General Strain Theory and Stability in Offending and Substance Use Over Time: A Dynamic Approach

Abstract

One of the hallmarks of a good theory is that it can explain the known facts. Therefore, it is surprising that little research has examined whether General Strain Theory (Agnew 1992, 2006) can account for the relative continuity in antisocial behavior that individuals generally display over their time. The current study fills this void in the criminological literature by testing the ability of General Strain Theory (GST), in combination with the broader stress literature, to account for stability in offending and substance use from adolescence to adulthood. Four mechanisms that Agnew (1997, 2006) argues lead to behavioral continuity--a direct effect, evocative and active selection, passive selection, and stressor and deviance amplification--are examined using structural equation modeling. Drawing from the broader stress literature and the life-course perspective, two additional pathways--stress proliferation and the moderating effect of past exposure to stressors--are tested. This research is conducted using two unique datasets, the Collaborative Perinatal Project and the Pathways to Adulthood Study, which together provide information on the lives of 1,758 high risk individuals from birth through adulthood. Support for GST explanations of behavioral continuity is mixed, with more support for the dynamic mechanisms that do not rely on negative emotionality and low constraint. Specifically, for both offending and substance use, there is no evidence to suggest that evocative and active selection or passive selection contribute to the stability of criminal behavior, however, stress proliferation and stressor and deviance amplification each explain a small portion of the association between adolescent and adult illegal behavior. In addition, the findings indicate that negative emotionality and low constraint condition the effect of stressors on criminal behavior, as does exposure to stressors in childhood. The findings for offending and substance use diverge only with regard to the direct effects of negative emotionality and low constraint: the direct effect of these variables on criminal behavior accounts for continuity in substance use, but not offending. It is argued that GST's emphasis on individual differences may be misplaced and that more attention should be directed to exploring the social processes through which stressors develop over time.