Background: Multivariate statistics can assist in refining the nosology and diagnosis of pervasive developmental disorders (PDD) and also contribute important information for genetic studies. The Autism Diagnostic Interview-Revised (ADI-R) is one of the most widely used assessment instruments in the field of PDD. The current study investigated its factor structure and convergence with measures of adaptive, language, and intellectual functioning.
Methods: Analyses were conducted on 1,861 individuals with PDD between the ages of 4 and 18 years (mean = 8.3, SD = 3.2). ADI-R scores were submitted to confirmatory factor analysis (CFA) and exploratory factor analysis (EFA). Analyses were conducted according to verbal status (n = 1,329 verbal, n = 532 nonverbal) and separately for algorithm items only and for all items. ADI-R scores were correlated with scores on measures of adaptive, language, and intellectual functioning.
Results: Several factor solutions were examined and compared. CFAs suggested that two- and three-factor solutions were similar, and slightly superior to a one-factor solution. EFAs and measures of internal consistency provided some support for a two-factor solution consisting of social and communication behaviors and restricted and repetitive behaviors. Measures of functioning were not associated with ADI-R domain scores in nonverbal children, but negatively correlated in verbal children.
Conclusions: Overall, data suggested that autism symptomatology can be explained statistically with a two-domain model. It also pointed to different symptoms susceptible to be helpful in linkage analyses. Implications of a two-factor model are discussed.