A randomized trial was performed in which imaginal exposure (IE) and cognitive therapy (CT) were compared in the treatment of chronic posttraumatic stress disorder (PTSD). Patients who continued to meet PTSD caseness at the end of a 4-week symptom-monitoring baseline period (n = 72) were randomly allocated to either IE or CT. There was a significant improvement in all measures over treatment and at follow-up, although there were no significant differences between the 2 treatments at any assessment. A significantly greater number of patients who showed worsening over treatment received IE, although this effect was not found at follow-up. Patients who worsened showed a greater tendency to miss treatment sessions, rated therapy as less credible, and were rated as less motivated by the therapist. It was concluded that either exposure or a challenge to cognition can result in symptom reduction, although neither resulted in complete improvement.