Failure of a "misattribution therapy" manipulation with a clinically relevant target behavior

To test the effects of attribution manipulations on a naturally occurring and clinically relevant problem behavior, 60 speech anxious subjects presented two speeches. Meaningless noise was presented during the first speech and absent from the second. Subjects were randomly assigned to three attribution conditions. Prior to the first speech, subjects in the Arousal condition were told that the effect of the noise was to increase physiological arousal. Control subjects were given no instructional set regarding the physiological effects of the noise. Sedation subjects were informed that the noise would have a decremental effect on physiological arousal. The results, which fail to replicate a reverse placebo effect previously reported in the attribution literature and support a placebo interpretation, especially among highly fearful subjects, suggest that successful misattribution is unlikely when subjects have had an adequate history of experience with the feared situation.