — — — — —

^a Race variables here refer to the dummy variables included in analyses that are described in the “Measures” section. Correlations between the race variables are not assessed because they’re inherent collinearity.

p ≤ .05.

Table 3.
Generalized Structural Equation Modeling of Covariate Effects on Wave 3 Victimization Risk: Model 1.

Variables Coefficient p Value 95% Confidence Interval

Cumulative victimization score at baseline .369 <.001 [0.256, 0.483]

Gender (0 = male; 1 = female) −.663 .027 [−1.125, −0.074]

Race

Black −.227 .322 [−0.675, 0.222]

Hispanic −.110 .637 [−0.568, 0.348]

Other race .022 .958 [−0.788, 0.832]

Socioeconomic status .006 .442 [−0.009, .020]

Age at Wave 3 −.111 .131 [−0.256, 0.033]

Proportion of Wave 3 spent in secure facilities −.682 .001 [−1.092, −0.272]

Wave 3 observation period length .003 .286 [−0.003, 0.008]

Constant −.280 .848 [−3.136, 2.576]

Table 4.
Generalized Structural Equation Modeling of Covariate Effects on Wave 3 Victimization Risk: Model 2.

Variables Coefficient p Value 95% Confidence Interval

Cumulative victimization score at baseline .307 <.001 [0.169, 0.444]

Sensation-seeking at Wave 2 <.001 .997 [−0.069, 0.069]

Impulse control at Wave 2 .071 .564 [−0.171, 0.313]

Deviant peer association at Wave 2 .594 <.001 [0.369, 0.819]

Gender (0 = male; 1 = female*) −.555 .107 [−1.229, 0.119]

Race

Black −.105 .698 [−0.637, 0.426]

Hispanic −.141 .615 [−0.690, 0.408]

Other race .001 .998 [−0.940, 0.943]

Socioeconomic status .003 .697 [−0.014, 0.020]

Age at Wave 3 −.112 .199 [−0.282, 0.059]

Proportion of Wave 3 spent in secure facilities −.951 <.001 [−1.436, −0.466]

Wave 3 observation period length .001 .731 [−0.005, 0.007]

Constant −.668 .721 [−4.327, 2.991]

Cumulative victimization→sensation−seeking .545 <.001 [0.410, 0.680]

Cumulative victimization→impulse control −.150 <.001 [−0.185, −0.115]

Cumulative victimization→deviant peer association .172 <.001 [0.140, 0.205]

Table 5.
Indirect Effects of Cumulative Victimization on Later Victimization Risk Operating Through Hypothesized Mediators (Bootstrap Repetitions = 500).

Indirect Effects Coefficient Bootstrapped Standard Error p Value 95% Confidence Interval

Sensation-seeking indirect effect >−.001 .002 .931 [−.004, .004]

Impulse control indirect effect −.001 .002 .636 [−.005, .003]

Deviant peer association indirect effect .013 .003 <.001 [.007, .019]

Total indirect effect .012 .003 <.001 [.006, .018]

Model 1 results indicated that greater victimization accumulation prior to baseline was associated with increased odds of experiencing revictimization during the Wave 3 observation period (coefficient = .369, p ≤ .001). Being female and spending more time in secured facilities were associated with lower risk of victimization at Wave 3 in this model. Model 2 results indicated that inclusion of mediating variables attenuated the direct effect of cumulative victimization on victimization risk at Wave 3 by about 20% (coefficient = .307, p ≤ .001). Greater levels of deviant peer association at Wave 2 also predicted greater victimization risk at wave three in this model (coefficient = .594, p ≤ .001). Neither sensation-seeking nor impulse control was associated with Wave 3 victimization risk. However, the paths running from pre-baseline victimization accumulation to sensation-seeking, impulse control, and deviant peer association were all indeed significant (cumulative victimization→sensation-seeking coefficient = .545, p ≤ .001; cumulative victimization→impulse control coefficient = −.150, p ≤ .001; cumulative victimization→deviant peer association coefficient = .172, p ≤ .001). Spending more days during the Wave 3 observation period in a secured facility was associated with lower victimization risk in this model.

The second phase of analyses entailed the computation of bootstrapped standard errors to test whether any of the specific indirect effects or the total indirect effect was significant using the Preacher and Hayes (2008) method. Results indicated that the specific indirect effect running through deviant peer association was indeed significant (coefficient = .013, p ≤ .001). Neither specific indirect effect running through impulse control nor sensation-seeking were significant. The total indirect effect was also found to be significant (coefficient = .012, p ≤ .001). Based on these obtained coefficients, the significance of the total indirect effect appears to be accounted for entirely by deviant peer association.³ This is because deviant peer association was the only significant mediator and accounted for a net of over 99% of the .012 total indirect effect. The discrepancy between the total effect being less than the deviant peer association effect was because of the ∼−.001 coefficient for impulse control that washed out part of the positive portion of this effect. This instrument was then reverse coded so that the indirect effects were all in the same direct and the effect was then .001. The .002 combined effects of sensation-seeking and impulse control accounted for less than 1% of the total indirect effect of .013 rounded to the thousandth. As such, it can be stated that more than 99% of the total indirect effect was accounted for by deviant peer association.

Discussion

The present study sought to provide more complete understanding of the relationship between cumulative victimization and later victimization risk. Results indicated that experiencing increased cumulative victimization prior to baseline predicted greater victimization risk at follow-up. While multiple psychological and sociological mediators were tested, only increased deviant peer association was found to significantly mediate this relationship. This mediation accounted for nearly 20% of this direct effect. There are numerous theoretical implications of these findings, as well as implications for professionals involved with providing treatment and services for victims.

Past research has indicated that greater levels of association with criminal peers may put an individual at risk for victimization (Hong et al., 2017; Rudolph et al., 2020). This study established that experiencing greater victimization accumulation also increases later association with deviant peers. It is here that further work is necessary. Understanding why it is that greater victimization accumulation results in greater association with deviant peers may aid in reducing risk for later victimization. One potential explanation may lie in the constructs that were used in this study already. Past research has indicated that trauma exposure is linked to disruptions in cognitive development, including that of sensation-seeking and impulse control (Efrati & Gola, 2019; Morris et al., 2020). The findings of the present study provided additional support for these relationships, as trauma exposure prior to baseline was associated with lower impulse control and higher sensation-seeking. One prominent postulate in the field of criminology is the “birds of a feather” hypothesis. This hypothesis predicts that the relationship between deviant peer association and offending is not due to socialization effects, as predicted by Akers’ (1973) social learning theory but instead is due to offenders coalescing around one another due to shared interests and characteristics. In this manner, it takes the selection side of the selection versus socialization debate. In terms of the relevance for victimization, it may be that those individuals high in sensation-seeking and/or low in impulse control may come together to form peer groups. Because both of these cognitive constructs are risk factors for offending (Armstrong et al., 2020; Forrest et al., 2019), this may lead to the selection of individuals with high sensation-seeking and/or low impulse control to coalesce around other deviant peers. With this research indicating that deviant peer association mediates the victimization–revictimization relationship, this may provide another link in the mediating chain. Increased cumulative victimization in early life may result in high levels of sensation-seeking and/or low impulse control which results in greater association with deviant peers later, which subsequently increases risk for revictimization. This would help to clarify the lack of mediating effects observed in this study. However, this remains speculative and is beyond the scope of the present research. Future studies should seek to clarify this mediating relationship using longitudinal data with at least four waves (Wave 1: Victimization→Wave 2: Cognitive Constructs→Wave 3: Deviant Peer Association→Wave 4: Re-victimization) to establish temporal ordering for these causal relationships.

While there are relevant theoretical implications and avenues for future research to explore, there are also potential clinical implications of these findings with applicability for victims’ services and treatment. These findings indicate that a focus on the social ties of victims is relevant for decreasing the risk of future harm being perpetrated upon those who have suffered cumulative victimization. Screening victims’ social networks at intake of treatment may provide important information for ensuring safety. Education regarding the potential dangers that individuals may face for repeat victimization due to those in their social networks may help individuals may informed decisions about potentially toxic relationships with criminal peers that could result in their future victimization. Adapting screening and intervention like this into existing victims’ services and treatment could aid in reducing the odds of future damage being done to individuals attempting to recover from traumatic experiences.

It should be noted that there were no significant gender nor race differences in victimization risk in the final model. The only significant difference was the finding that female participants reported elevated victimization risk at follow-up relative to male participants in Model 1. Considering that null findings were observed upon inclusion of hypothesized mediators, it may be that gender differences in these constructs helped account for this. Future research should seek to reexamine these processes with gender moderation in the future to determine the relevance of differences in victimization risk among male and female juvenile offenders.

One final consideration that should be made here is the relatively strong relationship between time spent in secured facilities and revictimization risk. Considering that greater deviant peer association and less time spent in secured facilities were both associated with greater revictimization risk, it may be that further analyses may yield a relationship between these three constructs. It may be that the engaging with more peers involved in antisocial behavior is a function of more time spent in the community and that spending more time with deviant peers is the reason why revictimization risk is higher. While further examination of these relationships was beyond the scope of this study, this does provide an area for researchers to examine in greater detail in the future.

While the present study provided a unique examination of the victimization–revictimization relationship, there remain several relevant limitations. First, the generalizability of these results is somewhat questionable, as this data set was comprised solely of adjudicated youth. These individuals are at greater risk for suffering exposure to violence, and it seems likely that the indicated nature of this sample may lead them to differ substantially from their peers in the general population (Baglivio et al., 2014; Kilpatrick et al., 2003). As such, extrapolation of results beyond this population is limited. Another limitation of the present study relates to the generality of the victimization measures utilized in analyses. While understanding how the cumulative victimization–revictimization relationship functions in a general sense is useful, the available measures lacked the capacity for parsing out the separate types of victimization. It may be that different forms of victimization influence revictimization risk to different degrees. For example, physical victimization may influence risk for future physical victimization, but not for future sexual victimization. The publicly available Pathways to Desistance data do not allow for a breakdown of separate victimization types, so these analyses are beyond the scope of the present study. Future research should seek to better understand the more specific aspects of these processes by leveraging data that allows for such examinations. A related limitation pertains to the fact that the cumulative victimization measure utilized in analyses comprised only experiences of direct physical and sexual victimization. Other research has also included other forms of vicarious trauma, like witnessed community violence (Finkelhor et al., 2005; Hamby et al., 2004), for assessing related research questions. This choice was made because of a desire to take a narrower definition of the term victimization and eschew more general forms of trauma, as the Finkelhor et al. (2005) study does with the concept of polyvictimizatio n , but this does remain as a limitation. While the present study chose to focus on more direct forms of victimization, sensitivity analyses were conducted that included additional measures of witnessed violence in the cumulative victimization score. Results were essentially analogous to those observed in the main analyses, providing robustness for these findings. It should also be noted again that the waves selected for analyses were based on the low amount of missing data and the short follow-up time period may have influenced results to some extent. Future research should investigate these relationships using a long follow-up period. It should also be noted that the indirect coefficient was very small, even if it was significant, drawing questions regarding its overall meaningfulness. It should also be noted that this study did not examine mental health variables as potentially relevant mechanisms despite the relationship between traumatic stress exposure and mental health issues (American Psychiatric Association, 2013). A final limitation pertains to the fact that mediation variables measured prior to baseline were not included in analyses. It may be that underlying levels of these mediators drove these findings, rather than their levels at wave two (ex: those who had more deviant peers at wave two already had more deviant peers prior to baseline). Unfortunately, the cumulative victimization measure used in analyses was a lifetime measure, so it was not possible to parse temporal ordering for the mediating variables prior to baseline. As such, their control likely would not have addressed this issue sufficiently. Future research should seek to reexamine these processes using novel data which is better able to address this problem.

The findings of the present study provide novel understanding of the relationship between cumulative victimization and revictimization risk. More severe victimization accumulation prior to baseline was associated with increased victimization risk later. Further, this relationship was significantly mediated by variation in deviant peer association but not psychological constructs. This provides indication that screening of individuals seeking victims’ services regarding their social relationships may provide useful information regarding potential for future retraumatization. Further, the results of this study lay out avenues for future research in not only testing the robustness of findings using more generalizable data and data that allows for greater specificity in analyses but also in examination the potential for longer chains of mediation. It may be that variation in dual systems model constructs resulting from trauma exposure helps to explain the variation in deviant peer association that drives future victimization risk. Future research should seek to study these relationships to better understand the processes that drive revictimization risk throughout the life course.

Variables	Mean/Proportion	Standard Deviation	Minimum	Maximum
Victimization at Wave 3 (0 = no; 1 = yes)	0.140	NA	0	1
Cumulative victimization score at baseline	1.575	1.457	0	6
Sensation-seeking at Wave 2	13.393	3.350	4	20
Impulse control at Wave 2	3.071	0.941	1	5
Deviant peer association at Wave 2	0.963	.875	0	4
Gender (0 = male; 1 = female)	0.136	NA	0	1
Race
Black	0.414	NA	0	1
Hispanic	0.335	NA	0	1
White	0.202	NA	0	1
Other race	0.048	NA	0	1
Socioeconomic status	51.409	12.299	11	77
Age at Wave 3	17.048	1.154	15	20
Proportion of Wave 3 spent in secure facilities	0.426	0.432	0	1
Wave 3 observation period length	175.192	30.277	61	245

Variables	1	2	3	4	5	6	7	8	9	10	11	12	13
1. Victimization (Wave 3)	—	.181*	−.062*	.070*	.197*	−.069*	−.070*	.039	.008	.017	−.020	.029	−.060*
2. Baseline Cumulative Victimization	—	—	−.231*	.235*	.287*	.118*	−.119*	.080*	.030	−.027	.156*	−.025	.164*
3. Impulse control	—	—	—	−.481*	−.275*	.048	.173*	−063*	−.035	−.025	.054	−.014	−.102*
4. Sensation-seeking	—	—	—	—	.270*	−.025	−.071*	−.028	.041	−.059	.023	−.007	.053
5. Deviant peer association	—	—	—	—	—	−.147*	−.056	.078*	.012	.074*	.044	−.035	.203*
6. Gender (0 = male; 1 = female)	—	—	—	—	—	—	−.036	−.026	.022	−.051	−.017	−.016	−.196*
7. Black ^a	—	—	—	—	—	—	—	—	—	−.114*	.022	−.025	.148*
8. Hispanic	—	—	—	—	—	—	—	—	—	.321*	−.012	.010	−.063*
9. Other Race	—	—	—	—	—	—	—	—	—	−.037	.046	.025	.015
10. SES	—	—	—	—	—	—	—	—	—	—	−.058*	.036	.036
11. Age	—	—	—	—	—	—	—	—	—	—	—	−.010	.030
12. Observation Period Length	—	—	—	—	—	—	—	—	—	—	—	—	−.035
13. Time Spent in Secured Facilities	—	—	—	—	—	—	—	—	—	—	—	—	—

Variables	Coefficient	p Value	95% Confidence Interval
Cumulative victimization score at baseline	.369	<.001	[0.256, 0.483]
Gender (0 = male; 1 = female)	−.663	.027	[−1.125, −0.074]
Race
Black	−.227	.322	[−0.675, 0.222]
Hispanic	−.110	.637	[−0.568, 0.348]
Other race	.022	.958	[−0.788, 0.832]
Socioeconomic status	.006	.442	[−0.009, .020]
Age at Wave 3	−.111	.131	[−0.256, 0.033]
Proportion of Wave 3 spent in secure facilities	−.682	.001	[−1.092, −0.272]
Wave 3 observation period length	.003	.286	[−0.003, 0.008]
Constant	−.280	.848	[−3.136, 2.576]

Variables	Coefficient	p Value	95% Confidence Interval
Cumulative victimization score at baseline	.307	<.001	[0.169, 0.444]
Sensation-seeking at Wave 2	<.001	.997	[−0.069, 0.069]
Impulse control at Wave 2	.071	.564	[−0.171, 0.313]
Deviant peer association at Wave 2	.594	<.001	[0.369, 0.819]
Gender (0 = male; 1 = female)	−.555	.107	[−1.229, 0.119]
Race
Black	−.105	.698	[−0.637, 0.426]
Hispanic	−.141	.615	[−0.690, 0.408]
Other race	.001	.998	[−0.940, 0.943]
Socioeconomic status	.003	.697	[−0.014, 0.020]
Age at Wave 3	−.112	.199	[−0.282, 0.059]
Proportion of Wave 3 spent in secure facilities	−.951	<.001	[−1.436, −0.466]
Wave 3 observation period length	.001	.731	[−0.005, 0.007]
Constant	−.668	.721	[−4.327, 2.991]
Cumulative victimization→sensation−seeking	.545	<.001	[0.410, 0.680]
Cumulative victimization→impulse control	−.150	<.001	[−0.185, −0.115]
Cumulative victimization→deviant peer association	.172	<.001	[0.140, 0.205]

Indirect Effects	Coefficient	Bootstrapped Standard Error	p Value	95% Confidence Interval
Sensation-seeking indirect effect	>−.001	.002	.931	[−.004, .004]
Impulse control indirect effect	−.001	.002	.636	[−.005, .003]
Deviant peer association indirect effect	.013	.003	<.001	[.007, .019]
Total indirect effect	.012	.003	<.001	[.006, .018]

Footnotes

Declaration of Conflicting Interests

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

Funding

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

ORCID iD

Thomas Wojciechowski

Notes

References

Akers

R. L.

(1973). Deviant behavior: A social learning approach. Wadsworth Publishing Company.

American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders: DSM-5. American Psychiatric Association.

Andershed

Kerr

Stattin

Levander

(2002). Psychopathic traits in non-referred youths: A new assessment tool. In Blauuw

Sheridan

(Eds.), Psychopaths: Current international perspectives (pp. 131–158). Elsevier.

Armstrong

T. A.

Boisvert

Wells

Lewis

(2020). Extending Steinberg’s adolescent model of risk taking to the explanation of crime and delinquency: Are impulsivity and sensation seeking enough? Personality and Individual Differences, 165, 110133.

Averdijk

Ribeaud

Eisner

(2019). Childhood predictors of violent victimization at age 17 years: The role of early social behavioral tendencies. The Journal of Pediatrics, 208, 183–190.

Baglivio

M. T.

Epps

Swartz

Huq

M. S.

Sheer

Hardt

N. S.

(2014). The prevalence of adverse childhood experiences (ACE) in the lives of juvenile offenders. Journal of Juvenile Justice, 3(2), 1–17.

Barnes

J. E.

Noll

J. G.

Putnam

F. W.

Trickett

P. K.

(2009). Sexual and physical revictimization among victims of severe childhood sexual abuse. Child Abuse & Neglect, 33(7), 412–420.

Berg

M. T.

Felson

R. B.

(2016). Why are offenders victimized so often. In Rennison

C. M.

Cuevas

C. A.

(Eds.), The Wiley handbook on the psychology of violence (pp. 49–65). Wiley.

Bornovalova

M. A.

Gwadz

M. A.

Kahler

Aklin

W. M.

Lejuez

C. W.

(2008). Sensation seeking and risk-taking propensity as mediators in the relationship between childhood abuse and HIV-related risk behavior. Child Abuse & Neglect, 32(1), 99–109.

10.

Brady

S. S.

Donenberg

G. R.

(2006). Mechanisms linking violence exposure to health risk behavior in adolescence: Motivation to cope and sensation seeking. Journal of the American Academy of Child & Adolescent Psychiatry, 45(6), 673–680.

11.

Brezina

Azimi

A. M.

(2018). Social support, loyalty to delinquent peers, and offending: An elaboration and test of the differential social support hypothesis. Deviant Behavior, 39(5), 648–663.

12.

Burt

C. H.

Sweeten

Simons

R. L.

(2014). Self-control through emerging adulthood: Instability, multidimensionality, and criminological significance. Criminology, 52(3), 450–487.

13.

Chen

Drabick

D. A.

Burgers

D. E.

(2015). A developmental perspective on peer rejection, deviant peer affiliation, and conduct problems among youth. Child Psychiatry & Human Development, 46(6), 823–838.

14.

Cole

Logan

T. K.

Shannon

(2008). Women’s risk for revictimization by a new abusive partner: For what should we be looking? Violence and Victims, 23(3), 315–330.

15.

Connolly

E. J.

Cooke

E. M.

Beaver

K. M.

Brown

(2020). Do developmental changes in impulsivity and sensation seeking uniquely predict violent victimization? A test of the dual systems model. Journal of Criminal Justice, 66, 101639.

16.

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.

17.

Dierkhising

C. B.

Ford

J. D.

Branson

Grasso

D. J.

Lee

(2019). Developmental timing of polyvictimization: Continuity, change, and association with adverse outcomes in adolescence. Child Abuse & Neglect, 87, 40–50.

18.

Dulin

P. L.

Passmore

(2010). Avoidance of potentially traumatic stimuli mediates the relationship between accumulated lifetime trauma and late-life depression and anxiety. Journal of Traumatic Stress: Official Publication of the International Society for Traumatic Stress Studies, 23(2), 296–299.

19.

Edalati

Krausz

Schütz

C. G.

(2016). Childhood maltreatment and revictimization in a homeless population. Journal of Interpersonal Violence, 31(14), 2492–2512.

20.

Efrati

Gola

(2019). The effect of early life trauma on compulsive sexual behavior among members of a 12-step group. The Journal of Sexual Medicine, 16(6), 803–811.

21.

Ellingson

J. M.

Corley

Hewitt

J. K.

Friedman

N. P.

(2019). A prospective study of alcohol involvement and the dual-systems model of adolescent risk-taking during late adolescence and emerging adulthood. Addiction, 114(4), 653–661.

22.

Evans

C. B.

Smokowski

P. R.

Cotter

K. L.

(2014). Cumulative bullying victimization: An investigation of the dose–response relationship between victimization and the associated mental health outcomes, social supports, and school experiences of rural adolescents. Children and Youth Services Review, 44, 256–264.

23.

Felson

R. B.

Berg

M. T.

Rogers

E. M.

Krajewski

(2018). Disputatiousness and the offender–victim overlap. Journal of Research in Crime and Delinquency, 55(3), 351–389.

24.

Finkelhor

Ormrod

R. K.

Turner

H. A.

Hamby

S. L.

(2005). Measuring poly-victimization using the Juvenile Victimization Questionnaire. Child Abuse & Neglect, 29(11), 1297–1312.

25.

Ford

J. D.

Elhai

J. D.

Connor

D. F.

Frueh

B. C.

(2010). Poly-victimization and risk of posttraumatic, depressive, and substance use disorders and involvement in delinquency in a national sample of adolescents. Journal of Adolescent Health, 46(6), 545–552.

26.

Forrest

Hay

Widdowson

A. O.

Rocque

(2019). Development of impulsivity and risk-seeking: Implications for the dimensionality and stability of self-control. Criminology, 57(3), 512–543.

27.

Golder

Logan

T. K.

(2011). Cumulative victimization, psychological distress, and high-risk behavior among substance-involved women. Violence and Victims, 26(4), 477–495.

28.

Grasso

D. J.

Greene

Ford

J. D.

(2013). Cumulative trauma in childhood. In Ford

J. D.

Courtois

C. A.

(Eds.), Treating complex traumatic stress disorders in children and adolescents: An evidence based guide (pp. 79–99). Guilford.

29.

Hamby

S. L.

Finkelhor

Ormrod

R. K.

Turner

H. A

. (2004). The Juvenile Victimization Questionnaire (JVQ): Administration and scoring manual (pp. 10–18). Crimes Against Children Research Center.

30.

Hanish

L. D.

Guerra

N. G.

(2000). Predictors of peer victimization among urban youth. Social Development, 9(4), 521–543.

31.

Higgins

G. E.

Jennings

W. G.

Tewksbury

Gibson

C. L.

(2009). Exploring the link between low self-control and violent victimization trajectories in adolescents. Criminal Justice and Behavior, 36(10), 1070–1084.

32.

Hong

J. S.

Kim

D. H.

Piquero

A. R.

(2017). Assessing the links between punitive parenting, peer deviance, social isolation and bullying perpetration and victimization in South Korean adolescents. Child Abuse & Neglect, 73, 63–70.

33.

Johnson

M. E.

(2018). Trauma, race, and risk for violent felony arrests among Florida juvenile offenders. Crime & Delinquency, 64(11), 1437–1457.

34.

Kennedy

A. C.

Bybee

Greeson

M. R.

(2014). Examining cumulative victimization, community violence exposure, and stigma as contributors to PTSD symptoms among high-risk young women. American Journal of Orthopsychiatry, 84(3), 284.

35.

Kilpatrick

D. G.

Saunders

B. E.

Smith

D. W.

(2003). Youth victimization: Prevalence and implications. Research in brief. U.S. Department of Justice, Office of Justice Programs.

36.

Laird

R. D.

Jordan

K. Y.

Dodge

K. A.

Pettit

G. S.

Bates

J. E.

(2001). Peer rejection in childhood, involvement with antisocial peers in early adolescence, and the development of externalizing behavior problems. Development and Psychopathology, 13(2), 337.

37.

Loeber

Kalb

Huizinga

(2001). Juvenile delinquency and serious injury victimization. U.S. Department of Justice, Office of Justice Programs, Office of Juvenile Justice and Delinquency Prevention.

38.

Moffitt

T. E.

(1993). Adolescence-limited and life-course-persistent antisocial behavior: A developmental taxonomy. Psychological Review, 100, 674–701.

39.

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(4pt1), 1267–1283.

40.

Monks

S. M.

Tomaka

Palacios

Thompson

S. E.

(2010). Sexual victimization in female and male college students: Examining the roles of alcohol use, alcohol expectancies, and sexual sensation seeking. Substance Use & Misuse, 45(13), 2258–2280.

41.

Morris

V. L.

Huffman

L. G.

Naish

K. R.

Holshausen

Oshri

McKinnon

Amlung

(2020). Impulsivity as a mediating factor in the association between posttraumatic stress disorder symptoms and substance use. Psychological Trauma: Theory, Research, Practice, and Policy, 12, 659–668.

42.

Mounier

Andujo

(2003). Defensive functioning of homeless youth in relation to experiences of child maltreatment and cumulative victimization. Child Abuse & Neglect, 27(10), 1187–1204.

43.

Mustanski

Andrews

Puckett

J. A.

(2016). The effects of cumulative victimization on mental health among lesbian, gay, bisexual, and transgender adolescents and young adults. American Journal of Public Health, 106(3), 527–533.

44.

Nedelec

J. L.

(2018). Individual differences and co-occurring victimization online and offline: The role of impulsivity. Personality and Individual Differences, 133, 77–84.

45.

Pantalone

D. W.

Horvath

K. J.

Hart

T. A.

Valentine

S. E.

Kaysen

D. L.

(2015). Traumatic revictimization of men who have sex with men living with HIV/AIDS. Journal of Interpersonal Violence, 30(9), 1459–1477.

46.

Pereda

Abad

Guilera

(2017). Victimization and polyvictimization of Spanish youth involved in juvenile justice. Journal of Interpersonal Violence, 32(21), 3272–3300.

47.

Pratt

T. C.

Turanovic

J. J.

Fox

K. A.

Wright

K. A.

(2014). Self-control and victimization: A meta-analysis. Criminology, 52(1), 87–116.

48.

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.

49.

Reingle

J. M.

(2014). Victim–offender overlap. In Miller

J. M.

(Ed.), The encyclopedia of theoretical criminology (pp. 1–3). Wiley.

50.

Rhyner

K. J.

Uhl

C. A.

Terrance

C. A.

(2018). Are teens being unfairly punished? Applying the dual systems model of adolescent risk-taking to sexting among adolescents. Youth Justice, 18(1), 52–66.

51.

Rudolph

K. D.

Lansford

J. E.

Agoston

A. M.

Sugimura

Schwartz

Dodge

K. A.

Petit

G. S.

Bates

J. E.

(2014). Peer victimization and social alienation: Predicting deviant peer affiliation in middle school. Child Development, 85(1), 124–139.

52.

Rudolph

K. D.

Monti

J. D.

Modi

Sze

W. Y.

Troop-Gordon

(2020). Protecting youth against the adverse effects of peer victimization: Why do parents matter? Journal of Abnormal Child Psychology, 48(2), 163–176.

53.

Schreck

C. J.

Stewart

E. A.

Fisher

B. S.

(2006). Self-control, victimization, and their influence on risky lifestyles: A longitudinal analysis using panel data. Journal of Quantitative Criminology, 22(4), 319–340.

54.

Shulman

E. P.

Smith

A. R.

Silva

Icenogle

Duell

Chein

Steinberg

(2016). The dual systems model: Review, reappraisal, and reaffirmation. Developmental Cognitive Neuroscience, 17, 103–117.

55.

Skogan

W. G.

(1981). Issues in the measurement of victimization (Vol. 74682). U.S. Department of Justice, Bureau of Justice Statistics.

56.

Skogan

W. G.

(1986). Methodological issues in the study of victimization. In Fattah

E. A.

(Ed.), From crime policy to victim policy (pp. 80–116). Palgrave Macmillan.

57.

Smith

M. D.

(1994). Enhancing the quality of survey data on violence against women: A feminist approach. Gender & Society, 8(1), 109–127.

58.

Steinberg

(2010). A dual systems model of adolescent risk-taking. Developmental Psychobiology: The Journal of the International Society for Developmental Psychobiology, 52(3), 216–224.

59.

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.

60.

Turner

H. A.

Finkelhor

Ormrod

(2006). The effect of lifetime victimization on the mental health of children and adolescents. Social Science & Medicine, 62(1), 13–27.

61.

Walker

H. E.

Freud

J. S.

Ellis

R. A.

Fraine

S. M.

Wilson

L. C.

(2019). The prevalence of sexual revictimization: A meta-analytic review. Trauma, Violence, & Abuse, 20(1), 67–80.

62.

Weinberger

D. A.

Schwartz

G. E.

(1990). Distress and restraint as superordinate dimensions of self-reported adjustment: A typological perspective. Journal of Personality, 58(2), 381–417.

63.

Widom

C. S.

Czaja

S. J.

Dutton

M. A.

(2008). Childhood victimization and lifetime revictimization. Child Abuse & Neglect, 32(8), 785–796.

64.

Wojciechowski

T. W.

(2018). The development of deviant peer association across the life-course and its relevance for predicting offending in early adulthood. Journal of Developmental and Life-Course Criminology, 4(1), 73–91.

65.

Wojciechowski

T. W.

(2020). Early life poly-victimization and differential development of anxiety as risk factors for the continuity of substance dependence in adulthood. Substance Use & Misuse, 55(8), 1347–1355.

66.

Wolfer

T. A

. (1999). “It happens all the time”: Overcoming the limits of memory and method for chronic community violence experience. Journal of Interpersonal Violence, 14(10), 1070–1094.

67.

Yoder

J. R.

Hodge

A. I.

Ruch

Dillard

(2019). Effects of childhood polyvictimization on victimization in juvenile correctional facilities: The mediating role of trauma symptomatology. Youth Violence and Juvenile Justice, 17(2), 129–153.