Measures of Relationship Obsessive-Compulsive Symptoms
A quick search on Google would show the term ROCD has been frequently used in the last several years mainly on peer-support OCD forums. Systematic research, however, requires precise definitions and valid measurement tools. Recently, two measures were developed and validated for this purpose: the Relationship Obsessive-Compulsive Inventory (ROCI), assessing relationship-centered OC symptoms (Doron, Derby et al., 2012a), and the Partner-Related Obsessive-Compulsive Symptoms Inventory (PROCSI), assessing partner-focused OC symptoms (Doron, Derby et al., 2012b). In accordance with recent evidence that OCD symptoms are better conceptualized in terms of dimensions rather than categories (e.g., Haslam, Williams, Kyrios, McKay, & Taylor, 2005; Olatunji, Williams, Haslam, Abramowitz, & Tolin, 2008), we designed the ROCI and the PROCSI to assess relationship-centered and partner-focused symptoms on a continuum, from mild preoccupation to severe and debilitating disorder. Our references to ROCD symptoms throughout this paper correspond to this dimensional view.
The ROCI was constructed to measure the severity of obsessions (i.e., preoccupation and doubts) and compulsions (i.e., checking and reassurance seeking) on three relational dimensions: one’s feelings towards a relationship partner (e.g., “I continuously reassess whether I really love my partner”), the partner’s feelings towards oneself (e.g., “I continuously doubt my partner’s love for me”), and the “rightness” of the relationship (e.g., “I check and recheck whether my relationship feels right”). Findings supported this three-factor structure above and beyond two alternative measurement models, but also suggested the existence of a higher-order general factor for relationship-centered OC symptoms. The ROCI performed well on most goodness of fit indices, and the total and subscale scores were highly reliable (Doron, Derby et al., 2012a).
The PROCSI was designed to measure obsessions (i.e., preoccupations and doubts) and neutralizing behaviors (i.e., checking) focused on the perceived flaws of one’s relationship partner in six character domains: physical appearance, sociability, morality, emotional stability, intelligence, and competence. Findings supported this six-factor structure above and beyond alternative measurement models, but also suggested the existence of a higher-order general factor for partner-focused OC symptoms. The PROCSI’s total and subscale scores were found to be internally consistent and had good test-retest reliability (Doron, Derby et al., 2012b).
ROCI and PROCSI scores seem to discriminate between ROCD and other OCD symptoms. In an ongoing study, we compared differences in ROCI and PROCSI scores between 17 clients presenting ROCD themes and 18 clients presenting other OCD themes. We also used the Mini International Neuropsychiatric Interview (MINI; Sheehan et al., 1998) to attain clinical diagnosis. Findings so far show significant differences between the two groups on the ROCI, F(1, 33) = 10.28, p =.003, ŋ² = .24, and the PROCSI, F(1, 33) = 5.42, p =.026, ŋ² = .14. ROCD clients scored higher on the ROCI (M = 2.10, SD = 0.67) than clients presenting other OCD symptoms (M = 1.16, SD = 1.02). This difference remained significant when controlling for severity of OCD and depression symptoms. Similarly, ROCD clients scored higher on the PROCSI (M = 1.33, SD = 0.56) than clients presenting other OCD symptoms (M = 0.78, SD = .79). Again, this difference remained significant when controlling for severity of OCD and depression symptoms. Thus, ROCD symptoms, as measured by the ROCI and the PROCSI, seem to be conceptually and empirically differentiated from other OCD symptom dimensions.
Nevertheless, as the ROCI and PROCSI are designed to assess obsessive-compulsive phenomena, small to moderate correlations are expected between these measures and tools assessing other OCD symptoms. Indeed, we have found moderate correlations between the ROCI and the Obsessive Compulsive Inventory-Revised (OCI-R; Foa et al., 2002). Specifically, the ROCI total score was moderately correlated with the OCI-R total score (r = .45) and subscale scores (rs ranged from .28 for neutralizing to .47 for obsessions; Doron, Derby et al., 2012a). Similarly, small to moderate correlations were found between the PROCSI total score and the OCI-R total score (r = .44) and subscale scores (rs ranged from .28 for ordering to .40 for obsessions; Doron, Derby et al., 2012b).