Structural equation analysis of a hypothesised symptom model in the autism spectrum


A. Boomsma, University of Groningen, Department of Statistics and Measurement Theory, Boumangebouw, Grote Rozenstraat 31, 9712 TG Groningen, The Netherlands; Tel: + 31 50 3636187; Fax: + 31 50 3636304; Email:


Background:  Several studies showed a different symptom structure underlying the spectrum of autistic-like disorders from the behaviour triad as mentioned in the DSM-IV. In the present study, a hypothesised symptom model for autism was constructed, based on earlier explorative findings, and was put to a confirmatory test.

Method:  Items from the Autism Diagnostic Interview-Revised (ADI-R) were used to examine the goodness of fit of the DSM-IV model, the hypothesised symptom model, and two additional models for autism. All models were tested in a group of 255 verbal and nonverbal individuals with minor to severe autistic symptomatology.

Results:  The DSM-IV model encountered estimation problems. Conversely, the hypothesised symptom model had no such problems and proved to have a better fit to the sample data than the two additional models for autism. However, some of the observed variables were weak indicators of the three latent factors in the model.

Conclusions:  The hypothesised symptom model appeared to be a plausible model in a group of individuals with a broad range of autistic behaviours and levels of functioning. Nevertheless, the stability of the model needs further examination in a larger group of individuals with disorders in the autism spectrum, and with varying degrees of intellectual functioning.