accounted for a large proportion (61%) of the phenotypic correlation between aggressive and non-aggressive DSM-IV criteria (Gelhorn et al., 2006). However, criterion-level heritability analyses indicated that specific DSM-IV conduct disorder criteria (e.g., truancy) showed little to no heritability but rather high levels of shared environmental influence (Gelhorn et al., 2005). Exploratory twin models similarly show that there are distinct etiological influences on conduct disorder criteria. For example, a series of multivariate twin models of DSM-III-R conduct disorder criteria in a sample of male-male twin pairs found evidence for two genetic factors, one shared environmental factor, and one nonshared environmental factor (Kendler et al., 2013). The first genetic factor included items related to rule-breaking (truancy, running away, and telling lies), and the second genetic factor included items related to overt aggression (hurting people and fighting Thus, both hypothesis-driven and data-driven approaches suggest that conduct disorder has a complex genetic architecture, meaning that there are multiple sources of genetic influence on the various behaviors that are encompassed in conduct disorder diagnoses.
