Measurement-Dependent Sex-Based Differences in Autistic Traits: Comparing Self-Report, Informant-Report, and Clinician Observation in Adults With Autism Spectrum Disorder.

Abstract

Autistic traits are commonly assessed using self-reported questionnaires, informant-reported measures, and clinician-administered diagnostic instruments; however, the convergence among these approaches and how they vary according to sex remain incompletely understood. We aimed to examine associations and discrepancies among self-reported autistic traits (Autism-Spectrum Quotient Japanese version [AQ-J]), informant-reported traits (Social Responsiveness Scale, Second Edition [SRS-2]), and clinician-rated autism trait intensity (Autism Diagnostic Observation Schedule, Second Edition [ADOS-2], Module 4) in adults with autism spectrum disorder (ASD; n = 219) and typically developing (TD; n = 130) individuals. We also evaluated whether these patterns differed based on sex across measures. AQ-J and SRS-2 scores were higher in ASD than in TD participants. Questionnaire-based measures showed minimal sex differences, whereas ADOS-2 scores were higher in males within the ASD group. Repeated-measures analyses revealed a significant measure-by-sex interaction (p < 0.001, partial η2 = 0.036). This reflected significantly lower clinician-rated trait intensity in females, despite non-significant questionnaire-based sex differences, with no overall sex difference in mean trait intensity. Self-reported, informant-reported, and clinician-rated assessments capture overlapping but distinct aspects of the autism phenotype, and the pattern of sex differences depends on the measurement modality used.