Measurement invariance and latent mean differences of the 20-item Toronto Alexithymia Scale across genders and across clinical and non-clinical samples

Abstract This study examined the measurement invariance and latent mean differences of the Chinese version of the 20-item Toronto Alexithymia Scale (TAS-20) across genders and across samples (undergraduates vs. clinical patients). A total of 1373 undergraduates and 1004 patients completed the TAS-20 in this study. Confirmatory factor analyses (CFA) confirmed that the three-factor model of the TAS-20 fitted the data well, both in undergraduate sample and in clinical sample and each gender subgroup. Multigroup CFA demonstrated the partial measurement invariance of TAS-20 across genders and across samples. Analyses of latent mean differences revealed that males exhibited higher levels of externally oriented thinking than females, in both clinical and undergraduate samples, and clinical patients scored higher on all the three factors of TAS-20 than undergraduates. The findings provided further support for the three-factor model of the TAS-20 in Chinese undergraduates and clinical patients. The results of measurement invariance indicated that the gender and sample differences in alexithymia could be directly evaluated by comparing TAS-20's raw scores. Furthermore, the results of latent mean differences suggested that males might adopt more externally-oriented thinking modes than females, and clinical patients displayed higher levels of alexithymia than undergraduates.