The Effects of Victimization on Offending: An Examination of General Strain Theory, Criminal Propensity, Risk, Protection, and Resilience

ABSTRACT

General strain theory (GST) proposes that criminal coping is most likely for victims who possess multiple individual and environmental risk factors that together create a strong propensity for offending. However, this conceptualization does not consider the potential buffering effects of multiple protective factors on the victimization/offending relationship. This study addresses this limitation using self-report data from waves I and II of the National Longitudinal Study of Adolescent to Adult Health (Add Health) (N = 13,346). Negative binomial logistic regression models assessed the conditioning effects of composite and individual measures representing criminal propensity/risk and protection/resilience on the victim/offending relationship. Results support GST’s premise that victimization increases involvement in offending but, contrary to GST, greater risk weakened the relationship between victimization and offending while greater protection strengthened this relationship. Findings highlight a need for further theoretical development and a greater understanding of the subjective experience of victimization and pathways associated with increased risk and resiliency.

KEYWORDS:

• General strain theory
• victimization
• offending
• criminal propensity
• risk factors
• protective factors

Disclosure statement

The authors report there are no competing interests to declare.

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/15564886.2022.2077495.

Notes

1. When assessing moderating effects, Moon and Morash (2017) combined victimization with other types of strain, such that the findings do not strictly reflect a moderating effect of propensity on the relationship between victimization and offending.

2. Differences exist between the study sample and excluded cases on the independent variables and study covariates, but they are small (all less than one-fourth of a standard deviation) and unlikely to significantly impact results. Further, analyses included Add Health sampling weights that adjust for differential attrition and nonrandom missingness in the data.

3. Agnew (2002) highlights the potential moderating role of negative emotionality. While we were unable to assess this characteristic as defined by Agnew (2002), we were able to assess the moderating role of depressive symptomology, which was the closest measure available in Add Health.

4. Perhaps more than any criminological construct, intelligence and/or grade point average (GPA) has consistently been operationalized to denote risk at low levels and protection at high levels (Bernat et al., 2012; Lösel & Farrington, 2012), and this dual role was retained in the current study.