Journal of Experimental Social Psychology 42 (2006) 18–27 www.elsevier.com/locate/jesp 0022-1031/$ - see front matter  2005 Elsevier Inc. All rights reserved. doi:10.1016/j.jesp.2005.01.005 Attributing the sources of accuracy in unequal-power dyadic communication: Who is better and why?  Judith A. Hall a,¤ , Janelle C. Rosip a , Lavonia Smith LeBeau b , Terrence G. Horgan c , Jason D. Carter d a Department of Psychology, 125 NI, Northeastern University, Boston, MA 02115, USA b The Pennsylvania State University, USA c The Ohio State University, USA d State University of New York, New Paltz, USA Received 14 July 2003; revised 2 November 2004 Available online 23 March 2005 Abstract A study was conducted to assess accuracy of deliberate nonverbal communication of aVective messages between individuals assigned to diVerent power roles within dyads. In phase 1, participants (N D 158) were assigned to unequal- or to equal-power roles and asked to send positive, negative, and neutral messages to their partner using nonverbal cues while the partner guessed which kind of message it was. In phase 2, naïve decoders (N D 294) made judgments of the videotapes from phase 1 to resolve the confounding of sender and decoder factors in the within-dyad communication paradigm. Results showed that subordinates were more accurate at decoding superiors than vice versa, and that this diVerence was due to subordinates sending less clear messages to superiors than superiors sent to subordinates. Comparison with the equal-power group’s expressions revealed that the subordinates’ expressions were also less clear than those sent by the equal-power group.  2005 Elsevier Inc. All rights reserved. Keywords: Accuracy; Power; Expression The sensitivity of people to each other’s nonverbal cues is obviously important in many contexts. Of partic- ular interest in social psychology has been the sensitivity of people with diVering levels of dominance, power, or status. Henley (1977; Henley & LaFrance, 1984; LaFrance & Henley, 1994) proposed that people with low status or low social power are motivated to be, and as a consequence are, especially socially sensitive, a view endorsed in various forms by many other authors (e.g., Crawford & Unger, 2000; Fiske, 1993; Keltner, Gruen- feld, & Anderson, 2003; Matlin, 2000; Snodgrass, 1985, 1992; Worchel, Cooper, & Goethals, 1991; Yoder, 2003). These two hypotheses—that people lacking in social power have superior social sensitivity and that the rea- son lies in their motivation to be sensitive—have seemed compelling to many writers. In fact, however, they are controversial. A number of writers have presented either theoretical arguments or empirical data that are unsup- portive or even contradictory to these hypotheses (e.g., Hall & Halberstadt, 1997; Hall, Halberstadt, & O’Brien, 1997; Overbeck & Park, 2001; Riggio, 2001; Snodgrass, Hecht, & Ploutz-Snyder, 1998). One problem has been that testing the hypotheses has proved methodologically complex. One approach has been to relate indicators of power (such as personality  This article is dedicated to the memory of Sara Snodgrass. This re- search was supported by a grant from the National Science Founda- tion to the Wrst author. The authors are grateful to Sara Eltzroth, Jenna Lavery, Alycia Piccone, and Andrea Sparko for their help in conducting this study, Patricia Noller for sharing her communication task with us and giving permission to adapt it, and Marianne Schmid Mast for providing helpful comments on the manuscript. * Corresponding author. Fax: +1 617 373 8714. E-mail address: hall1@neu.edu (J.A. Hall).