Background: There has been considerable acknowledgement in treatment outcome research that, although the assessment of treatment integrity is essential in many respects, it requires great effort as well as resources and is therefore often neglected. Aims: In order to fill this gap, the Cognitive Therapy Competence Scale for Social Phobia (CTCS-SP) was developed, based on the Cognitive Therapy Scale, to measure therapist competence in delivering cognitive therapy for social phobia. The aim of the present study was to investigate interrater reliability, internal consistency and retest reliability of the scale. Method: Raters evaluated therapist competence from 161 videotaped sessions (98 patients) selected from 234 cognitive treatments within a multi-centre study. Results: Interrater-reliability was found to be high for the overall score (ICC = .81) and moderate for individual items (ICC = .62–.92). Internal consistency and retest reliability were also found to be high (Cronbach's alpha = .89; (ICCretest = .86). Conclusions: The results indicate that the CTCS-SP is highly reliable. As even individual items yield satisfactory reliability, the scale can be used in various fields of research, including the measurement of changes in skill acquisition and the impact of competence on outcome criteria.