This paper reports the psychometric properties of a CBT for psychosis adherence scale, the Revised Cognitive Therapy for Psychosis Adherence Scale (R-CTPAS). The scale's factor structure, inter-rater reliability and concurrent validity were analysed in a sample of 67 audiotaped sessions of CBT for psychosis. The concurrent validity of the scale was examined through comparison with the Cognitive Therapy Scale (CTS, Young and Beck, 1980). Principal components analysis of the trial data suggested three factors: “engagement/assessment work”, “relapse prevention work” and “formulation/schema work”. Satisfactory levels of inter-rater reliability were established between rater dyads. Moderate correlations with the CTS provided an indication of concurrent validity. The R-CTPAS is concluded to be a reliable and useful instrument that can assess adherence to CBT for psychosis using the Fowler, Garety and Kuipers (1995) therapy model.