A confirmatory factor analysis of a self-report version of the Liebowitz Social Anxiety Scale

Abstract

One of the most popular measures of social phobia is the Liebowitz Social Anxiety Scale (LSAS; Liebowitz, 1987). The LSAS is a 24-item semi-structured interview measure of fear and avoidance experienced in a range of social and performance situations. Recently, the LSAS has been modified to a self-report version (LSAS-SR) by several independent groups (Cox, Ross, Swinson, & Direnfeld, 1998; Fresco et al., 2001; Mancini, Van Ameringen, & Oakman, 1999). A self-report version offers ease of administration, but it may differ from the structured interview version in its psychometric properties. We conducted confirmatory factor analyses of the self-report version of the LSAS using data from a sample of 188 outpatients with anxiety disorders. The structure and psychometric properties of the LSAS-SR are highly similar to that of the LSAS and robust across groups of patients with a variety of primary anxiety disorders. We argue in favor of adopting the 4-factor model for the LSAS proposed by Safren et al. (1999) instead of the models implied by the scoring instructions for the LSAS.