Does willpower actually run out?

What is the willpower depletion idea?

Ego depletion is the theory that self-control draws on a limited shared resource. Resist a biscuit now and you have less capacity to hold your temper later, because the same tank supplies both. It produced the familiar advice to spend your willpower early and expect to fail in the evening.

It was one of the most influential ideas in psychology for two decades, and it is intuitive, because everybody recognises the feeling of running out of restraint by nine at night.

What happened when it was properly tested?

It disappeared. A preregistered multisite study across 36 laboratories with 3,531 participants found an effect of d = 0.06, with a confidence interval crossing zero. The Bayesian analysis concluded the data were four times more likely under the null hypothesis than under the effect.

Vohs, Schmeichel and more than 150 collaborators published this in Psychological Science in 2021. The design matters as much as the result: the protocol was registered before any data was collected, which removes the freedom to analyse until something appears.

The original meta-analysis of published studies reported d = 0.62, a moderate to large effect. Later analyses correcting for publication bias and questionable research practices brought that down toward zero, which is where the preregistered test landed.

The authors state their own caveats. Nearly a third of participants were excluded under preregistered criteria, more than expected. Only four task versions were tested out of many in the literature. Exploratory analysis of the full sample still showed only d = 0.08.

Why did it look so convincing?

Because studies finding an effect were published and studies finding nothing largely were not. Run enough small experiments on a real-sounding hypothesis and some will produce a significant result by chance, and those are the ones that appear.

This is not a story about one bad idea. It is the mechanism behind a large share of psychology findings that shrank when tested properly, and it is why the strongest sources on this site are preregistered replications and analyses that correct for publication bias.

So why does self-control feel harder later?

Something real is happening, and depletion of a resource is a poor account of it. Better-supported explanations include shifting motivation, where effort feels less worth it as the day goes on, and shifting attention, where you become more alert to rewards and less to goals.

Ordinary tiredness matters too. Self-reported fatigue appeared as a moderator in the exploratory analyses, which needs replicating, and it points at something simpler than a depleting tank.

Belief may play a part as well. If you expect willpower to run out, you are more likely to stop when it feels hard, which is a self-fulfilling account rather than a physiological one.

What should you do differently?

Stop budgeting willpower and start reducing what needs it. The reliable finding across behaviour research is that changing your circumstances beats managing your resolve, which is why removing the app works better than resisting the app.

Decide in advance instead of in the moment. If-then plans have strong evidence behind them precisely because they remove the decision at the point where you are tired and least equipped to make it.

And treat an evening lapse as information rather than proof of a depleted self. It usually points at a badly designed environment or a plan you never actually made.