Feel confident walking into that interview! Raise your hands up, hold your chest high, and, whatever you do, don’t cross your arms as you wait. You want to make sure the pose you’re striking inspires confidence and communicates that confidence to your (hopefully) future employer.
We’ve all received this advice and probably incorporated it into our store of common knowledge at some point or another. Perhaps we’ve actively used it during a big meeting or important presentation. Much credit is owed to a 2010 paper on “power posing,” which found that “posing in high-power nonverbal displays (as opposed to low-power nonverbal displays) would cause neuroendocrine and behavioral changes for both male and female participants.” In other words, assuming a specific stance can cause you not only cause you to behave more confidently, but also induces an actual change in chemical reactions in the brain. The authors also concluded that high-power posers felt more powerful and were more likely to take risks. In academic parlance, they asserted that these stances caused “advantaged and adaptive psychological, physiological, and behavioral changes.” The purported phenomenon spawned a TED talk from one of the paper’s authors, Amy Cuddy, whose captivating words helped disseminate their ideas into popular culture. Sounds like some pretty great results from just shifting one’s body a bit!
Well, not so fast. It turns out one of the co-authors, Dana Carney, declared in no uncertain terms earlier this year that she does not “think power poses are real.”
In fact, Carney posted a manifesto on her website explaining her current view of the original paper, Carney reports a wave of evidence that has convinced her of serious confounders in the study. Chief among them: the sample size of 42 was “tiny,” and the “data are flimsy. The effects were small and barely there in many cases.” Uri Simonsohn, a behavioral scientist at the Wharton School of Business, also spoke to NPR about the 2010 study, specifically how its shortcomings demonstrate how the standards for rigorous scientific evaluations have changed over the years. He too expressed concern for the small sample size and how priors can infect the integrity of an intervention. Carney wrote further that certain research decisions seem in retrospect like “p-hacking,” or manipulating variables (even if unintentionally) to achieve a desired level of statistical significance.
But that’s not all! Carney goes on to outline the “obviously clear” confounds in the original paper, which are the main reason she now disavows the six-year-old findings. (She also claims not to have spoken about power poses for five years, since she first became skeptical of their significance). First, both experimenters involved with the experiment were aware of the hypothesis that power poses would increase feelings of authority and a taste for risk-taking (in this experiment, the risk-seeking was a gambling task that participants could win or lose). The testosterone effect, which is the one related to power, was measured after risk-seeking participants were told they had won. Other research, though, shows that winning itself increases testosterone levels. So, the observed elevation might have been a product of #winning, not posing.
The now infamous study did randomize which participants were assigned to “power poses” and which were not. Carney’s current views on the study highlight some flaws in the methodology, namely the need for rigor even when randomization is used. Cuddy was surprised that Carney chose to discount the study on account of its flaws, explaining “[l]ike all scientists, I understand that my field evolves as new evidence replicates some effects and not others. My scientific evaluation of the various effects will update as new information comes in.”
Notwithstanding Cuddy’s laudable view, there is almost no denying the problems with the original paper. From sample size to administrative integrity to statistical measures, it was plagued by methodological shortcomings that can’t save any RCT. Relying on one flawed, even if randomized, study cannot be sufficient, even if the results “feel” right. Thus, Cuddy might follow her co-author’s lead and update her priors more definitively. That would be an intellectual power pose worth embracing.