Shortened version of TAS using IRT: The 10-item scale for adolescents

Abstract

Alexithymia is a multidimensional construct characterized by difficulty in identifying and describing feelings. Considering that the assessment of alexithymia in adolescents is very difficult, the aim of the present study was to examine the TAS-20 psychometric properties and its efficiency among a sample of Italian adolescents. In order to achieve this goal, the Rasch model, partial credit (PCM) specification with joint maximum likelihood estimation, was used. The Rasch results were used to select a set of items with high validity and discriminating power for adolescents, without any loss in measurement invariance. The final brief 10-item version of the scale, TAS-10, resulted to be valid and reliable: its correlation with the original 20-item score was remarkably high and the network with convergent and divergent measures gave the expected relationships. Furthermore, the concordance of the classification obtained from the 20 and 10-item versions was adequate, but the latter had the noticeable quality of reducing the quota of possible alexithymic adolescents, distinguishing non-alexithymic or alexithymic persons from those who are not clearly classified. Finally, the TAS-10 is definitively more suitable for adolescents in clinical and research practice.