Factorial structure of the 20-item Toronto Alexithymia Scale in a large sample of somatoform patients

Although a strong association between alexithymia and somatization has been postulated in numerous studies, no systematic study has investigated the psychometric properties of the 20-item Toronto Alexithymia Scale (TAS-20) in a sample of patients with somatoform disorder yet. The purpose of this study was to ensure a valid assessment by the German version of the TAS-20 in somatoform samples. We investigated whether the original three-factor model proposed by Bagby et al. (1994a), which is widely used in clinical research and practice, is replicable in a large sample of somatoform patients (n=806). Using confirmatory factor analysis (CFA) the goodness-of-fit of the originally proposed factor structure was compared to three factor models generated with exploratory factor analysis (EFA) and other factorial solutions derived from the literature. Our results demonstrate that the original three-factor model is not replicable in somatoform patients. Instead, the four-factor model by Franz et al. (2001b) described the data best. However, none of the models met all criteria of confirmatory factor analysis. Our results indicate that the three-factor model is not robust in the German version of the TAS-20. At this state of research we recommend to use the TAS-20 sum-score as a measure of alexithymia in somatoform patients in clinical practice.