Variance Composition, Reliability, Validity, and Measurement Invariance of the Toronto Alexithymia Scale-20 in...

Area of Science:

• Psychology
• Psychometrics

Background:

• The Toronto Alexithymia Scale-20 (TAS-20) is widely used in research on men and masculinities and in clinical settings.
• Confirmatory factor analyses (CFAs) have often not supported the scale's intended factor structure.
• Prior work on the TAS-20's psychometric properties requires extension.

Purpose of the Study:

• To assess the dimensionality and reliability of the TAS-20.
• To provide evidence for the validity of the TAS-20 in men.
• To examine measurement invariance of the TAS-20 across genders.

Main Methods:

• A sample of 913 men and women was used to assess variance composition and calculate model-based dimensionality and reliability.
• Convergent and concurrent validity in men (N = 505) was examined using relationships with latent variables.
• Measurement invariance by gender was assessed.

Main Results:

• Evidence for unidimensionality was found, supporting the use of the TAS-20 total score but not subscales.
• Model fit was less than desired.
• Convergent and concurrent validity for the TAS-20 in men was established.
• Metric invariance was found between cisgender men and women.

Conclusions:

• The TAS-20 total score can be supported for use, but subscale use is not recommended based on current psychometric properties.
• The scale is understood similarly by cisgender men and women.
• Scale developers should consider procedures to improve the TAS-20's psychometric properties, particularly model fit.