In this article, we describe a pilot study that investigated the effectiveness of group cognitive behavioral therapy (CBT) for auditory hallucinations. Eleven inpatients with either chronic schizophrenia or schizoaffective disorder participated in 2 CBT groups of differing treatment duration (i.e., 7 versus 20 sessions). The results showed that participation in both groups was associated with significant positive changes in the participants' beliefs about their voices and with a trend for reduced negative reactions to the voices. These changes were not a function of premorbid cognitive functioning. Finally, duration of treatment did not affect participants' beliefs or distress associated with the voices. Implications for future clinical research in this area are discussed.