The existing scale for assessing competence in cognitive therapy (CTS) dates from 1988 and only the previous version of 1980 has been validated to any extent. A revised version, the CTS-R, was devised to improve on the CTS by: eliminating overlap between items, improving on the scaling system, and defining items more clearly. Kolb's well-known educational model was used as a guideline. In the new 14-item scale, three new items measure general therapeutic flair, the facilitation of emotional expression, and therapist's non-verbal behaviours (optional). We hypothesized that the CTS-R would prove more user friendly and demonstrate satisfactory reliability and validity. Twenty-one mental health professionals undergoing training in cognitive therapy provided 102 video-tapes of therapy with 34 patients, reflecting three stages of therapy. The tapes were rated by four expert raters, in a balanced design. The CTS-R showed high internal consistency and adequate average inter-rater reliability. Reliability for individual items varied widely among pairs of raters. Validity was demonstrated by improved ratings of competence for trainees who saw patients early and later during the course of training. Although raters found the CTS-R a more useful tool than the CTS and satisfactory reliability and validity were demonstrated, more refinement is needed in item definition. The study has led to modifications in the CTS-R, which are in the process of evaluation.