Attachment and Parental Sensitivity: Findings and Research Needs
Early childhood studies of attachment do not have an analogous outcome for nonviolent offending. So while those studies indicate that insecure attachment is associated with physically aggressive behavior, they do not help us compare effects on nonviolent offending. The evidence from other studies suggests that attachment to parents is not a differential predictor of violence. The evidence on two particular measures, parental loss or separation and caregiver sensitivity, was inconclusive. There were some findings on parental loss which strongly favored an effect on violence, and theoretically, loss should have a differential effect on violence. This might indicate that very significant problems with parent attachment are required to trigger violent outcomes, and we see this as consistent with both the developmental literature and our expectations. Thus, further research should tap into the extent of attachment and bonding problems, ensuring that part of the sample has had significant disruptions or bonding problems, and also make comparisons between violent and nonviolent offenders. We provided an analysis of Add Health data which confirmed our conclusion that, as of now, there is no evidence to suggest that violent offenders have lower levels of attachment than nonviolent ones, or that attachment is more strongly associated with frequency of violence than nonviolent offending.
Because of the mixed findings in studies that measured attachment using the strange situation, we caution those studying early childhood attachment and physical aggression not to combine subcategories of insecure attachment, which appear to have differential correlations with physically aggressive behavior. In addition, among studies of attachment in young children, it appears that the association between physical aggression and secure attachment may not exist for females. This remains to be tested in a more carefully designed study.
We encourage additional studies of parental sensitivity. These could be used to illuminate the particular parenting practices that reduce violent behavior in children. Because there is no consistent analog for “nonviolent offending” in preschool and elementary school-aged children, using longitudinal studies to predict violent and nonviolent offending in adolescence would be advantageous for tests of the differential etiology hypothesis. There are not many studies of this kind, and such a study could inform other lines of research as well. Better understanding of the parental practices that do (or do not) impede the manifestation of violent behavior over the child’s life course would have policy applications in parent training programs.
Although most measures of bonds to parents, even weak ones, are negatively associated with violence in adolescents, better measures of parent behavior associated with bonding could be used to help determine the types of behaviors that make adolescents feel more bonded, information that could also be used to improve parent interventions.