The extant literature regarding psychometric properties of measures of subclinical adult autism traits is sparse. The current study investigated the validity and internal consistency reliability of the 26-item Autism Quotient (AQ-26), a self-report measure of autism traits, in a nonclinical young adult sample. Results indicated significantly better model fit with an 18-item measure (AQ-18), with optimal fit indices. Otherwise, AQ-18 and AQ-26 findings were similar and are discussed together. A three-factor model best fit the data. Scores demonstrated similar correlations in expected directions with the Empathy Quotient and Broad Autism Phenotype (BAP) Questionnaire. Expected differences were also observed when college major, gender, and BAP/non-BAP groups were compared. Despite somewhat weaker evidence for the 26/18-Details/Patterns factor, findings indicate that the AQ-26 and AQ-18 are promising measures of the three traits of subclinical autism spectrum symptomatology.