LEARNING TO DETECT DECEPTION FROM THREE COMMUNICATION CHANNELS Miron Zuckerman Richard Koestner Michele J. Colella ABSTRACT: A Lie Detection Test--a videotape of senders delivering truthful and deceptive messages--was administered to subjects via one of three channels: Face only, speech only, or face plus speech. Feedback information, identifying the messages as truthful or deceptive, was given to some subjects (learning con- dition) but not to others (control). It was found that subjects in the learning con- dition performed better overall relative to control subjects, regardless of the channel presented to them. In addition, however, accuracy of subjects in the learning condition improved progressively over the course of the test (relative to control), for the speech only and face plus speech channels but not for the face only channel. This effect was interpreted in terms of the limited number of deception cues offered by the face. In a previous study, Zuckerman, Koestner and Alton (1984) examined whether individuals can learn to become more accurate lie detectors. Several learning procedures were tested. Specifically, subjects received information identifying messages as truthful or deceptive, either after they recorded their judgments of a particular item or before the itemwas presented. The number of items for which information was provided was also varied. In general, the results showed that the more information given to sub- jects about deceivers' messages, the more accurate the detection of lies enacted by the same deceivers. This effect was not moderated by whether the information was given before or after the items were presented. On the other hand, there was no evi- dence that the increase in accuracy generalized to accuracy of Requests for reprints should be sent to Miron Zuckerman, Department of Psychology, University of Rochester, Rochester, NY 14627 Journal of Nonverbal Behavior 9(3), Fall 1985 1 88 © 1985 Human Sciences Press