Evaluation of the factor structure of the Autism-Spectrum Quotient

The Autism-Spectrum Quotient (AQ) is a self-report scale developed as a screener for Autism Spectrum Disorders and also to measure autistic traits found in the general population. Research has examined its factor structure, resulting in several different measurement models. The current study empirically tested previous models of the AQ using confirmatory factor analysis (CFA) with data from 522 university students. For all models tested, several goodness-of-fit indices failed to reach adequate fit criteria. Given the inadequate fit found for the previous AQ models, for the second part of our study we randomly divided our sample into two groups to derive and test an alternative factor structure for the AQ. With data from the first group (n = 300), a principal components analysis was conducted which revealed a 28-item solution with five factors: (1) social skills, (2) communication/mindreading, (3) restricted/repetitive behaviour, (4) imagination, and (5) attention to detail. Employing data from the second group (n = 222), this 5-factor structure was tested using CFA and was partially confirmed. Findings are discussed with regards to the adequacy of the AQ for assessing autistic traits in the general population.