Objective: This study compared the practice of cognitive-behavioral therapy ( CBT) for psychosis across research and routine clinical settings. Methods: An observer-rated adherence measure was used to compare the content of 40 therapy sessions of clients with positive psychotic symptoms. Twenty therapist-client dyads came from a research setting in the United Kingdom and 20 from three clinical settings, two in the United Kingdom and one in the United States. In the research setting CBT was provided by research clinical psychologists and trained local therapists. In the clinical settings CBT was part of a case management service by trained therapists. Results: Therapist adherence to CBT for psychosis did not differ between the research and clinical settings. However, clinicians in the research settings scored significantly higher on items for schema work ( z=-1.98, p<.05), relapse prevention interventions ( z=-2.08, p<. 05), and formulating a model of relapse ( z=-2.61, p<. 01). Conclusions: CBT for psychosis conducted in clinical settings was more strongly characterized by assessment of symptoms and work on coping strategies and less so by relapse prevention and schema-level work. Relapse prevention interventions and schema work could be considered more challenging for therapists to undertake. The findings suggest that therapists working in routine clinical settings are able to establish good therapeutic relationships with people with psychosis and to work on assessing and coping with their psychotic symptoms. However, some therapeutic approaches may be more challenging in this context.