The classification agreement of the Autism Diagnostic Interview-Revised (ADI-R) and the Autism Diagnostic Observation Schedule (ADOS) was examined in 129 children and adolescents (aged 7-18 years) who were evaluated for autism. Participants received a diagnosis of autism or non-autism based on the ADI-R. Linear discriminant analysis revealed adequate concordance between the ADI-R and ADOS, with 75% of the participants being correctly classified using the ADOS. Classification accuracy significantly improved to 84% when a measure of adaptive functioning (i.e., the Vineland Adaptive Behavior Scales) was included in the analysis. The findings suggest that when clinicians obtain discrepant information on the ADI-R and ADOS, assessment of an individual's adaptive functioning may reduce diagnostic errors.