Psychometric properties and longitudinal invariance of the session rating scale in Chinese clinical samples

Abstract

The Session Rating Scale (SRS) is a widely used clinical tool to measure the client-therapist working alliance. This study investigated the psychometric properties, the cut-off value, and longitudinal invariance of the SRS in a Chinese clinical population. The analyses were conducted separately in a sample of college students in counseling (n = 403) and in a sample of clients in outpatient therapy (n = 246). Participants completed the Outcome Rating Scale (ORS) before each session and the SRS after each session. To test concurrent validity, a subset of participants also completed the Working Alliance Questionnaire at the end of each session. In both samples, the results indicated high internal consistency (α = .92 ~ 0.97) and adequate test-retest reliability over four sessions (university sample: r = .69 ~ .78; outpatient sample: r = .52 ~ .66). Construct validity was evident in a one-factor structure, and concurrent validity was established based on a strong correlation with the Working Alliance Questionnaire (r = .64 ~ .70). In addition, regression analysis indicated that early alliance ratings (at the third session) on the SRS predicted post-ORS scores. The cut-off value for the SRS in the Chinese context was established as 34. The longitudinal measurement invariance was tested by a longitudinal confirmatory factor analysis. Full scalar invariance of the SRS was supported. This study supported the use of the SRS in China and that a single-factor solution stayed stable over time, providing preliminary evidence for subsequent mean comparisons.