Applying Generalizability Theory to the Self-Compassion Scale to Examine State and Trait Aspects and Generalizability of Assessment Scores

Abstract

Self-compassion has become increasingly recognized for its benefits to human health and well-being, but continued work is required to achieve its accurate assessment in the context of research and psychological treatment. The Self-Compassion Scale (SCS) is widely referred to as a trait measure, although the scale’s ability to differentiate between state and trait has not been thoroughly examined using suitable methods. The aim of this study was to use generalizability theory (G Theory) to distinguish state and trait aspects and evaluate generalizability of assessment scores. The present study utilized G Theory in a longitudinal nested design, using an existing dataset of Australian-based educators (n = 124) randomized into intervention (mindfulness/self-compassion) and control group and assessed using the SCS-Short Form (SCS-SF) at baseline, after 8 weeks, and after 6 weeks. The robust psychometric properties of the SCS-SF as a trait measure were confirmed by strong temporal stability and generalizability of scores across occasions and sample population (G = 0.89) with only a minor proportion of state-related variance (SCI = 0.11). The majority of individual items clearly measured self-compassion as a trait. The study findings unequivocally demonstrated that the SCS-SF is a valid and reliable measure of trait self-compassion with scores generalizable across sample population and occasions. If significant changes of the SCS-SF scores are observed in individuals as a result of intervention, such changes are likely to be long lasting.