Psychometric validation and measurement invariance of the self-compassion scale-short form (SCS-SF) across gender, clinical population, and cultures.

BACKGROUND

The concept of self-compassion, rooted in Eastern philosophies, is closely related to psychological well-being and is considered a key skill in alleviating psychological disorders. The short form (SCS-SF) of Self-Compassion Scale was developed and used by many scholars. The current study aimed to conduct an adaptation study of the SCS-SF in Turkish and to analyze its psychometric properties in clinical and non-clinical samples.

METHODS

In the adaptation process, the confirmatory factor analysis (CFA) for the general population (n = 545) and clinical population (n = 246) was conducted. We also investigated evidence for convergent validity (n = 274) and conducted a test-retest reliability study with a sample of 53 participants. Additionally, we evaluated the measurement invariance of the SCS-SF across different gender groups, clinical versus non-clinical samples, and cross-cultural samples (comparing Türkiye and the USA, with a sample size of 125 in the USA).

RESULTS

Confirmatory factor analysis showed a six-factor model for the non-clinical sample, yielding fit indices of TLI = 0.920, CFI = 0.953, and RMSEA = 0.095. Also, the six-factor model exhibited a good fit for the clinical sample, with TLI = 0.968, CFI = 0.981, and RMSEA = 0.045. Reliability analysis indicated Cronbach's alpha values were between 0.77 and 0.88, alongside a strong test-retest reliability coefficient of r = 0.878 (p < 0.001). Furthermore, measurement invariance was confirmed across clinical and non-clinical groups, as well as across gender groups and among Turkish and American cultural contexts, thereby affirming the scale's applicability across varied populations.

CONCLUSION

The adapted SCS-SF scores demonstrated good reliability. The CFA showed a six-dimensional structure and there was evidence for convergent validity and measurement invariance across gender groups as well as between samples from the USA and Türkiye. Consequently, the current study indicates that the adaptation was successful, and the Turkish version can be used for both clinical and non-clinical samples. However, limitations include the reliance on self-report measures and potential cultural bias in the interpretation of self-compassion constructs.