The Psychological Record, 1978, 28,123-128. ATTRIBUTION AND INSOMNIA: A REPLICATION FAlLURE DA VID HEFFLERl and STEPHEN A. LISMAN State University 0/ New York at Binghamton We attempted to replicate the negative placebo effect described by Storms and Nisbett (1970), whereby placebos described as arousal pills or relaxing pills resulted in reports of decreased or increased sleep onset latencies, respectively. Design improvements included lengthening both assessment and "treatment" periods and balancing instructions for both groups. Our negative results, consistent with other recent replication failures, suggest the need for caution in the c1inical application of attribution theory. Comprehensive summaries of current clinical and research activity in the field of behavior therapy have traced clearly the development and usage of treatments derived from attribution theory (e.g., Goldfried & Davison, 1976; Kopel & Arkowitz, 1975; Valins & Nisbett, 1971). Such reviews typically note that, while attributional manipulations and subsequent behavior change have been the focus of numerous laboratory analog studies, a particular study by Storms and Nisbett (1970) represents the first effort to utilize attributional manipulations for clinically relevant behavior change. Indeed, we believe that Storms and Nisbett's success at diminishing reported latency to fall asleep in insomniacs represents a cornerstone study in this field, since it so strongly encouraged therapists to begin considering seriously the potential of attributional approaches to clinical problems. To summarize briefly, Storms and Nisbett (1970) simply construed insomnia as a problem affected by labelling bedtime arousal as anxiety or sleeplessness. Under the ruse of assessing dreams, they recruited a group of insomniac subjects, administered a placebo pill to each one, and told them one of two possible types of side-effects the pill might have. Half the subjects were told that the pill would arouse them, and half that it would relax them. Consistent with their predictions, Storms and Nisbett found that those given arousal instructions reported decreased time to fall asleep. The authors concluded that (a) improved subjects had benefited from the opportunity to re-attribute their arousal to the pill, whereas (b) those who worsened had assumed that their emotions (arousal) were unusually intense, since they couldn't fall asleep even after taking an arousal-reducing agent. Considering the importance of Storms and Nisbett's study, it is notable that for five years following its appearance, no replication appeared in the literature. Then, in an effort to examine alternative explanations for the "reverse placebo" effect, Kellogg and Baron (1975) used Storms and 'Now at WilJard Psychiatrie Center, WilJard, N.Y. 14588. Reprint requests should be addressed to Stephen A. Lisman, Department of Psyehology, SUNY ·Binghamton, Bingharoton, New York 13901. 0033-2933/78/1300/0123$00.10 © 1978 The Psychological Record