A growing body of research indicates that cognitive-behavioral (CB) treatments are superior to no treatment for panic. However, few compelling differences exist between treatments, questioning whether the addition of CB elements is necessary for effective treatments. The current study examined two questions. First, does the addition to treatment of CB elements enhance outcome? Second, can a more general construct such as self-efficacy account for treatment outcome? Nineteen subjects with panic disorder were assigned to one of two treatments (panic education or guided imaginal coping). There were minimal differences in outcome between groups. Subjects became more efficacious and efficacy appeared to lead to reductions in catastrophic thoughts, but not panic symptoms. Results are discussed in terms of the potential role of personal control in the treatment of panic and in terms of the necessity to conduct more powerful comparisons to clarify the role of CB strategies in the treatment of panic. © 1991 Plenum Publishing Corporation.