An ambitious search of Criminal Justice Abstracts and PsycINFO databases was conducted. We combed these databases in an effort to acquire all studies that would inform our central research question. To that end, we combined search terms related to all potential independent variables with a list of outcome terms when narrowing the search was necessary. The outcome terms were: delinquency, crime, criminal, violent, violence, property, nonviolent, non-violent, theft, status, aggression, aggressive behavior, and burglary. The search terms used to capture abuse and trauma were: maltreatment, abuse (not substance abuse or drug abuse), physical punishment, harsh punishment, physical discipline, harsh discipline, and neglect. In addition, any item that fit our criteria which we came across through other means was added to the list. The reference lists of important studies and numerous reviews were also examined for possible sources (e.g., Conaway & Hansen, 1989; Hansen, Conaway, & Christopher, 1990; Malinosky-Rummell & Hansen, 1993).
The master list of potential items for review included close to 1,000 items, the titles of which were evaluated for possible “tier 1” items which we deemed had the highest likelihood of reporting associations between abuse/trauma and antisocial behavior. Because there are an enormous number of studies of child abuse where violence is the sole outcome, and reviews of such studies have been published elsewhere (e.g., Falshaw, Browne, & Hollin, 1996; Gershoff, 2002 [corporal punishment]; Haapasalo & Pokela, 1999; Howing, Wodarski, Kurtz, Gaudin, & Herbst, 1990; Jonson-Reid, 1998; Lemmon, 2006 [maltreatment recurrence]; McGrath, Nilsen, & Kerley, 2011 [sexual victimization]; Thornberry, Knight, & Lovegrove, 2012), we decided to narrow our own analysis to studies that inform our central research question: whether abuse has a relationship with violence above and beyond its association with general antisocial behavior. The final sample of studies reviewed here include three main types: those that examine normative samples and present findings on violent and nonviolent antisocial behavior separately, those that examine violent offending among offenders, and those that compare violent to nonviolent offenders. We will also comment on a few studies where violent outcomes are examined using multivariate models controlling for antisocial behavior or nonviolent offending.
When evaluating studies for inclusion, we included those that tested effects of “maltreatment” or mixed forms of abuse, physical abuse, sexual abuse, neglect, or “trauma.” We also included studies of physical discipline, “harsh discipline,” and corporal punishment, deciding not to interpose any ideas we might have about the distinction between abuse and proper discipline. The severity of physical discipline is not often clear from the measures as described, anyway. Regarding childhood sexual abuse, we relied mainly on the author’s definition. We did not further seek out studies of “rape,” as most of these are about adult victims, though in some studies child sexual abuse included rape. While we focused on physical abuse perpetrated by parents and caregivers, most studies of child sexual abuse do not make this distinction, and neither did we. Thus, our tables include studies of parent or caregiver physical abuse (including corporal punishment), caregiver neglect, and any kind of childhood or adolescent sexual abuse victimization. We include studies published in English. We do not include studies of emotional abuse, psychological abuse, verbal abuse, or sibling abuse.
With regard to trauma, we selected all studies reporting effects on physical aggression or violence, or nonviolent antisocial behavior, because the number is small and less is known about this topic. We elected to rely mainly on the authors’ use of the term, and selected studies where the authors were looking at associations between trauma and antisocial outcomes. In a small number of cases, studies where the subjects’ exposure to community violence, victimization, or witnessing violence, was seen as a form of trauma; we included those studies. Because exposure to interparent violence opens the door to a much larger project, we deemed it beyond the scope of this chapter and did not include studies that examined witnessing of interparent violence on its own, with one exception where the authors narrowed it to witnessing “extreme violence” between parents (Lewis, Shanok, Pincus, & Glaser, 1979). In some cases, however, “trauma” as defined by the authors included family violence exposure as part of an index of violence exposure.
The resulting set of studies includes myriad operationalizations of trauma. Although we realize that these operationalizations have little in common in some cases, we made an effort to include only those where the measure is likely to reflect actual traumatic impact. For example, in a few cases, the authors used scales of trauma with many items including such events as natural disasters, accidents, and violent victimization (e.g., Byrd & Davis, 2009). Others used scales of victimization apart from or including family victimization (e.g., Fagan et al., 1983; Maschi et al., 2008; Neller, Denney, & Pietz, 2006; Ramoutar & Farrington, 2006). McNulty and Bellair (2003) defined trauma as witnessing a shooting or stabbing in the past year. Peacock et al. (2003) used an indicator of level of exposure to violence in neighborhoods including the extent that the participant was bothered by the events.
As with the other chapters, our goal is to compare the effects of abuse on physically aggressive and violent behavior to its effects on criminal but nonviolent outcomes. Thus, we exclude studies where these cannot be disentangled. We do not include alcohol or drug outcomes in the “nonviolent offending” category. We did not treat “sex offending” as equivalent to violent offending unless the outcome was specified as violent sex offending. We do not include studies where “weapon carrying” is the only violent outcome.