BACKGROUND: Cognitive behavioural therapy has been established as an effective treatment for residual psychotic symptoms but a substantial proportion of people do not benefit from this treatment. There has been little direct study of predictors of outcome, particularly in treatment targeting auditory hallucinations. METHOD: The Psychotic Symptom Rating Scales (PSYRATS) and Positive and Negative Syndrome Scale (PANSS) were administered pre- and post-therapy to 33 people with schizophrenia-related disorders receiving CBT for auditory hallucinations in a specialist clinic. Outcome was compared with pre-therapy measures of insight, beliefs about the origin of hallucinations, negative symptoms and cognitive disorganization. RESULTS: There were significant improvements post-treatment on the PSYRATS and PANSS Positive and General Scales. Improvement on the PSYRATS was associated with lower levels of negative symptoms, but was unrelated to overall insight, delusional conviction regarding the origins of hallucinations, or levels of cognitive disorganization. CONCLUSIONS: Lack of insight and presence of formal thought disorder do not preclude effective cognitive-behavioural treatment of auditory hallucinations. There is a need to further understand why negative symptoms may present a barrier to therapy.