The TAS-9-C is a short form of TAS-20 validated for assessing alexithymia in adolescent populations. Tam & Wong (2012) reported the invariance of the one-factor model of TAS-9-C; yet this resulting model has yet been tested against its original three-factor model of the TAS-20. The objective of this study was to test the goodness-of-fit of both models in three different Chinese samples.

Data were obtained from adolescents (n = 1294), healthy adults (n = 196) and adult psychiatric outpatients (n = 243). Confirmatory factor analyses (CFAs) tested the fit of a one-factor model and a correlated model that comprised three correlated first-order factors (including difficulty identifying feelings, difficulty describing feelings and externally-oriented thinking) to the data. The factorial invariance of the TAS-9-C between independent samples was investigated using multigroup CFAs.

The CFAs results demonstrated a better fit of the correlated three-factor model (CFI > 0.776) over the one-factor model in both adolescent and adult samples (CFI > 0.871). While the correlated model showed evidence of strong factorial invariance among different adolescents from different grade levels, configural invariance was evidenced in the adolescent, healthy adults and adult psychiatric outpatients samples.

The three meaningful clusters of TAS-9-C mirrored the original factor structure of TAS-20. The use of subscales representing different aspects of alexithymia adds incremental value to the measurement using full scale alone among Chinese across developmental stages and settings. Configural invariance among these samples suggests that growth-related change leads to different developmental trajectories for adult alexithymia. Future research is necessary to ascertain this.