Assessing alliance ruptures and resolutions: Reliability and validity of the Collaborative Interactions Scale-revised version

Objective: This study presents the revised version of the Collaborative Interactions Scale (CIS) [Colli, A., & Lingiardi, V. (2009). The Collaborative Interactions Scale: A new transcript-based method for the assessment of therapeutic alliance ruptures and resolutions in psychotherapy. Psychotherapy Research, 19(6), 718–734.], an observer-rated measure for the assessment of therapeutic-alliance ruptures and resolutions. Intensive use of the previous scale resulted in three criticisms: (i) excessive time required to perform evaluations, (ii) the low occurrence of some items, and (iii) the scale’s low capacity to capture some patient–therapist interactions in fine detail. In this study, we aimed to describe the scale revision process and evaluate interrater reliability and scale validity by comparing sessions of patients with and without personality disorders (PDs). Method: Three raters conducted a blind evaluation of a sample of 60 sessions (180 segments; 3607 narrative units) with 30 patients (15 had a PD diagnosis and 15 had a DSM-5 clinical syndrome diagnosis without a PD). Results: Interrater reliability results ranged from acceptable to excellent and were comparable to those of the former version. Patients with PDs showed a greater number of alliance ruptures and a smaller number of collaborative processes than patients without PDs. Moreover, therapists presented more negative interventions with the PD sample than with the non-PD sample. Conclusions: The results indicate that the revised CIS is a reliable rating system that is useful for both empirical research and clinical assessments.