Does Childhood Victimization Matter: A Longitudinal Study of Substance Use and Criminal Activity From Adolescence to Young Adulthood

Background and Purpose: Childhood victimization is a severe social problem that potentially results in long-lasting consequences for adolescents and young adults (Thomson et al., 2002). Previous research indicates that childhood victimization is associated with subsequent substance use problems and criminal behavior supporting the main propositions of both developmental theory (Moffitt, 1993) and social learning theory (Akers, 1998). The purpose of this study is to build on existing literature by using a multilevel growth approach to examine the effects of repeated bullying victimization (RBV) in childhood on substance use and criminal activity among adolescents over time. Additionally, the study investigates whether gender and race/ethnicity moderate the association between RBV in childhood and youth substance use and criminal activity.

Methods: This study used data from the National Longitudinal Survey of Youth 1997 (NLSY 97), a national representative survey of 8,984 U. S. youth. To examine behaviors from ages approximately 12-24, we utilized the longitudinal data from wave 1, 3, 5, and 7 of the NLSY97. The final sample included 5,301 adolescents at wave 1 and 15,491 in total observations. At wave 1, participants were 50% female and 20% Hispanic. The average age of participants was 14 with a range of 12 to 16. Hierarchical linear modeling (HLM) was used to understand the influence of RBV in childhood on substance abuse and criminal activity. Our outcomes were measured as count data. Thus, the multilevel growth model with an overdispersed Poisson sampling distribution was used (Pires & Jenkins, 2007).

Results: The longitudinal analyses showed three important findings. First, the results of the unconditional growth model revealed that the event rate ratios (ERRs) of substance use (i.e., cigarette use, alcohol use, alcohol binges, and marijuana use) increased over time (ERR = 1.74, 1.92, 1.77, and 1.34; p < .001, respectively) whereas, the ERR of criminal activity decreased over time (ERR = .50, p < .001). RBV in childhood was positively associated with the ERRs of cigarette use, marijuana use, and criminal activity over time after controlling for other variables (ERR = 1.48, 1.35, and 1.48; p < .001, respectively). The conditional growth model with interacting covariates demonstrated that the positive effect of RBV in childhood on youth alcohol use was stronger for females than males and for Hispanics than non-Hispanic Whites.

Conclusions and Implications: Our findings demonstrate that RBV in childhood affects later substance abuse and criminal consequences, particularly for women and Hispanics. Using a large sample of adolescents from a representative U.S. longitudinal dataset, this study extends the current literature and support theoretical frameworks. This study contributes to the useful knowledge on youth substance use and criminal activity and its application in social work research and practices. In addition, this study can inform school social workers attempting to reduce adolescents’ bullying behaviors as well as for developing and implementing school-based bullying prevention and intervention programs for adolescents and young adults. Bullying intervention programs should consider victim demographic characteristics when targeting services.