False feedback about placebo effectiveness: Consequences for the misattribution of speech anxiety

Abstract

Emotional reactions can be reduced by giving subjects a placebo to which they can misattribute arousal symptoms. Some emotional states have yielded disappointing results in misattribution research, however, including speech anxiety. To resolve the inconsistent findings, M. Ross and J. M. Olson (1981, Psychological Review, 88, 408–437) proposed an expectancy-attribution model of the effects of placebos, which identified several preconditions for misattribution. The present experiment tested whether subjects would misattribute speech anxiety when these requirements were satisfied. Subjects read a speech in front of a camera while allegedly being exposed to "subliminal noise". Subjects were either told that subliminal noise makes people feel unpleasantly aroused or pleasantly relaxed. Immediately prior to reading the speech, subjects were provided with bogus feedback indicating that the subliminal noise was having either a weak or a strong impact on their symptoms. Results showed that arousing noise subjects who received feedback that the noise was having a strong impact on their symptoms made fewer speech dysfluencies while reading the speech than arousing noise subjects who received feedback that the noise was having little impact. Arousing noise subjects in the strong impact condition also made fewer speech errors than relaxing noise subjects in the strong impact condition. Finally, within the arousing noise groups, subjects' perceptions of the impact of the noise on their symptoms correlated significantly with their rate of speech errors. These findings support the expectancy-attribution model of placebo effects.