Los Angeles, 1973. A lecture hall at the University of Southern California School of Medicine. In the seats: psychiatrists, psychologists, educators, graduate students. On the stage, a man introduces himself as “Dr Myron L. Fox,” an expert in the application of mathematics to human behavior. For an hour he delivers a lecture on mathematical game theory applied to physician education. Then he takes questions for thirty minutes.
Dr Fox does not exist. The man on stage is a television actor who read a single article on the topic before walking up to the podium. He was coached to do one thing: speak with warmth, humor, and confidence, while stringing together jargon, invented terms, and statements that contradict each other from one sentence to the next. The lecture is delivered three times to a total of 55 people. Every single audience rates it highly. Nobody detects the hoax (Wikipedia).
What the experiment actually measured
The study appeared in 1973 in the Journal of Medical Education, authored by John Ware, Donald Naftulin, and Frank Donnelly. Their hypothesis: teaching evaluations depend more on the speaker’s personality than on the substance of what is being taught. Law professor Deborah Merritt later summarized the finding by pointing out that the actor’s nonverbal behavior alone was enough to mask a confused, meaningless lecture.
The result now has a name: the Dr Fox effect. Substance still matters, but past a certain threshold of expressiveness, a room full of specialists can rate an empty speech highly, learn nothing from it, and still walk out convinced it was excellent.
I spent my career inside executive committees, at Apple, at Sony, as CTO of Neopost with more than 1,100 engineers reporting up through my organization. I have watched that 1973 lecture hall replay itself hundreds of times, wearing different clothes. My book devotes a full chapter to the biases that govern our judgment in meetings: the halo effect, the authority bias, the pull toward whoever speaks with the most confidence. (My book)
Your leadership meetings are running the same scene
A tall, well-dressed person who speaks clearly gets judged as more competent and more credible by a team. A survey of Fortune Global 500 CEOs found they stand, on average, 6 feet tall, about 2.5 inches above the average American man. Each additional inch of height carries a measurable salary premium, in China and in the United States alike. That is the halo effect at work: physical presence and verbal ease bleed into our judgment of real competence.
A second bias compounds it, the one I call by its American acronym, HIPPO, for “highest paid person’s opinion.” A manager speaks first in a meeting, and the room aligns behind that opinion before anyone has had a real chance to challenge it. The strength of the argument matters less than the rank of whoever is making it.
The fix is procedural, not moral: whoever holds the highest rank in the room speaks last, once every other view has been heard. That single change in sequencing protects a decision from these two biases far better than any appeal to open-mindedness.
Together, these two biases explain something I have watched unfold since ChatGPT arrived: dozens of self-proclaimed “experts” who can produce a striking image in a generation tool, talk fluently about prompts and productivity gains, and have never once carried the responsibility of a transformation inside an organization of several thousand people. They are the Dr Fox of the corporate keynote circuit: the delivery is flawless, the fluency is real, and none of it has been checked against experimentation. Assumptions are dangerous and should be killed through measurement, not applauded because they were well told.
The same mechanism shapes how a company hires a consultant, books a keynote speaker, or evaluates a strategic plan in committee. A 2012 Harvard Business Review study by John Antonakis and his colleagues at the University of Lausanne identified twelve verbal and nonverbal tactics, metaphor, anecdote, rhetorical question, an expressed moral conviction, contrast, that raise how charismatic a leader appears independently of the quality of their decisions. In their experiment, executives trained in these tactics saw their perceived leadership scores rise by roughly 60% in the eyes of their audience, while their actual competence had not moved at all (Harvard Business Review).
An organization gets into trouble the moment it mistakes the pleasure of listening for proof of being right. A weak plan delivered with polish will clear an executive committee more easily than a strong plan delivered badly. That is the exact opposite of what sound governance is supposed to produce.
Voting doesn’t escape the same glaze
If the Dr Fox effect stayed inside lecture halls and boardrooms, it would remain a management footnote. It does not stay there.
In 2005, psychologist Alexander Todorov and his team at Princeton showed that competence judgments made in a single glance at a candidate’s face predicted the outcome of U.S. Senate races better than chance, 68.8% accuracy for the 2004 races, and the accuracy tracked the actual margin of victory (Science). Two years later, with Charles Ballew, he repeated the experiment on the 2006 gubernatorial and Senate races with the same result: under one second of exposure to a candidate’s face is enough to predict roughly 70% of winners, before a voter has any idea what that candidate stands for (PNAS).
The most unsettling version of this work may be the one John Antonakis and Olaf Dalgas published in Science in 2009. They showed more than 680 adults pairs of photographs of winners and runners-up from the 2002 French parliamentary elections, with no other information attached. Competence judgments based on the photograph alone predicted the real outcome with 72% accuracy. The researchers then showed the same photographs to children aged 5 to 13 and simply asked them who they would pick as captain of their boat for a game. The children’s choices tracked the adults’ almost exactly (Science).
Plato already sensed this in The Republic, comparing the public to a crew that picks its captain for bearing rather than for knowledge of navigation. Experimental science confirmed it twenty-four centuries later: perceived competence drawn from an appearance or a tone of voice carries more weight in an election than actual competence does.
As the next election cycle approaches, this finding deserves to be taken seriously, whichever side of the aisle one favors. A candidate who masters the twelve tactics Antonakis catalogued, the personal anecdote, the rhetorical question, the confident look into the camera, gains a real edge in perceived competence, regardless of how solid their platform is. This holds for every party and every ticket: the effect has shown up in every democracy where researchers have gone looking for it.
The discipline I ask of my clients, measure before, during, and after any decision, applies just as well inside a voting booth. Asking what a candidate has actually done, actually costed, actually delivered, before letting the polish of a performance do the deciding, is a small habit of democratic hygiene with an outsized payoff.
The essentials
Remember three things:
- The Dr Fox effect remains an active bias today, in your leadership meetings, your hiring decisions, and your choice of vendors, wherever the polish of a pitch gets mistaken for proof that it is sound.
- The halo of presence, height, delivery, confidence, and the authority bias, the highest paid person’s opinion, reinforce each other and deserve to be named out loud before any important decision.
- The same mechanism operates at the ballot box: work from Todorov, Ballew, Antonakis, and Dalgas shows that competence perceived in a fraction of a second, from a face or a tone, carries more weight on an election outcome than most people assume.
A decision that cannot be checked against experimentation stays a hypothesis, no matter how well it was told.
References
- (My book) https://philippeboulanger.com/book/
- (Wikipedia) https://fr.wikipedia.org/wiki/Effet_Dr._Fox
- (Harvard Business Review) https://hbr.org/2012/06/learning-charisma-2
- (Science) https://www.science.org/doi/10.1126/science.1110589
- (Science) https://www.science.org/doi/10.1126/science.1167748
- (PNAS) https://www.pnas.org/doi/10.1073/pnas.0705435104
