Psychometric properties of the Social Anxiety Scale (SAS) using psychophysical and categorical scaling

This study evaluated the psychometric properties of a newly developed instrument, the Social Anxiety Scale (SAS) within a non-clinical Swedish adult population (n: 384), in which psychophysical (SAS-cMax) and categorical (SAS-C) scaling methods were compared. Unlike many common rating scales, the SAS items were designed to align with the perceived intensity of symptom severity. With a reduced number of items, both scaling methods (SAS11-cMax and SAS13-C) demonstrated robust conceptual validity, with a three-factor model encompassing "fear of interaction in groups and with strangers", "fear of speaking in front of others and being in the center of attention", and "fear of performing tasks while being observed". In addition, the SAS showed an excellent internal consistency (SAS11-cMax, alpha = 0.93 and SAS13-C, alpha = 0.93), as well as good to satisfactory convergent and discriminant validity. ROC-analyses revealed an excellent accurcay of both scaling methods to distinguish between high and low social anxiety levels. A score of 25.3 cM ("moderate social anxiety") yielded the most optimal sensitivity (0.80) and specificity (0.80) of the SAS11-cMax. A score of 18.5 points yielded the most optimal sensitivity (0.85) and specificity (0.87) of the SAS13-C. Overall, its concluded that the SAS exhibits sound psychometric properties regardless of scaling method. Further validation with other social anxiety scales and within a clinical population, is needed to establish its effectiveness as a screening tool for social anxiety.