Due to ongoing criticism of the current diagnostic system, hybrid models of psychopathology, involving the combination of the categorical approach to diagnoses with shared symptom dimensions that are common across various disorders, have received increasing attention in the literature for the depressive and anxiety disorders. One of the few empirically-derived hybrid models was recently developed for social anxiety disorder (SOC) and subsequently led to a self-report measure entitled the Multidimensional Assessment of Social Anxiety (MASA). Although initial support has been provided for the psychometric properties of the MASA, additional research is needed. The present study investigated the internal consistency, convergent and discriminant validity, and sensitivity to detect treatment changes of the MASA scales in 116 participants with SOC that completed the measure at pretreatment and posttreatment during a trial of group cognitive behavioral therapy (CBT) for SOC. The MASA scales demonstrated acceptable sensitivity to detect symptom changes during treatment, convergent validity with related measures of SOC, and internal consistency. In addition, the observed pattern of symptom changes across the MASA scales appeared to be most consistent with the goals/techniques involved in CBT for SOC (e.g., largest improvements in behavioral avoidance; smallest improvements in substance use). Together, these findings provide additional support for psychometric properties of the MASA scales as well as the growing movement towards transdiagnostic assessment and treatment practices.