Psychometric properties of the Chinese Self-Compassion Scale and its short-form: Insights on pathways to mental well-being

Abstract

Background: Self-compassion is receiving increasing research attention. There is lack of consensus on a paucity of psychometrically robust measures of self-compassion. The only self-report measurement of self-compassion is the Self-Compassion Scale developed by Neff (2003). Existing validation studies showed diverged opinions on its psychometric structure. Neff (2003) proposed a 6-factor structure with a composite score; while López et al. (2015) proposed a two-factor model, formed by the positively and negative formulated items respectively. The current study examines the factor structure, reliability, and construct validity of the SCS and its short-form in a community sample. Methods: Two independent samples of 455 and 292 Chinese adults were recruited via convenience sampling. In addition to the Chinese SCS, the first sample completed a battery of mental health measures. Factor structure of the SCS and SCS-SF were evaluated using confirmatory factor analyses. Convergent and divergent validity of the SCS were examined via the relationships of the SCS factors with covariates and mental health measures in a path analysis. Results: The 6-factor correlated model and the 6-factor model with 2 higher-order factors provided adequate fit to the SCS and cross-validated well in an independent sample. The two higher-order factors of the positively formulated “Self-Compassionate Behaviors” and the negatively formulated “Self-Uncompassionate Behaviors” showed good internal consistency and test-retest reliability. The factor “Self-Compassionate Behaviors” significantly predicted hope, perspective taking, empathic concern, and positive affect while the “Self-Uncompassionate Behaviors” factor significantly predicted personal distress, anxiety, depression, general health problems, and negative affect. For the SCS-SF, the 2-factor correlated model but not the 1-factor model was found to provide a good fit in terms of model fit and parsimony. Conclusion: This is the first study to assess the factor structure of the Self-Compassion Scale in the non-Western world, utilizing a non-clinical community sample. The present study suggested a two second-order factor structure for the Self-Compassion and its short-form. The structure with the best fit was comprised of the six first-order factors, and two second-order factors of “Self-Compassionate Behaviors” and “Self -Uncompassionate Behaviors”, with distinct pattern of correlations with other psychological measures. The findings were consistent with evidence in neuroscience suggesting that the default threat-defense system and the self-soothing system are two separate systems in mood regulation. Such differentiation would allow the SCS to be adopted as both an assessment of strengths and vulnerabilities to well-being.