The first study surveyed a national sample of professionals to determine rankings and agreement on rankings for abuse severity across a number of dimensions. Scales on the ADI were constructed based on the results. The second study examined interrater reliability of the instrument based on semistructured interview data. The third study examined the instrument's construct validity by factor analysis. The findings suggest that considerable agreement exists on how professionals rank abuse features in terms of severity and that ADI scales based on these rankings can be reliably coded from interview data. Also, the instrument was found to have a factor structure consistent with expectations. Generally, the findings of these three studies suggest that the ADI has adequate psychometric properties. In considering potential uses of the instrument, a number of points are noted. First, the construct of abuse severity is probably a matter of subjective interpretations as much as objective behavioral characteristics of the abuse. Second, the reference criteria on which the ADI is based are consensual. Third, although the instrument apparently is amenable for use in coding abuse characteristics and severity from informants or sources other than interviews with nonabusive parents, guidelines for use with other informants have not been developed nor has reliability been established. Finally, certain psychometric limitations of the ADI are noted. 6 tables, 11 references, and appended ADI and ADI semistructured interview and scoring guidelines