The purpose of this study was to examine the measurement properties of the Social Responsiveness Scale in an accelerated longitudinal sample of 4-year-old preschool children with the complementary approaches of categorical confirmatory factor analysis and Rasch analysis. Measurement models based on the literature and other hypothesized measurement models which were tested using categorical confirmatory factor analysis did not fit well and were not unidimensional. Rasch analyses showed that a 30-item subset met criteria of unidimensionality and invariance across item, person, and over time; and this subset exhibited convergent validity with other child outcomes. This subset was shown to have enhanced psychometric properties and could be used in measuring social responsiveness among preschool age children with Autism Spectrum Disorders.