Items are scored in five subscales: Irritability, Lethargy/Social Withdrawal, Stereotypic Behavior, Hyperactivity/Non-compliance, and Inappropriate speech. Internal consistency was reported as good by Karabekiroglu and Aman  (Cronbach’s alphas from 0.68 to 0.90) and by Kaat, Lecavalier and Aman  (alphas from.77 to.94). Inter-rater reliability (between similar raters) and test-retest reliability were not assessed. Brinkley et al.  and Kaat, Lecavalier and Aman  demonstrated that the ABC had good structural validity; the latter very large study (n = 1893) found that 90% of items matched the standard ABC factor structure, though the model fit was ‘marginal’ (Root Mean Square Error of Approximation (RMSEA) was .086). Sigafoos et al.  also showed that the ABC had good structural validity with five factors, though due to the small sample size (n = 32), the Sigafoos paper was judged to be of poor methodological quality. Karabekiroglu and Aman  showed that the ABC distinguished between clinical subgroups. Kaat, Lecavalier and Aman  found, as expected, that irritability and hyperactivity decreased with age. Kaat, Lecavalier and Aman  and Karabekiroglu and Aman  found predicted significant correlations with related constructs measured by the Child Behaviour Checklist and the Autism Behaviour Checklist, and Kuhlthau et al.  with a measure of child quality of life, the Child Health and Illness Profile—Child Edition.