Psychometric evaluation of the Comprehensive Autistic Trait Inventory (CATI) in autistic and non-autistic adults

Abstract

Autistic traits may be assessed for a variety of purposes including as a screener before seeking autism assessment, and as a measure of autistic-like traits as expressed in the general population for use in research. For such tools to be useful in research and clinical contexts, it is paramount that autistic trait inventories efficiently and accurately assess autism as it is currently defined and that they are valid in different demographic groups. The present study assesses the psychometric properties of the Comprehensive Autistic Trait Inventory (CATI), a relatively recently developed measure of autistic traits that examines six trait domains, including sensory sensitivity and social camouflage) using 42 self-report statements. To properly examine the CATI, we recruited large samples of autistic (n = 1,322) and non-autistic (n = 1,279) participants to complete the scale. Factor fit for the six subscale factor structure was appropriate as was total-scale and subscale reliability. Sensitivity and specificity supported the use of the CATI to discriminate autistic and non-autistic participants and total-scale classification threshold developed. Additional analyses examined differences between diagnosed (n = 737) and self-identifying (n = 585) autistic participants, with little to differentiate the groups. Gender differences for the subscales were noted across both autistic and non-autistic groups with a trend for higher scores to be observed in participants of a gender minority (n = 336) compared to cis-males (n = 1,205) and cis-females (n = 1,059), and gender-specific thresholds additionally computed. Lastly, the effect of age on autistic traits was examined with traits in non-autistic participants generally decreasing in older age groups and more complex pattern across the subscales noted for autistic participants. Overall, our results indicate that the CATI is a practical measure of autistic traits in both non-autistic and autistic participants and should be of use to researchers, clinicians, and the general public alike.