In 1998, Roy Baumeister ran an experiment that would become one of the most cited studies in all of psychology. He brought participants into a room filled with the aroma of freshly baked chocolate chip cookies. Some were told to eat the cookies. Others were told to resist the cookies and eat radishes instead.
Afterward, both groups attempted an unsolvable puzzle. The radish group gave up 60% faster.
Baumeister’s conclusion: willpower operates like a muscle. Use it on one task — resisting cookies — and you have less of it for the next task. He called the phenomenon “ego depletion,” and it launched a research empire.
A Theory That Shaped a Decade
Within five years of that cookie study, ego depletion became one of the most influential frameworks in behavioral science. An initial meta-analysis by Martin Hagger and colleagues, published in 2010, surveyed 198 tests of the ego depletion effect and found a medium effect size of d = 0.62 — a robust, meaningful number.
Baumeister, a professor at Florida State University at the time, co-authored a bestselling book with journalist John Tierney called Willpower: Rediscovering the Greatest Human Strength. The book sold hundreds of thousands of copies and became a staple on corporate reading lists.
The practical implications seemed enormous. If willpower was depletable, then organizations needed to structure work environments to conserve it. Decision fatigue became a mainstream concept. Companies redesigned workflows to reduce the number of choices employees faced. Wellness programs focused on “willpower conservation.” Leaders were told to make their most important decisions in the morning, before depletion set in.
“Baumeister’s work gave us a mechanistic explanation for something everyone intuitively felt,” said Michael Inzlicht, a psychology professor at the University of Toronto who initially built his own research on the ego depletion model. “You feel drained after a hard day. Baumeister told us why.”
The Replication That Changed Everything
In 2016, Martin Hagger — the same researcher who had authored that supportive meta-analysis — coordinated a massive multi-lab replication effort. Twenty-three laboratories across the world, working from a single pre-registered protocol, tested the ego depletion effect with 2,141 participants.
The result: an effect size of d = 0.04, with a 95% confidence interval that spanned from -0.07 to 0.15. In plain language, the study found no evidence that ego depletion existed at all.
The finding landed like a bomb in the field. A second, even larger replication led by Kathleen Vohs — one of Baumeister’s longtime collaborators — tested 3,531 participants across 36 labs. It also failed to find a meaningful ego depletion effect.
“I was genuinely shaken,” Inzlicht wrote on his blog. “I had spent years building on this foundation. Now I wasn’t sure the foundation was real.”
The Debate That Followed
Baumeister did not concede. He argued that the replication studies used tasks that differed from his original 1998 protocol, and that the coordinators had rejected his preferred methodology. In his view, the replications tested something adjacent to ego depletion but not the phenomenon itself.
“Failures to replicate ego depletion are mostly operational failures,” Baumeister and Vohs wrote in a 2016 response. “They never tested the hypothesis.”
Critics found this unconvincing. If a theory cannot survive reasonable variation in how it is tested, they argued, the theory is too fragile to be useful. A robust psychological effect should show up across different tasks, populations, and contexts — not only when the original author controls the protocol.
The controversy highlighted a broader problem in social psychology known as the replication crisis. Ego depletion was not the only prominent theory to stumble. Power posing, stereotype threat, and priming effects all faced similar challenges during the same period. But ego depletion’s fall was particularly visible because the theory had been so widely adopted outside academia.
What the New Science Suggests
The collapse of the simple depletion model did not mean that self-control was suddenly easy or unlimited. It meant the mechanism was wrong.
Newer research, led by scholars like Inzlicht and Veronika Job at the University of Zurich, suggests that what looks like depletion may actually be a shift in motivation. When people perform a difficult task, they do not run out of willpower. They become less willing to allocate effort to a second task because their motivation shifts.
Job’s work showed that people who believed willpower was limited experienced more depletion. Those who believed willpower was abundant did not. This “mindset” theory of self-control reframed the entire question. Depletion was not a biological inevitability. It was, at least in part, a self-fulfilling prophecy.
“The resource model told us willpower is a gas tank that empties,” Inzlicht said in a 2019 lecture. “The new model says it’s more like an emotion. It fluctuates based on what you care about, what you believe, and what you expect.”
This shift mirrors what other researchers have found about how motivation works — and how easily it can be disrupted by the wrong kind of incentive.
What Managers Built on the Old Science
The organizational structures inspired by ego depletion were not all wrong. Reducing unnecessary decision-making is still good practice. Simplifying workflows still helps. The morning meeting still probably is better than the 4 p.m. meeting, if only because people are more alert.
But the underlying logic needs updating. If willpower is not a finite tank, then the goal is not conservation. It is alignment.
The new research suggests that people perform sustained self-control best when the task connects to something they genuinely value. An employee who finds meaning in their work does not run out of willpower at 2 p.m. An employee doing busywork for a goal they do not believe in will “deplete” by 10 a.m. — not because their tank is empty, but because their motivation has shifted elsewhere.
This reframes the manager’s job. Instead of designing environments that minimize willpower expenditure, leaders should focus on ensuring that the work itself is connected to purpose. That means making progress visible, removing meaningless friction, and making sure people understand why their tasks matter — precisely what the best coaching conversations are designed to do.
The Broader Lesson
Ego depletion’s trajectory is a case study in how science is supposed to work — and why relying on a single theory to build organizational policy is always risky.
Baumeister’s original insight — that self-control has limits — remains directionally useful. People do experience fatigue. Cognitive load is real. Decision quality does decline under sustained pressure.
But the specific model that explained these observations, the one that generated the corporate programs and the bestselling books and the morning-routine advice, did not survive scrutiny. The effect size went from d = 0.62 to d = 0.04. That is not a minor correction. It is a collapse.
For leaders, the takeaway is twofold. First, be cautious about building policy on any single study, no matter how famous. The half-life of a psychology finding is shorter than the half-life of the organizational structure built on top of it.
Second, update your model. The new science of self-control is more nuanced, more empowering, and more actionable than the old one. Willpower is not a gas tank. It is a signal — a signal about whether the work in front of you feels worth the effort.
If your team is running out of willpower by midafternoon, the problem may not be their biology. It may be that the work you are asking them to do has too much friction and too little meaning.
That is a problem no amount of willpower conservation can solve — and one that only better leadership can fix.