boys to meet criteria for ODD than girls. We suspect that these study differences may largely reflect differing treatments of ODD/CD comorbidity: by definition, our ‘ODD only’ group excluded cases with comorbid CD, whereas studies examining overall prevalence rates frequently do not. Second, parental use of the child as a scapegoat was elevated in ODD (whether comorbid with CD or not) but not in CD as was also highlighted in other analyses of the GSMS (Shanahan et al., 2008). It is possible that this represents a parental response to child oppositonality rather than a specific risk factor for ODD.
