Our own symptom count based analyses found ODD symptoms at wave 1 predicted later CD symptoms but did not identify the sex difference indicated in the diagnosis based analyses. ODD symptoms at wave 1 were similarly predictive of later CD symptoms for boys and girls even after control of wave 1 CD symptoms. This discrepancy between diagnosis and symptom based analyses may be relevant to the debates on the utility of sex-specific diagnostic thresholds for CD (Zoccolillo, 1993). A dimension of CD severity is implied by the DSM-IV specification of mild, moderate and severe CD. Increased prominence for a dimensional approach has been considered for DSM-V (Moffitt et al., 2008) and there is increased recognition that diagnostic thresholds may be arbitrary (e.g., Pickles et al., 2001). In clinical practice, however, dichotomous decision making will continue to be required, for example in deciding whether and when to provide treatment. Therefore dichotomous diagnosis is likely to remain a prominent feature of DSM-V (Angold & Costello, 2009). Identification of thresholds that are effective for both boys and girls is an important challenge for research.
