There are several studies demonstrating that early childhood disorders, including externalizing disorders, may impact an individual’s premorbid personality consistent with the scar model (e.g., Roberts, Walton, Bogg, & Caspi, 2006; Shiner & Masten, 2002). In contrast, the current study did not examine reciprocal effects between measures of temperament and externalizing behavior. Rather, the focus was on the more common pathway from early temperament to later externalizing behavior for the purposes of simplicity and clarity. Future studies should explore potential bidirectional relationships between temperament and problem behavior, with a particular focus on what neural structures might be implicated in creating such associations. Although measures included in the current study are well-known and had good reliability in the current sample, previous work demonstrates that mediating variables should be measured with a reliability of at least 0.90, otherwise measurement error may introduce bias into estimates of mediation effects (Hoyle & Robinson, 2003). Moreover, measures should be selected that have the potential to clarify the mechanistic structure of these connections. With the present variables, we can only do some imprecise estimation, and suggest some plausible, but untested associational linkages.
