Short- and Mid-Term Effects of Violent Victimization on Delinquency

Abstract

The present study investigates how adolescents’ experiences of violent victimization exert short- and mid-term effects on their involvement in delinquency. The study compares and contrasts delinquency trajectories of youths whose experiences of violent victimization differ. A multilevel growth-curve modeling approach is applied to analyze data from five waves of the National Youth Survey. The results show that, although delinquency involvement increases as youths experience violent victimization, delinquency trajectories differ with the type of violent victimization, specifically, parental versus non-parental victimization. Violent victimization by parents produced a sharp initial decline in delinquency (short-term effect) followed by a rapid acceleration (mid-term effect). In turn, non-parental violence showed a stable trend over time. The findings have important implications for prevention and treatment services.

Keywords

violent victimization growth-curve model delinquency trajectory

The increased concerns about youths who experience violent victimization either at home or elsewhere have produced a large volume of empirical research in various disciplines to identify its negative impacts. The research literature on victimization suggests that the impacts include a wide variety of psychological and mental problems, such as low self-esteem, high anxiety, depression (Afifi, Brownridge, Cox, & Sareen, 2006; Banducci, Hoffman, Lejuez, & Koenen, 2014; Peltonen, Ellonen, Larsen, & Helweg-Larsen, 2010; Turner, Finkelhor, & Ormrod, 2006; Wood, Foy, Goguen, Pynoos, & James, 2002), and poor social relationships (Silva, Graña, & González-Cieza, 2014), as well as substance abuse (Banducci et al., 2014; Kaukinen & DeMaris, 2005; Lo, Kim, & Church, 2008; Wekerle & Wall, 2002). Numerous studies have also reported a strong relationship between youths’ experience of violent victimization and their later delinquency involvement (Agnew, 2002; Agnew, Brezina, Wright, & Cullen, 2002; Begle et al., 2011; Fang & Corso, 2007; Herrera & McCloskey, 2003; Jackson, Hanson, Amstadter, Saunders, & Kilpatrick, 2013; Lansford et al., 2007; Mersky & Reynolds, 2007; Mulder, Brand, Bullens, & van Marle, 2011; Smith & Ecob, 2007; Watts & McNulty, 2013).

Most studies of a possible violent victimization–delinquency link have utilized either a cross-sectional (Elze, Stiffman, & Dore, 1999; Flannery, Singer, & Wester, 2001) or a retrospective/prospective longitudinal approach (Begle et al., 2011; Coleman & Stewart, 2010; Fang & Corso, 2007; Jackson et al., 2013; Lansford et al., 2002; Mersky & Reynolds, 2007; Silva et al., 2014; Widom, 1989). The cross-sectional approach in general suffers from weak inference about any causal relationship. And although longitudinal approaches have the methodological advantage, previous longitudinal studies on this topic are not exempt from criticism either. Researchers have focused mostly on long-term consequences of violent victimization, measuring/analyzing behavioral outcomes at only one time point (often several years following victimization). When the focus is on links between victimization and subsequent psychological/behavioral changes in the long term, potential short- and/or mid-term psychological and behavioral changes following such trauma may be overlooked.

Several studies have reported that effects of violent victimization vary according to individual characteristics, showing different patterns of internalization and externalization of psychopathology (Keys et al., 2012; Li & Godinet, 2014; Peltonen et al., 2010; Silva et al., 2014; Topitzes, Mersky, & Reynolds, 2012). Also, studies using the developmental perspective have reported that individuals show different trajectories of antisocial behavior depending on the type and severity of violent victimization they experience (Li & Godinet, 2014; Light, Rusby, Nies, & Snijders, 2013; Thompson & Tabone, 2010). However, relative to the victimization–delinquency link, little is known about temporal process/dynamic explaining the manifestation of violent victimization’s consequences in the short term or mid-term. The present study aimed to contribute to the victimization literature (a) by analyzing short- and mid-term consequences of violent victimization through a multilevel growth-curve modeling technique and (b) by comparing delinquency trajectories of youths who experienced different types of violent victimization: victimization by parents (or parental physical abuse) and victimization by others (or non-parental violent victimization).

General Strain Theory (GST) and Violent Victimization as a Source of Strain

Agnew’s (1992) GST provides a clear causal process for the victimization–delinquency link. The theory proposes that youths’ involvement in delinquency is directly related to “negative affective states” they experience (Agnew, 1992). Agnew claims that negative affective states (i.e., anger, frustration, disappointment, helplessness, fear, depression, and so on) are the result of strain caused by (a) failure to achieve positively valued goals, (b) a disjunction of expectations and achievements, (c) removal of positively valued stimuli from the individual, and (d) presentation of negative stimuli.

Many empirical studies have examined links between delinquency and stressful/negative life events, strain, and/or negative affective states. In large part, they have reported that maltreated children are more likely to have depression and other mental health problems (Banducci et al., 2014; DeMaris & Kaukinen, 2005; Peltonen et al., 2010; Turner et al., 2006). In particular, violent victimization during childhood leads to greater anger, anxiety, depression, and aggression, and consequently to increased delinquency (Agnew, 2002; Agnew et al., 2002; Baron, 2004; Hay & Evans, 2006; Manasse & Ganem, 2009; Turner et al., 2006; Wallace, Patchin, & May, 2005; Watts & McNulty, 2013). However, experiencing victimization itself may not be a sufficient condition for later delinquency involvement. In one of the early studies with abused children, Widom (1989) reported that a majority of abused children do not become delinquents or criminals, less than one third having a juvenile arrest record. Noting that not all victimized youths become delinquents or criminals, Agnew (2001) identifies the conditions under which victimization or negative affective states resulting from strain lead to delinquency, such as when strains “are seen as unjust, are seen as high in magnitude, are associated with low social control, and create some pressure or incentive to engage in criminal coping” (Agnew, 2001, p. 351).

Utilizing Agnew’s GST, some studies have focused on the mediating role that negative affective states may play in the victimization–delinquency link (Aseltine, Gore, & Gorden, 2000; Sigfusdottir, Gudjonsson, & Sigurdsson, 2010). Other studies have reported that such factors as temperament, neurological dysfunction, family problems, association with delinquent peers, and neighborhood characteristics may mediate (or moderate) victimization’s effects on delinquency, by influencing the magnitude of negative affective states, the level of social control, or the opportunities available for delinquency (Jaffee, Caspi, Moffitt, & Taylor, 2004; Lahey, Waldman, & McBurnett, 1999). Thus, as Agnew (2001) suggests, parental and non-parental physical victimization’s effects on delinquency may be mediated to an extent that even very intense negative affective states are adequately regulated or controlled by relevant surrounding conditions.

The Present Study and Research Objectives

Using data from the first five waves of the National Youth Survey (NYS), the present study examining violent victimization and delinquency had two goals. The first was to better understand potential short- and mid-term effects of violent victimization on juvenile delinquency. A long-term link between victimization and delinquency has been reported by numerous studies, but little published research has examined victimization’s short- and mid-term effects on the delinquency of youths with differing victimization experiences (Finkelhor, 1995). Several researchers have reported that victims of violence either internalize or externalize their psychopathological problems (Keys et al., 2012; Li & Godinet, 2014; Peltonen et al., 2010; Silva et al., 2014; Topitzes et al., 2012) and that those who internalize these problems may exhibit no overt behavioral problems for a short period following victimization (Li & Godinet, 2014; Light et al., 2013; Thompson & Tabone, 2010). Analyzing longitudinal data from short- and mid-term developmental perspectives should lead to more comprehensive understanding of dynamics shaping the victimization–delinquency link (Macmillan, 2001).

The present study’s second goal was to learn whether youths’ delinquency trajectories differed with the type of violent victimization experienced; types investigated were parental physical abuse versus non-parental violent victimization. We asked, “Is the delinquency trajectory of victims of parental physical abuse different from that of victims of violence by others?” Compared with victims of other violence, parental physical abuse victims may experience more dramatic and more intense strain, leading to stronger negative affective states and exerting a stronger impact on delinquency. Moreover, victims of domestic physical abuse have a relatively high likelihood of being further victimized outside the home (Rich, Gidycz, Warkentin, Loh, & Weiland, 2005). Few previous studies have incorporated parental physical victimization and violent victimization by others to produce a single investigation distinguishing each victimization type’s unique/independent effects on delinquency and/or evaluating the combined effects of various types (Jackson et al., 2013). In examining delinquency trajectories, the present study considered adolescents’ relationships with their parents and peers, exploring these relationships’ roles in the victimization–delinquency link. Again, we acknowledged that victimization may or may not lead to delinquency, given the surrounding circumstances (Agnew, 2001; Ford, Chapman, Mack, & Pearson, 2006; Mersky & Reynolds, 2007). The role played in delinquency by violent victimization by parents may be mediated by children’s associations with delinquent peers. In addition, non-parental victimization’s effects on delinquency may be attenuated by a positive relationship between parent and child.

We intended this study to contribute to the victimization literature by bridging a gap we observed between correlational cross-sectional study and long-term longitudinal study of the topic. Our study investigated delinquency trajectories among youths with differing violent victimization experiences, applying a multilevel growth-curve modeling approach rarely used in research on victimization. We anticipated that examining short- and mid-term delinquency trajectories following violent victimization would provide information valuable in developing and implementing sound intervention programs.

Method

Data

This study used data from the NYS (1976-1987). That longitudinal study began in 1976 and used a nationally representative sample of 1,725 American adolescents aged 11 to 17. As Table 1 shows, the sample comprised 918 males (53%) and 807 females (47%); there were, moreover, 1,361 White respondents (79%) and 364 non-White respondents (21%). The original NYS data included information about these adolescents collected over 11 years, via 7 survey waves. The first five waves were conducted annually. A 3-year gap separated Wave 5 from Wave 6; a 4-year gap separated Wave 6 from Wave 7. The NYS data described adolescents’ delinquency- and victimization-related information, as well as their demographics, family and peer relations, attitude toward deviance, and community involvement. The present study analyzed data from the five annual waves of NYS, identifying short-term and mid-term changes in respondents’ delinquent behavior year by year. (Detailed information on NYS data collection procedures is at http://www.icpsr.umich.edu/icpsrweb/ICPSR/series/88.)

Table 1.

Descriptive Statistics of Variables at Wave 1 (N = 1,725).

Variable	n	%	M	SD
Age			13.87	1.95
Gender
Male	918	53.2
Female	807	46.8
Race
White	1,361	78.9
Non-White	364	21.1
Location
Urban	446	25.9
Non-urban	1,276	74
Victimization by parents (parental physical abuse)
No	1,621	94
Yes	98	5.7
Victimization by others
No	1,184	68.6
Yes	534	31
Self-reported delinquency			12.26	5.658
Peer delinquency			11.9	4.23
Relationship with parent			15.33	3.3

Measures

Delinquency

The present study measured, via respondents’ self-reports, the frequency of engagement in 11 delinquent behaviors such as buying/selling stolen goods, breaking and entering/demaging someone’s property, stealing (money/motor vehicle), selling drugs (marijuana/“hard” drugs), gang fighting, attacking someone, and forcing sexual relations against someone’s will. For each item, offered responses ranged from 1 (never) to 9 (2-3 times daily). Values for all items were summed to obtain a final measure, delinquency involvement. This delinquency index demonstrated adequate reliabilities, ranging from .78 to .85 for the five survey waves.

Time

The variable Time indicated how many years had passed since NYS’s first wave; values ranged from 0 to 4, corresponding to Wave 1, Wave 2, Wave 3, Wave 4, and Wave 5 of NYS. Another variable, Time², captured a potential non-linear trend of delinquency in the growth model.

Victimization

At the first wave, violent victimization by parent (or parental physical abuse; in this study, the terms are used interchangeably) was measured as a dummy variable, by the single item “Had you been beaten up by mother or father in the past year?” (0 = no, 1 = yes). We also included a measure of violent victimization by others. Previous research suggested youths who experienced physical abuse at home were more likely to be violently victimized outside the home as well (Rich et al., 2005). Including a measure for violent victimization by others allowed us to examine both discrete and joint effects, on delinquency, wielded by various types of violent victimization. Victimization by others was measured via two items, asking whether the respondent had been beaten up by a person other than mother or father and whether the respondent had been attacked with a weapon (gun, knife, bottle, or chair) by a person other than mother or father. A value of 0 was assigned when both items generated a response of no. A value of 1 was assigned when at least one item generated a response of yes.

Relationship with parents and peer delinquency

The present study involved first-wave measures for relationship with parents, obtained with a four-item index. The index explored (a) extent of communication between a respondent and the respondent’s parent(s), (b) parental ability to comfort the respondent when unhappy, (c) parental opinion that the respondent does/does not perform well, and (d) degree to which a respondent gets along with parent(s). A 3-point response scale was used, as follows: 1 (not well at all), 3 (OK), and 5 (very well). By summing the responses for the four items, respondents’ relationships with parents were measured. The index demonstrated moderate stability (α = .67).

Peer delinquency was measured at the first wave, via an eight-item index (with reliability of .78) that gauged the prevalence of delinquency among a respondent’s friends. The eight items noted friends’ participation in property damage, marijuana use, theft of something worth less than $5, hitting someone, breaking into a vehicle, and selling “hard” drugs. The 5-point response scale ranged from 1 (none of them) to 5 (all of them). For each respondent, we summed the values for all eight items to obtain peer delinquency’s final measure. We also used control variables such as age, gender, race, and residential location. Descriptive statistics for the study’s variables appear in Table 1.

Multilevel Growth-Curve Models

This study applied a multilevel growth-curve modeling approach, developing two sets of multilevel models that examined whether victimization experience at Wave 1 produced various short- and mid-term effects on delinquency. Model A is an unconditional baseline model with only two time variables (Time, Time²) at Level 1 and with no between-individual factors at Level 2:

Y_{t i} (Delinquency) = Π_{0 i} + Π_{1 i} ({Time}_{t i}) + Π_{2 i} ({Time}_{t i}^{2}) + ε_{t i},

Π_{0 i} = β_{00} + γ_{0 i},

Π_{1 i} = β_{10} + γ_{1 i},

Π_{2 i} = β_{20} + γ_{2 i},

where Π_0i is Person i’s mean delinquency at Wave 1, Π_1i is Person i’s growth rate (or instantaneous growth rate) at Wave 1, and Π_2i is the acceleration rate in each subsequent wave. Model A provides information about changes in average delinquency involvement across all of the survey waves.

Model B adds to Model A all of the between-individual variables, to examine between-individual variables’ effects on mean delinquency (the intercept) and also on coefficients describing linear (Time) and curvilinear (Time²) relationships between time and delinquency. Model B, then, examines whether violent victimization at Wave 1 (by parent or by others) produces any differences in (a) initial mean delinquency at Wave 1 (Π_0i), (b) instantaneous growth rate at Wave 1 (Π_1i), and (c) acceleration (Π_2i) in subsequent years, when other Wave 1 characteristics are controlled:

Y_{t i} (Delinquency) = Π_{0 i} + Π_{1 i} ({Time}_{t i}) + Π_{2 i} ({Time}_{t i}^{2}) + ε_{t i},

\begin{array}{l} Π_{0 i} = β_{00} + β_{01} (Male) + β_{02} (White) \\ + β_{03} (Age) + β_{04} (Urban) + β_{05} (Vic_Parents) + \\ β_{06} (Vic_Others) + β_{07} (Peer_Del) \\ + β_{08} (Relation_Par) + γ_{0 i}, \end{array}

\begin{array}{l} Π_{1 i} = β_{10} + β_{11} (Male) + β_{12} (White) + β_{13} (Age) \\ + β_{14} (Urban) + β_{15} (Vic_Parents) + \\ β_{16} (Vic_Others) + β_{17} (Peer_Del) + β_{18} (Relation_Par) + γ_{1 i}, \end{array}

\begin{array}{l} Π_{2 i} = β_{20} + β_{21} (Male) + β_{22} (White) + β_{23} (Age) \\ + β_{24} (Urban) + β_{25} (Vic_Parents) + \\ β_{26} (Vic_Others) + β_{27} (Peer_Del) + β_{28} (Relation_Par) + γ_{2 i} . \end{array}

The mean delinquency signified by β₀₀ indicates average delinquency at Wave 1, with other variables equal to zero. The coefficients β₀₁, β₀₂, and β₀₄ reflect differences in mean delinquency involvement between male and female, White and non-White, and urban and non-urban, respectively. The coefficients β₀₅ and β₀₆ indicate differences in delinquency involvement between those who were victimized by parents and by others, respectively. The coefficients β₀₃, β₀₇, and β₀₈ represent the effects, on respondents’ own delinquency, of age, relationship with parents, and peer delinquency. In addition, the mean growth rate at Wave 1 and the acceleration rate are signified by β₁₀ and β₂₀, respectively. β₁₀ and β₂₀ provided a baseline for the study’s evaluation of short- and mid-term effects of victimization. Concerning interpretation of other coefficients associated with instantaneous growth rate (Π_1i) and acceleration (Π_2i), they are identical to those just detailed, substituting “instantaneous growth rate” or “acceleration rate” for “mean delinquency.”

Results

Table 2 presents the data on the prevalence of youths’ violent victimization by parents and by others. Of respondents completing the Wave 1 survey, 98 reported they had experienced parental physical abuse in the past year, whereas 534 reported experiencing non-parental violent victimization in the past year. About 63% of those experiencing parental physical abuse (62 out of 98) reported they had also been victimized by others; in contrast, about 12% of those reporting victimization by others (62 out of 534) said they had also experienced parental physical abuse. It is interesting that youths physically abused by parents at home were more likely to also experience violent victimization away from home, whereas most youths victimized away from home were not subjected to domestic violence (χ² = 50.25, p < .001).

Table 2.

Cross-Tabulation of Violent Victimization by Types.

	Victimization by Others
	No	Yes	χ²	Significance
Victimization by parents (parental physical abuse)
No	1,148	472
			50.25	p < .001
Yes	36	62

To learn whether the respondents’ delinquency trajectories exhibited linear patterns or non-linear ones, we conducted an initial analysis using an unconditional quadratic growth model. Results from the analysis appear in Table 3 (under Model A). The Wave 1 mean delinquency (β₀₀) for the respondents was 12.220. Instantaneous growth rate (β₁₀)—which indicates a short-term effect—was −0.248 (p < .05), signifying a decrease in delinquency from Wave 1 to Wave 2. In addition, moving from Wave 2 to Wave 5, the mean acceleration rate (β₂₀)—indicating mid-term effects—was 0.044. The coefficient for the curvilinear relationship (acceleration) showed marginal statistical significance (p = .055), indicating a substantial non-linear relationship.

Table 3.

Quadratic Growth-Curve Modeling of Delinquency (Victimization Variables as Between-Individual Variables).

	Model A		Model B
Fixed Effect	Coefficient	SE	Coefficient	SE
Mean delinquency
Intercept	12.220	0.130***	11.666	0.349***
Male			0.160	0.096
White			0.317	0.163^†
Age			0.018	0.025
Urban			0.240	0.148
Victimization by parents			0.650	0.458
Victimization by others			0.655	0.122***
Peer delinquency			0.266	0.035***
Relationship with parents			−0.029	0.019
Mean instantaneous growth rate (Time)
Intercept	−0.248	0.113*	−0.626	0.377
Male			0.348	0.112**
White			−0.259	0.217
Age			0.047	0.027
Urban			−0.062	0.166
Victimization by parents			−0.801	0.415^†
Victimization by others			0.131	0.168
Peer delinquency			−0.053	0.026**
Relationship with parents			−0.010	0.016
Mean acceleration rate (Time²)
Intercept	0.044	0.023^†	0.222	0.096*
Male			−0.056	0.028*
White			0.064	0.05
Age			−0.017	0.007*
Urban			0.026	0.041
Victimization by parents			0.242	0.107*
Victimization by others			−0.041	0.041
Peer delinquency			0.002	0.006
Relationship with parents			0.002	0.004
Random Effect	Variance Component	χ²	Variance Component	χ²
Intercept	23.040	12,199***	1.522	2,095***
Time (instantaneous growth rate)	16.210	6,682***	1.272	1,845***
Time² (acceleration rate)	0.570	4,681***	0.079	1,868***
Level 1 effect, r_it	4.219		2.650

†

p < .1. *p < .05. **p < .01. ***p < .001.

The obtained negative coefficient for instantaneous growth rate along with the positive, substantial coefficient for mean acceleration indicate that, for the sample as a whole, delinquency significantly decreased from Wave 1 to Wave 2, but with time finally increased. For example, instantaneous growth rate at Wave 2 (Time = 1) was −0.16 (e.g., −0.248 + 2 [0.044] [2 − 1]), and at Wave 3, Wave 4, and Wave 5 was −0.072, 0.016, and 0.104, respectively (the reader may refer to Bryx and Raundenbush [1992)], Formula 6.14 on page 143). That is, mean delinquency declined until Wave 3, then increased from Wave 3 on, suggesting a curvilinear relationship between time passed and delinquency. ANOVA components suggest that significant variations remained in mean delinquency, mean instantaneous growth rate, and mean acceleration rate.

In Model B, between-individual variables were added to Level 2 equations to examine those variables’ effects on mean delinquency, on instantaneous growth rate at Wave 1, and on acceleration rate. Our between-individual variables included gender, race, age, urban residence, and Wave 1 measures of parental physical abuse, victimization by others, relationship with parents, and peer delinquency. The results showed that youths with experience of victimization by others at Wave 1 had a higher mean delinquency than those without such experience (M difference = .655, p < .001). No significant difference in delinquency at Wave 1 was observed between youths with experience of parental physical abuse and youths without it, when other variables were controlled. Peer delinquency was significantly and positively associated with youths’ own delinquency involvement (p < .001).

Gender, parental physical abuse, and peer delinquency were significantly associated with instantaneous growth rate. Males’ instantaneous growth rate at Wave 1 was 0.348 higher than females’ (p < .01). Males’ mean growth rate at Wave 1 was −0.278, whereas females’ was −0.626. This indicates that in the first year, males’ delinquency involvement decreased much more slowly than females’. In addition, respondents with higher peer delinquency at Wave 1 showed a significantly faster decrease in delinquency from Wave 1 to Wave 2.

Illustrating a short-term effect of violent victimization, youths who experienced parental physical abuse tended to demonstrate a slightly faster decrease in delinquency from Wave 1 to Wave 2. The coefficient of the instantaneous growth rate for victimization by parent (−.801) maintained its marginal significance (p = .053), even after controlling for victimization by others and several individual differences. However, the instantaneous growth rate of victimization by others (.131) was not statistically significant. This suggests that youths with and without experiences of victimization by others showed similar changes in delinquency involvement from Wave 1 to Wave 2. Thus, the short-term effect on delinquency made by parental physical abuse differed from that generated by violent victimization by others.

The mean acceleration rate was positive and significant, suggesting that youths’ delinquency significantly increased as time passed, exhibiting a rate of 0.222 units per year. The obtained negative gender coefficient indicates that the mean acceleration rate for females was 0.056 units higher than the mean acceleration rate of males (p < .05). Females’ delinquency involvement increased faster than males’ over subsequent years, despite the fact that a steeper initial decrease (Wave 1 to Wave 2) was observed for females. Age had a significant negative relationship with mean acceleration (p < .05), meaning older youths exhibited a significantly slower increase in delinquency, over time, than younger ones did.

Illustrating a mid-term effect of victimization, parental physical abuse at Wave 1 produced a significant, positive mean acceleration rate (coefficient = .242, p < .05). That is, victims of parental physical abuse showed a faster rise in delinquency over time than youths not abused by their parents, after controlling for victimization by others and several individual characteristics. Parental physical abuse yielded no significant initial difference in mean delinquency, and its short-term effect was to induce a slightly faster decrease in delinquency from Wave 1 to Wave 2. Parental abuse was associated, however, with a more negative mid-term consequence: a faster increase in delinquency over time. However, the statistically insignificant difference in mean acceleration rate (coefficient = −.041, p > .05) between youths with and without victimization by others at Wave 1 suggests that they shared similar delinquency trajectories, although youth who had been victimized by others maintained a higher level of delinquency over time. ANOVA components showed the addition of between-individual variables to Model A to reduce a large portion of variances in mean delinquency (93% of variance eliminated; [23.04 − 1.522] / 23.04), in instantaneous growth rate at Wave 1 (92% of variance eliminated), and in acceleration rate (86% of variance eliminated).

Figure 1 represents the overall, 5-year delinquency trend for all the respondents. Figures 2 and 3 illustrate short- and mid-term effects on delinquency wielded by parental physical abuse and by victimization by others. As Figure 2 shows, youths’ delinquency trajectories differed with the type of violent victimization they experienced. Most interestingly and importantly, youths who experienced parental physical abuse showed a quite different trajectory than youths who experienced victimization by others and youths with no victimization experience. Youths who at Wave 1 had been victimized only by others over time showed a delinquency pattern that was similar to that of youths who had not been victimized at all, although their average delinquency was substantially higher than those not victimized. Youths who were victimized only by their parents exhibited a trajectory similar to youths who were victimized both by parents and by others, although the average delinquency involvement of the latter was substantially higher. Such patterns remained consistent even after we controlled for individual characteristics such as age, gender, race, location of residence, relationship with parents, and peer delinquency (Figure 3).

Figure 1.

Overall delinquency trajectory.

Figure 2.

Delinquency trajectories by types of victimization.

Figure 3.

Delinquency trajectories by types of victimization, controlling for individual characteristics.

Discussion and Implications

Drawing on Agnew’s GST, the present study identified violent victimization as a negative stimulus that likely constitutes stress and negative affective states leading to delinquency. Because circumstances including the magnitude of strain, degree of control in strain-generating situations, and available opportunities for delinquency are central to the link between violent victimization and delinquency, we controlled related variables (peer delinquency, parental relationship, and demographic variables) as we examined the victimization–delinquency link (Agnew, 2001).

The present study intended to fill certain gaps in the literature on victimization and delinquency. For instance, previous studies had not properly addressed important questions about potential short- and mid-term effects of violent victimization (of various types) on delinquency, due largely to the retrospective cross-sectional nature of used data (Fishman, Mesch, & Eisikovits, 2002; Wallace et al., 2005) as well as to a lack of sophisticated analytic techniques even in prospective longitudinal studies (Fagan, 2005; Lansford et al., 2007; Smith, Ireland, & Thornberry, 2005).

The present study aimed to improve on previous research conceptually and methodologically, by applying GST and short- and mid-term developmental perspectives and using a multilevel growth-curve modeling approach. Its results generally support GST, in that the strain generated by violent victimization increased respondents’ level of delinquency later in life. Results also showed, however, that the effects of such strain can vary over time and by type of victimization. “Our study indicated that strain’s negative effects may not appear in the first year following victimization but become more salient as time goes by. The short- and mid-term effects of victimization on delinquency are illustrated in Figures 2 and 3.”

Findings of the study included differing delinquency trajectories for youths experiencing different types of violent victimization. Parental physical abuse at Wave 1 produced a marginally significant rate of instantaneous growth in delinquency (a short-term effect), as well as a significant rate of mean acceleration (a mid-term effect), over the five survey waves. Interestingly, trajectories for respondents victimized by parent only and respondents victimized by both a parent and another person(s) were similar, although among the latter, average delinquency levels were higher across the study period. Moreover, we found that over time, the delinquency trajectory associated with non-parental violent victimization resembled the delinquency trajectory associated with no victimization (although victimization by other raised, on average, measures for delinquency involvement). Comparing our four subsamples (youths reporting no victimization, reporting victimization by others, reporting victimization by parents, or reporting victimization by parents and by others) clearly shows that short- and mid-term effects of violent victimization differ with the type of victimization, whether or not individual characteristics are controlled.

Despite no significant initial difference in mean delinquency and a steep decrease between Waves 1 and 2, youths who experienced parental physical abuse exhibited delinquency levels that, over time, increased more quickly than those of other youths. The implication is that, in time, negative effects generated by victimization by parents become more serious than do effects stemming from other types of violent victimization. Why parental victimization failed to produce a difference in initial delinquency involvement is unclear. One possible explanation for the similar initial mean delinquency of youths experiencing such victimization and of youths not experiencing it may lie in the cross-sectional nature of the Wave 1 data. It is possible that abusive behavior constitutes parents’ response to their children’s delinquency or wrongdoing (Heck & Walsh, 2000). Variations in parenting may underlie the observed weak relationship between parental physical abuse and youths’ delinquency at Wave 1.

Our results showed a substantial decrease in delinquency in the first year following Wave 1 parental physical abuse, along with a significantly accelerated increase in delinquency as time passed. This may, first of all, reflect the possibility that delinquency resulted in severe parental physical discipline and very close monitoring and control of the child. Such parental behavior would limit a child’s opportunities to participate in delinquency, producing temporal decrease in delinquency (Clawford & Novak, 2002; LaGrange & Silverman, 1999; Unnever, Cullen, & Pratt, 2003). Second, youths physically abused by their parents would experience relatively severe trauma, inviting psychological problems of the kind that inhibit social interaction and activity (Silva et al., 2014), thereby limiting, for a short while, the youths’ opportunities to commit delinquency. As GST proposes, over time, relatively more intense negative affective states may develop within youths victimized by parents (Agnew, 2002). The negative affective states curb the creation of positive social bonds (with parents and significant others) and/or facilitate associations with delinquent peers. With the passage of time, however, the capability and opportunity to commit delinquency may be renewed, at which point, any internalized problem may begin to manifest in increased delinquency.

For further understanding of the victimization–delinquency link, our model included as controls several demographic and other factors. The obtained significant results for race and gender are worth discussing. First, Whites were found to report marginally higher levels of delinquency than non-Whites (p < .10), controlling for other variables. Although previous studies generally suggest that non-Whites—especially Blacks—are more likely than Whites to be involved in delinquent acts, our finding of marginally higher self-report delinquency involvement for Whites compared with non-Whites is not surprising (Elliott, 1994) and may be explained by the particular self-report items in our delinquency index, as well as by the fact that all respondents except Whites were comprised by one single, undifferentiated race category (Elliott & Ageton, 1980).

Second, males and females reported different patterns in delinquency involvement over time: In the first year, males’ delinquency decreased more slowly than females’ delinquency, and in subsequent years, females’ involvement increased more quickly than males’ involvement. It may be due to female victims’ tendency to internalize the problems, resulting in higher likelihood of depression/anxiety and running away (Francis, 2014). More opportunities to participate in delinquency following running away may help explain why, as respondents grew older, increases in delinquency among females exceeded increases among males.

Our study’s findings offer some valuable insights about preventing delinquency in youths victimized at home or outside home. Youths who were victimized by others and youths not victimized exhibited similar delinquency trajectories. Both groups over time showed a stable trend in their delinquency involvement, although delinquency among victimized youths exhibited a higher delinquency involvement than youths not victimized. This may imply, ironically, that trauma and other negative consequences of victimization outside home (not at the hands of parents) are easily identified, which may result in the development of adequate treatment programs (e.g., treatment via establishing positive relationships with parents and significant others; Aceves & Cookston, 2007; Teague, Mazerolle, Legosz, & Sanderson, 2008). It is important to help parents develop strong attachment along with good supervision skills, ensuring both adequate monitoring of children and appropriate parental intervention should victimization occur outside home.

As for parental violent victimization, the initial abuse alone may produce long-lasting developmental consequences. Previous studies have associated even minor physical discipline (“spanking”) with children’s later delinquency (Slade & Wissow, 2004; Straus, 1996). Parents should thus be taught to use non-coercive disciplinary actions with children rather than physical force. When physical force/abuse has been used, appropriate treatment of the victim may minimize negative effects. However, often, child protective agencies curtail interventions with victimized children once the number of overt problematic behaviors of the children drops. The present findings suggest that, during this period, victims of parental violence may be internalizing the resultant problems rather than engaging in overt problematic behaviors (i.e., delinquency). Lack of instantaneous behavior problems following parental abuse should not, then, be considered a sign of improvement. Instead, continuing intervention/treatment should be granted during this period.

Limitations

The present study contributed to the literature on victimization and delinquency by addressing important questions about victimization’s short- and mid-term developmental consequences. Its application of a multilevel growth-curve modeling approach and simultaneous comparisons of delinquency trajectories associated with different types of violent victimization make it a meaningful addition to the literature. Four limitations, however, should be mentioned. First, the study used youths’ Wave 1 self-reports of victimization experiences as a baseline measure explaining delinquency in subsequent years. Self-reporting might well leave out incidents of violent victimization by parents, given the potential consequences of such criminal behavior to the entire family. However, perhaps it is easier to report violent victimization in a survey than via official channels. Furthermore, youths’ victimization could possibly have predated Wave 1. A lack of data on prior victimization means that great caution must be taken in generalizing this study’s findings. Second, the present study did not control respondents’ victimization experience over time. It is likely that youths who were victimized are more likely to be re-victimized. Thus, it would be useful to examine how repetitive victimization following the first victimization affects delinquency over time. Third, in our sample, 98 out of 1,725 youths reported experience of parental violent victimization. Although statistical inferences can indeed be drawn from the subsample of 98 victims (we believe that it is still far too many child victims), additional research involving a larger subsample would provide more robust results. Fourth and finally, the present study measured parental victimization using a single survey item on physical abuse, and it rendered parental victimization a dichotomous variable. In contrast, it might well be useful to measure parental victimization using more sophisticated approaches.

In conclusion, although not without its limitations, the present study, in tracking how the initial (or at least early) experience of victimization affects delinquency as time goes by, demonstrated that the first cut may indeed be the deepest. Its findings should bring home to child protective professionals the need to pay careful attention to young victims of violence and ensuring trauma, so that their healthy development is not permanently disrupted.

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.

Author Biographies

Young S. Kim, PhD, is an associate professor of criminology at Eastern Michigan University. His major research interests include social contexts and crime/delinquency, life-course approach to crime, and application of advanced research methods.

Celia C. Lo, PhD, is a professor in the Department of Sociology and Social Work at Texas Woman’s University. Her research interests include the sociology of drugs and alcohol, disparities in health and in health-risk behaviors, and drugs and crime.

References

Aceves

M. J.

Cookston

J. T.

(2007). Violent victimization, aggression, and parent–adolescent relations: Quality parenting as a buffer for violently victimized youth. Journal of Youth and Adolescence, 36, 635-647.

Afifi

T. O.

Brownridge

D. A.

Cox

B. J.

Sareen

(2006). Physical punishment, child abuse, and psychiatric disorders. Child Abuse & Neglect, 30, 1093-1103.

Agnew

(1992). Foundation for a general strain theory of crime and delinquency. Criminology, 30, 47-87.

Agnew

(2001). Building on the foundation of general strain theory: Specifying the types of strain most likely to lead to crime and delinquency. Journal of Research in Crime & Delinquency, 38, 319-361.

Agnew

(2002). Experienced, vicarious, and anticipated strain: An exploratory study on physical victimization and delinquency. Justice Quarterly, 19, 603-632.

Agnew

Brezina

Wright

J. P.

Cullen

F. T.

(2002). Strain, personality traits, and delinquency: Extending general strain theory. Criminology, 40, 43-71.

Aseltine

R. H.

Gore

Gorden

(2000). Life stress, anger, and anxiety, and delinquency: An empirical test of general strain theory. Journal of Health & Social Behavior, 41, 256-275.

Banducci

A. A.

Hoffman

Lejuez

C. W.

Koenen

K. C.

(2014). The relationship between child abuse and negative outcomes among substance users: Psychopathology, health, and comorbidities. Addictive Behaviors, 39, 1522-1527.

Baron

S. W.

(2004). General strain, street youth and crime: A test of Agnew’s revised theory. Criminology, 42, 457-483.

10.

Begle

A. M.

Hanson

R. F.

Danielson

C. K.

McCart

M. R.

Ruggiero

K. J.

Amstadter

A. B.

. . . Kilpatrick

D. G.

(2011). Longitudinal pathways of victimization, substance use, and delinquency: Findings from the National Survey of Adolescents. Addictive Behaviors, 36, 682-689.

11.

Bryx

A. S.

Raundenbush

S. W.

(1992). Hierarchical linear models. Newbury Park, CA: Sage.

12.

Clawford

L. A.

Novak

K. B.

(2002). Parental and peer influences in adolescent drinking: A relative impact of attachment and opportunity. Journal of Child & Adolescent Substance Abuse, 12(1), 1-26.

13.

Coleman

Stewart

L. M.

(2010). Prevalence and impact of childhood maltreatment in incarcerated youth. American Journal of Orthopsychiatry, 80, 343-349.

14.

DeMaris

Kaukinen

(2005). Violent victimization and women’s mental and physical health: Evidence from a national sample. Journal of Research in Crime & Delinquency, 42, 384-411.

15.

Elliott

D. S.

(1994). Serious violent offenders: Onset, developmental course, and termination. Criminology, 32, 1-21.

16.

Elliott

D. S.

Ageton

S. S.

(1980). Reconciling race and class differences in self-reported and official estimates of delinquency. American Sociological Review, 45, 95-110.

17.

Elze

D. E.

Stiffman

A. R.

Dore

(1999). The association between types of violence exposure and youths’ mental health problems. International Journal of Adolescent Medicine and Health, 11, 221-255.

18.

Fagan

A. A.

(2005). The relationship between adolescent physical abuse and criminal offending: Support for an enduring and generalized cycle of violence. Journal of Family Violence, 20, 279-290.

19.

Fang

X. M.

Corso

P. S.

(2007). Child maltreatment, youth violence, and intimate partner violence: Developmental relationships. American Journal of Preventive Medicine, 33, 281-290.

20.

Finkelhor

(1995). The victimization of children: A developmental perspective. American Journal of Orthopsychiatry, 65, 177-193.

21.

Fishman

Mesch

G. S.

Eisikovits

(2002). Variables affecting adolescent victimization. Findings from a National Youth Survey. Western Criminology Review, 3(2), 1-19.

22.

Flannery

D. J.

Singer

M. I.

Wester

(2001). Violence exposure, psychological trauma, and suicide risk in a community sample of dangerously violent adolescents. Journal of the American Academy of Child & Adolescent Psychiatry, 40, 435-442.

23.

Ford

J. D.

Chapman

Mack

Pearson

(2006). Pathway from traumatic childhood victimization to delinquency: Implications for juvenile and permanency court proceedings and decisions. Juvenile & Family Court Journal, 57(1), 13-26.

24.

Francis

K. A.

(2014). General strain theory, gender, and the conditioning influence of negative internalizing emotions on youth’s risk behaviors. Youth Violence and Juvenile Justice, 12, 58-76.

25.

Hay

Evans

M. M.

(2006). Violent victimization and involvement in delinquency. Journal of Criminal Justice, 34, 261-274.

26.

Heck

Walsh

(2000). The Effects of maltreatment and family structure on minor and serious delinquency. International Journal of Offender Therapy and Comparative Criminology, 44, 178-193.

27.

Herrera

McCloskey

L. A.

(2003). Sexual abuse, family violence, and female delinquency: Findings from a longitudinal study. Violence and Victims, 18, 319-334.

28.

Jackson

C. L.

Hanson

R. F.

Amstadter

A. B.

Saunders

B. E.

Kilpatrick

D. G.

(2013). The longitudinal relation between peer violent victimization and delinquency: Results from a national representative sample of U.S. adolescence. Journal of Interpersonal Violence, 28, 1596-1616.

29.

Jaffee

Caspi

Moffitt

Taylor

(2004). Physical maltreatment victim to antisocial child: Evidence of an environmentally mediated process. Journal of Abnormal Psychology, 113, 44-55.

30.

Kaukinen

DeMaris

(2005). Age at first sexual assault and current substance use and depression. Journal of Interpersonal Violence, 20, 1244-1270.

31.

Keys

K. M.

Eaton

N. R.

Krueger

R. F.

McLaughlin

K. A.

Wall

M. M.

Grant

B. F.

Hasin

D. S.

(2012). Child maltreatment and the structure of common psychiatric disorder. British Journal of Psychiatry, 200, 107-115.

32.

LaGrange

T. C.

Silverman

R. A.

(1999). Low self control and opportunity: Testing the general theory of crime as an explanation for gender differences in delinquency. Criminology, 37, 41-72.

33.

Lahey

Waldman

McBurnett

(1999). Annotation: The development of antisocial behavior. Journal of Child Psychology and Psychiatry, 29, 669-682.

34.

Lansford

J. E.

Dodge

K. A.

Pettit

G. S.

Bates

J. E.

Crozier

Kaplow

(2002). A 12-year prospective study of the long-term effects of early child physical maltreatment on psychological, behavioral, and academic problems in adolescence. Archives of Pediatric & Adolescent Medicine, 156, 824-830.

35.

Lansford

J. E.

Miller-Johnson

Berlin

L. J.

Dodge

K. A.

Bates

J. E.

Petti

G. S.

(2007). Early physical abuse and later violent delinquency: A prospective longitudinal study. Child Maltreatment, 12, 233-245.

36.

Godinet

M. T.

(2014). The impact of repeated maltreatment on behavioral trajectories from early childhood to early adolescence. Children and Youth Services Review, 36, 22-29.

37.

Light

J. M.

Rusby

J. C.

Nies

K. M.

Snijders

T. A.

(2013). Antisocial behavior trajectories and social victimization within and between school years in early adolescence. Journal of Research on Adolescence, 24, 322-336.

38.

C. C.

Kim

Y. S.

Church

(2008). The effects of victimization on drug use: A multilevel analysis. Substance Use & Misuse, 43, 1340-1361.

39.

Macmillan

(2001). Violence and the life course: The consequences of victimization for personal and social development. Annual Review of Sociology, 27, 1-22.

40.

Manasse

M. E.

Ganem

N. M.

(2009). Victimization as a cause of delinquency: The role of depression and gender. Journal of Criminal Justice, 37, 371-378.

41.

Mersky

J. P.

Reynolds

A. J.

(2007). Child maltreatment and violent delinquency: Disentangling mail effects and subgroups. Child Maltreatment, 12, 246-258.

42.

Mulder

Brand

Bullens

van Marle

(2011). Risk factors for overall recidivism and severity of recidivism in serious juvenile offenders. International Journal of Offender Therapy and Comparative Criminology, 55, 118-135.

43.

Peltonen

Ellonen

Larsen

H. B.

Helweg-Larsen

(2010). Parental violence and adolescent mental health. European Child & Adolescent Psychiatry, 19, 813-822.

44.

Rich

C. L.

Gidycz

C. A.

Warkentin

J. B.

Loh

Weiland

(2005). Child and adolescent abuse and subsequent victimization: A prospective study. Child Abuse & Neglect, 29, 1373-1394.

45.

Sigfusdottir

I. D.

Gudjonsson

G. H.

Sigurdsson

J. F.

(2010). Bullying and delinquency: The mediating role of anger. Personality and Individual Differences, 48, 391-396.

46.

Silva

T. C.

Graña

J. L.

González-Cieza

(2014). Self-reported physical and emotional abuse among youth offenders and their association with internalizing and externalizing psychopathology: A preliminary study. International Journal of Offender Therapy and Comparative Criminology, 58, 590-606.

47.

Slade

Wissow

(2004). Spanking in early childhood and later behavior problems: A prospective study of infants and young toddlers. Pediatrics, 113, 1321-1330.

48.

Smith

C. A.

Ireland

T. O.

Thornberry

T. P.

(2005). Adolescent maltreatment and its impact on young adult antisocial behavior. Child Abuse & Neglect, 29, 1099-1119.

49.

Smith

D. J.

Ecob

(2007). An investigation into causal links between victimization and offending in adolescents. British Journal of Sociology, 58, 633-659.

50.

Straus

(1996). Spanking and the making of a violent society: The short- and long-term consequences of corporal punishment. Pediatrics, 98, 837-848.

51.

Teague

Mazerolle

Legosz

Sanderson

(2008). Linking childhood exposure to physical abuse and adult offending: Examining mediating factors and gendered relationships. Justice Quarterly, 25, 313-348.

52.

Thompson

Tabone

J. K.

(2010). The impact of early alleged maltreatment on behavioral trajectories. Child Abuse & Neglect, 34, 907-916.

53.

Topitzes

Mersky

J. P.

Reynolds

A. J.

(2012). From childhood maltreatment to violent offending: An examination of mix-gender and gender-specific models. Journal of Interpersonal Violence, 27, 2322-2347.

54.

Turner

H. A.

Finkelhor

Ormrod

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

55.

Unnever

J. D.

Cullen

F. T.

Pratt

T. C.

(2003). Parental management, ADHD, and delinquency involvement: Reassessing Gofftredson and Hirschi’s General Theory. Justice Quarterly, 20, 471-500.

56.

Wallace

L. H.

Patchin

J. W.

May

J. D.

(2005). Reactions of victimized youth: Strain as an explanation of school delinquency. Western Criminology Review, 6, 104-116.

57.

Watts

S. J.

McNulty

T. L.

(2013). Childhood abuse and criminal behavior: Testing a general strain theory. Journal of Interpersonal Violence, 28, 3023-3040.

58.

Wekerle

Wall

A. M.

(2002). Introduction: The overlap between relationship violence and substance abuse. In Wekerle

Wall

A.-M.

(Eds.), The violence and addiction equation: Theoretical and clinical issues in substance abuse and relationship violence (pp. 1-21). New York, NY: Brunner–Routledge.

59.

Widom

C. S.

(1989). Child abuse, neglect, and violent criminal behavior. Criminology, 27, 251-271.

60.

Wood

Foy

Goguen

Pynoos

James

C. B.

(2002). Violence exposure and PTSD among delinquent girls. Journal of Aggression, Maltreatment & Trauma, 6(1), 109-126.