Objectives: This study investigated whether empirically derived dimensions of autistic behavior are consistent with the content-valid construction of the autistic behavior domains according to ICD-10 and DSM-IV (social interaction, communication and repetitive, stereotyped behavior).
Methods: A principal component exploratory factor analysis routine with varimax-rotation and extraction of factors following the Scree criterion was run using data from the Autism Diagnostic Interview-Revised (ADI-R) of N = 262 individuals exhibiting autism or autistic features.
Results: A three-factor solution consisting of two socio-communicative and one language dimension and accounting for 46.1% of the total variance was found to best describe the data. These factors yielded only vague correspondence with the idea of behavior domains described in ICD-10 and DSM-IV. In addition, factor loadings of items representing repetitive, stereotyped patterns were generally weak.
Conclusions: The factor-analytic approach to autism indicates a conception of the disorder divergent from that defined in the contemporary psychiatric classification systems, especially regarding the area of repetitive, stereotyped behavior.