Abstract
Objective: This analysis used methods of structural equation modeling to assess the extent to which comorbidity between conduct and affective disorders could be explained by (1) common or correlated causal factors that influenced both outcomes or (2) reciprocal causation between these conditions. Method: Data were obtained during the course of a 16-year longitudinal study of a birth cohort of New Zealand children. The data analyzed comprised measures of conduct and affective disorders at ages 15 and 16 years and data on a series of antecedent childhood factors. Results: Structural equation modeling suggested that a substantial component of the comorbidity between conduct and affective disorders arose because the risk factors associated with the development of conduct disorders in teenagers overlapped and were correlated with the risk factors for adolescent affective disorders; of the shared variance between conduct disorder and affective disorders, more than two thirds was explained by common risk factors. These conclusions were replicated using diagnostically scored measures and methods of categorical data analysis. Model extensions suggested an absence of direct causal pathways between conduct and affective disorders. Conclusions: A substantial amount of the correlation and comorbidity between conduct and affective disorders arises because the risk factors and life pathways that predispose adolescents to one outcome are also associated with the risk factors and life pathways that predispose adolescents to the other outcome. Nonetheless, even after control for common causal factors, there was evidence of some unexplained comorbidity between conduct and affective disorders.