Comparative Analysis of Psychometric Measures for Autism Spectrum Disorder Across Age and Sex

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder that has expanded in research focus over recent years. As diagnoses of the condition continue to impact a broader demographic profile, understanding the intersection of current evaluative psychometric tools and demographic factors becomes imperative. This paper analyzes the impact of age and sex on the age-related patterns of various behavioral domains used in the screening or diagnosis of autism, utilizing ASD populations between the ages of 5 and 30 pooled from the Autism Brain Imaging Data Exchange I & II (ABIDE I & II) databases. This study had three objectives. First, separate Generalized Additive Models (GAMs) were fitted to each behavioral domain to estimate the relationship between age and scores on individual assessments. Second, multiple linear regression models were used to assess sex differences in each behavioral domain with age as a covariate. Third, separate sex-specific GAMs were compared to a shared age-based GAM to determine whether developmental trajectories significantly differed in shape across males and females. The findings show that Vineland ABC had the largest age effect, with a 20.7% deviance explained, followed by the Social Communication Questionnaire (SCQ Total), with a 7.3% deviance explained. The sex effects across all behavioral domains were statistically insignificant, and none of the domains showed a meaningfully different sex-specific trajectory shape after false discovery rate correction. Overall, the study suggests that age may play a role in shaping the score trajectories on evaluative psychometric instruments used in autism screening and diagnosis. On the other hand, sex differences show no evidence of impacting developmental trajectories.