The power of a big TED talk

Tom Loncar's article on power posing research (June edition) gave a useful overview of the work in this area. However, I feel that some important points were neglected.

We really ought to question the role that TED talks play in giving global exposure to new ideas that have received little peer appraisal and which may not have been the subject of replication attempts. In early 2017 I had just moved from academia to a not-for-profit organisation, where my line manager was advising his staff to adopt power poses. As the only psychologist present, I was the only person aware that this work had just become controversial.

Loncar does not refer to the content of Cuddy's 2012 TED talk, in which she told her audience that a minor 'tweak' to their posture could 'significantly change the way your life unfolds' – an overblown claim even by the terms of the research that she had published by that point. Furthermore, the only dependent variables mentioned in Cuddy's TED talk were risk tolerance, testosterone and cortisol. In the other research cited by Loncar, no support was found for the impact of power posing on any of these variables. The subjective feeling of power, although included in Cuddy's published work, was apparently not of sufficient importance to be included in her talk.

Cuddy's talk also referred to research on 'facial feedback', which had found that people's affective responses became more positive after they had inserted a pencil between their teeth to create an artificial smile. Unfortunately, a 17-lab registered replication study in 2016 failed to reproduce this finding. Whilst some subsequent work suggests that the inclusion of camera recording had obliterated the facial feedback effect, if true, that merely shows how fragile the effect is.

Loncar notes that Cuddy's former co-researcher Dana Carney has distanced herself from their power posing work. What he doesn't mention is that Carney also described the various ways in which the original work with Cuddy was p-hacked (this happened before the discipline as a whole began to fully realise the problem of 'questionable research practices'). Carney also states that risk-taking was the primary dependent variable, whereas 'The self-report DV was p-hacked in that many different power questions were asked and those chosen were the ones that "worked."'

Loncar reports that Cuddy feels 'vindicated' by more recent work which shows that self-reported feelings of power can be impacted by postural manipulations. But this is surely clutching at straws, given that it is the avoidance of contractive postures that appears to have the impact rather than the use of the expansive postures that Cuddy promoted in her TED talk. This is, of course, useful knowledge but barely on the same scale as the original claims. It's akin to the finding – as reflected in NHS guidance – that taking vitamin supplements isn't really of any benefit unless you are actually deficient in a particular vitamin or vitamins (and possibly a better diet would be more helpful anyway).

The power pose isn't the only recent heavily hyped idea which has turned out to be controversial. Carol Dweck's (2006) theory of growth mindset – also a TED talk subject – has run into the same problem of mixed results from other researchers, and with a similar apparent resolution: recent work indicates that it is 'at-risk' children and low-SES children who may get some benefit from a growth mindset intervention, not all children. And it's not even clear if adults can benefit at all. A somewhat lower-profile idea, workplace resilience training, is plagued by a particular lack of clarity, with a range of different interventions and dependent variables reported in the literature, many of which return non-significant results. This hasn't stopped some researchers talking enthusiastically about such training.

Unfortunately, in a highly competitive market for jobs and promotions, where one's prospects may be boosted by prominent media exposure, many academics have learned that they may benefit from the slick presentation of even the most tentative early research findings. But so long as academics are afflicted by the perverse incentives that encourage such behaviour, they also cannot be too shocked if they find themselves under a critical spotlight by flagging up their work in this way. If the discipline wants to get itself out of the mess it is currently in, then it needs to fully embrace the rigours of pre-registration, power analysis, data-sharing, adversarial collaboration, and so on.

David Hardman
London