Psychometric Evaluation of the Spanish 10-Item Autism Spectrum Quotient Children’s Questionnaire in a Sample of Ecuadorian Children

PURPOSE: The 10-item Autism Spectrum Quotient (AQ-10) is widely used as a brief screening tool for autism spectrum disorder (ASD). Its psychometric properties have not been examined in Spanish-speaking South American populations. This study evaluated the reliability, dimensionality, item functioning, and scalability of the parent-report Spanish AQ-10. METHODS: Data were drawn from Proyecto Guagua, a community-based study in Ecuador. An age-appropriate subsample of 295 children (ages 5-11 years) was assessed using the parent-report Spanish AQ-10. Reliability was assessed with Kuder-Richardson-20 (KR-20), standardized α and tetrachoric α. Dimensionality was examined using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). Item discrimination and difficulty were evaluated using exploratory item response theory (IRT), and scalability was assessed using Mokken scaling. RESULTS: Internal consistency was modest under binary scoring (KR-20 = 0.49; standardized α = 0.54) but improved when tetrachoric correlations were used (α = 0.73). CFA indicated suboptimal fit for a one-factor model. Item 2 consistently displayed negative discrimination (-0.104) and negligible factor loadings (~ -0.03). IRT analyses showed strong discrimination for a subset of items, particularly Items 3 and 6, whereas most items exhibited weak discrimination. Mokken scaling produced a low scalability coefficient (H < 0.30). In the analytic sample, 6.44% (n = 19) scored at or above the clinical cut-off of ≥ 6. CONCLUSIONS: The Spanish AQ-10 did not meet conventional psychometric standards in this Ecuadorian community sample. These findings warrant local cultural adaptation and formal validation against clinical diagnoses before the AQ-10 can be recommended as ASD screening tool.