Abstract
In 2010, we published an article in which two experiments demonstrated that expansive (vs. contractive) nonverbal displays produced subjective feelings of power and increased risk tolerance (Carney, Cuddy, & Yap, 2010). One of these experiments demonstrated that such displays increased subjective feelings of power, risk tolerance, and testosterone, and decreased cortisol. Our two experiments were the eighth and ninth to be reported in the literature on the embodied effects of nonverbal expansiveness—seven experiments on this topic were published prior to 2010. Since our article in 2010, 24 additional experiments on the effects of expansive postures have been published (see Table 1). Embodiment and the long-standing discussion of mind-body connection has its experimental roots in William James’s (1890/1950) theories of emotion and ideomotor action. Since then, many studies have demonstrated the bidirectional link between nonverbal behavior and human thought and feeling (see Laird & Lacasse, 2014). One such study was conducted by Ranehill et al. (2015), who reported a conceptual replication of one of our experiments: They found an effect of expansive posture on subjective feelings of power, but no effect of posture on risk tolerance, testosterone, or cortisol.
Comparison of Studies on the Effects of Nonverbal Expansiveness
Note: All results reported were significant unless specified otherwise; comparisons between nonverbal expansiveness versus contractiveness (or neutral control posture). Reports demonstrating causal effects of other power- and pride-related nonverbal behaviors were excluded (e.g., making fists, pride postures, crossing arms, tilting the head up, making an angry face, lowering the voice). Also excluded were the hundreds of published experiments on effects of expanded body posture as an expression of power or dominance and on effects such as perceptions, attributions, and social interaction.
We offer four comments that we hope elucidate the similarities and differences among the 33 published experiments (harvested from the literature through extensive keyword searches and cross-referencing of published articles) and the newly published research of Ranehill et al. We also highlight the specific differences between our experiment and that of Ranehill et al. Unpublished findings were excluded in Table 1. Ranehill et al.’s commentary, with the review presented here, serves as an excellent springboard for identifying potential moderators of the psychological effects of nonverbally expansive (vs. contractive) posture.
Thirty-Three Published Results on Expansive Posture
In Carney et al. (2010), we reported that nonverbal expansiveness (vs. contractiveness) increased subjective feelings of power, risk taking, and testosterone, whereas it decreased cortisol. Using a conceptually similar paradigm (see differences in Table 2), Ranehill et al. reported no effect of nonverbal expansiveness on risk taking, testosterone, or cortisol—only an increase in subjective feelings of power. Prompted by Ranehill et al.’s commentary, we list in Table 1 all published tests (to our knowledge) of expansive (vs. contractive) posture on psychological outcomes. The work of Ranehill et al. joins a body of research that includes 33 independent experiments published with a total of 2,521 research participants. Together, these results may help specify when nonverbal expansiveness will and will not cause embodied psychological changes.
Comparison of Ranehill et al. (2015) and Carney, Cuddy, and Yap (2010)
Differences Between the Ranehill et al. and Carney et al. Studies
Table 2 lists the methodological differences between the Ranehill et al. (2015) and Carney et al. (2010) studies. The summary of the literature reported in Table 1 suggests that all significant results were obtained using paradigms with complex, detailed cover stories when participants were unaware of the hypothesis of the experiment, which suggests that awareness of the hypothesis may be a moderator. And many, but not all, significant results were obtained with paradigms situated in a social context, which suggests social context as a moderator. By “social context,” we mean there was either a social interaction with another person (e.g., participant or experimenter) during the posture manipulation or participants were engaging in a real or imagined social task. Indeed, Cesario and McDonald (2013) found direct evidence that social context (present vs. absent) moderated the effect of expansive posture such that effects were found only when the participant was in a social context.
Tables 1 and 2 taken together suggest that there are three differences between Ranehill et al.’s research and our previously published experiment that may account for the varied results. First, in our two experiments, we were careful to conceal experimental purpose with a detailed cover story; in their experiment, Ranehill et al. told participants the purpose of the study—to investigate effects of posture on hormones. Second, our two experiments involved a social task during the postural manipulation; Ranehill et al.’s experiment did not. Finally, in our experiments, we used postural manipulations that were comfortable, easy, and short in duration; Ranehill et al.’s experiment employed postures that were three times as long as those reported in our 2010 paper.
Contributions of Ranehill et al
Some of the variables listed in Tables 1 and 2 suggest future directions for research. One key moderator may be awareness of the hypothesis of the experiment; virtually all of the published reports demonstrating significant effects of expansive posture used elaborate cover stories to distract participants from the goal of the experiment. As is common in economics research, Ranehill et al. did not use any deception in the experiment, and participants were told that the study examined how physical position affects hormone levels and behavior. Investigating the effects of awareness of what one is doing seems like an interesting and useful avenue for future research—one with practical implications.
Another avenue for future research is the length of time participants hold the expanded posture. Time in posture was rarely reported and is not listed in Table 1. In extensive pilot testing, we had settled on 1 min for each of two postures because longer expressions of the feet-on-the-desk pose were uncomfortable and difficult if held longer than 1 min. Ranehill et al. (2015) tripled the amount of time participants held all postures—including the uncomfortable ones. Although it may make intuitive sense that longer time in the posture would increase effects, holding some postures for too long may cause discomfort, become awkward, or habituate a body to the effects of the posture. Length of time in posture should be directly tested.
Finally, the experimenters’ blindness to the experiment’s hypotheses was impossible to determine from most articles and is not listed in Table 1. Ranehill et al. used experimenters blind to the hypothesis, and we did not. This is a critical variable to explore given the impact of experimenter bias and the pervasiveness of expectancy effects.
Looking Forward
Although we hope that Tables 1 and 2 will assist in moving forward the study of nonverbal expansiveness, at present, direct replications are needed of Carney et al. (2010) and many of the other reports in Table 1. Note that in other disciplines, such as human physiology, similar results as those we obtained have shown that holding an expansive yoga-style pose for 2 to 3 min significantly increases blood-serum levels of testosterone and decreases blood-serum levels of cortisol (Minvaleev, Nozdrachev, Kir’yanova, & Ivanov, 2004). For the purposes of a direct replication of Carney et al., all materials can be obtained from the first author or downloaded from her Web site (http://faculty.haas.berkeley.edu/dana_carney/PRS%20Materials%20-%20to%20replicate.zip).
Footnotes
Acknowledgements
We would like to thank the following colleagues for providing helpful corrections and suggestions to this manuscript: Jill Allen, Manzarin Banaji, R. Bhaskar, Vanessa Bohns, Joe Cesario, Adam Galinsky, Li Huang, Johannes Michalak, Mike Norton, Brian Nosek, Lora Park, Jeff Pfeffer, Simone Schnall, Peter Strelan, Milica Vasiljevic, Mario Weick, Keith Welker, Scott Wiltermuth, Brian Winters at the Association for Psychological Science, and members of the Social and Nonverbal Behavior Lab at the University of California, Berkeley.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.