The Cognitive Therapy Scale is a well-established tool for assessing skills in delivering cognitive therapy, but has been subject to criticism. It has recently been updated by two groups, producing the Revised version or CTS-R (Blackburn, James, Milne and Reichelt, 2000) and a version designed for therapy of psychosis, the CTS-Psy (Haddock et al., 2001). The present study made a direct comparison of these scales to evaluate their inter-rater reliability, the extent to which they measure the same therapist qualities, and their utility for assessment of skills in trainee therapists working with different client-types. Twenty-six trainees submitted tapes of therapy with clients suffering either personality disorder or psychosis. Each tape was rated by two independent assessors on each of the two scales. Results suggest the scales are both fairly easy to use and produce highly similar estimates of student competence. Client diagnosis has no significant influence on the scores obtained by the therapist. However, inter-rater reliability is relatively low. It is concluded that safeguards are needed where these scales are used as a training outcome measure.