Self-Compassion and Obsessive–Compulsive Symptoms in a Non-Clinical Sample: Insights from Network Analysis and Canonical Correlation Analysis

Abstract

Abstract Objectives The goal of the present study was to investigate the hypothesized complex interactions between dimensions of self-compassion and obsessive-compulsive symptoms (OCD) in a non-clinical sample. Method Four hundred and twenty-four individuals (273 women) ranging in age from 18 to 83 years (M = 41.00; SD = 14.50) participated in the cross-sectional survey. The participants completed the Self-Compassion Scale (SCS) and the Obsessive-Compulsive Inventory—Revised (OCI-R). Results Certain dimensions of self-compassion were found to be associated with specific OCD symptoms, with positive dimensions (except Common humanity) correlating with reduced Obsessing, Hoarding, and Ordering, and negative dimensions correlating with all OCD symptoms. Canonical correlation analysis indicated three significant linkages: (1) between Isolation, Over-identification, and Obsessing, (2) between Self-judgment, Over-identification, Ordering, and Checking, and (3) between Self-judgment and Neutralizing. Exploratory graph analysis challenged the structure of the SCS and the OCI-R, suggesting two dimensions of self-compassion, compassionate self-reference and uncompassionate self-reference; and two dimensions of OCD, obsessing and other OCD symptoms. The Obsessing subscale appeared as a bridge between uncompassionate self-reference and OCD symptoms. Compassionate self-reference was negatively correlated with uncompassionate self-reference and obsessing, whereas uncompassionate self-reference was positively correlated with obsessing and OCD symptoms. Conclusions Investigation of the associations between self-compassion and psychopathology should avoid aggregating the uncompassionate and compassionate dimensions of self-compassion. The significant associations in the network of self-compassion and OCD symptoms were between uncompassionate self-reference, obsessing, and OCD symptoms. Preregistration This study is not pre-registered.