Factor Structure and Psychometric Properties of the Child Social Preference Scale-3 in Children With Autism

Considerable variability in social engagement among individuals with autism is well documented. Since multiple processes may contribute to this heterogeneity, validating tools to assess these differences is crucial. Originally developed in the general population, the Child Social Preference Scale (CSPS-3) aims to assess distinct forms of social disengagement arising from different combinations of approach and avoidance motivations and holds promise for delineating variability in social behaviors within autism spectrum disorder (ASD). This study investigated the factor structure and psychometric properties of the CSPS-3 in a sample of 689 children diagnosed with autism (Mage = 11.23, SD = 3.56; 76% male). The results suggest that a bifactor model, consisting of a general factor and three subscales (shyness, unsociability, and social avoidance), provided the best fit to the data, with the general factor accounting for most of the variance. While the subscales demonstrated adequate internal consistency, their construct reliability and stability varied, with much of the reliable variance attributed to the general factor. The structure was consistent across age and sex subgroups, and the subscales showed distinct patterns of associations with key clinical correlates. These findings support the CSPS-3's utility in assessing diverse forms of social disengagement in the autism population, while indicating that the subscales could be refined to better capture their unique aspects.