It's 11pm and you're looking at the app you swore you'd open every night. The streak died on Tuesday. And the sentence that surfaces is the same one everybody reaches for: I just don't have the willpower.
That sentence feels like an observation. It's actually a theory. A very specific, very famous scientific theory about how self-control works: willpower is a limited resource, a kind of fuel tank or battery, and every act of restraint drains it a little. Skip the donut at breakfast, bite your tongue in the meeting, and by evening there's nothing left in the tank for journaling or the gym or being patient with your roommate.
Here's the thing. That theory came from one 1998 experiment. And when scientists tried, at massive scale and on purpose, to reproduce its effect, they found approximately nothing.
where the battery idea came from
The myth has a proper name: ego depletion. In 1998, Roy Baumeister and colleagues published a paper.pdf) in the Journal of Personality and Social Psychology with a setup that's since become legend. Hungry participants sat in a room that smelled of fresh-baked chocolate chip cookies. Some got to eat the cookies. Others were told to resist them and eat radishes instead. Then everyone worked on puzzles that were secretly unsolvable. The radish-eaters gave up much sooner, and the researchers concluded that resisting the cookies had drained some inner resource, leaving less self-control for the puzzles. Willpower, they argued, works like a muscle that tires.
It's a great story. It's intuitive, it's flattering (your failures aren't your fault, your tank was empty), and it spread everywhere: bestselling books, productivity advice, the tip that a sugary snack can refill your willpower. The psychologist Michael Inzlicht, who spent years working within this framework himself, notes that something like 600 studies appeared to support ego depletion, and the field handed out a major career award for the line of work.
Then people started checking.
twenty-three labs, one very small number
In 2016, Martin Hagger led a preregistered replication across 23 labs, published in Perspectives on Psychological Science. Everyone agreed on the protocol in advance. Nobody could quietly shelve a disappointing result. The finding: an effect so close to zero it was statistically indistinguishable from no effect at all. Not "smaller than we thought." Not "depends on the person." Basically nothing.
This wasn't a lone shot, either. A meta-analysis by Evan Carter and colleagues, discussed in the same paper, had already corrected the existing literature for publication bias and concluded that the true effect was probably around zero, and that the evidence simply doesn't support the popular idea of self-control running on a limited fuel supply.
How do 600 studies point one way and the truth point another? Inzlicht's own reckoning is worth reading in full, because it's rare and a little moving: a researcher who built part of his career on an effect writing, publicly, that it wasn't real. The short version is that small studies, flexible methods, and journals that only publish positive results can manufacture an entire literature around a phantom.
Even the candy-bar version of the myth fell apart under scrutiny. Veronika Job, Gregory Walton, and colleagues published a study in PNAS showing that sugar only appeared to "restore" self-control in people who already believed willpower was a limited resource. People who didn't hold that belief performed fine without the glucose. Which is a strange result for a battery. Batteries don't care what you believe about them.
the part where it gets genuinely complicated
There's a tempting replacement story: the depletion effect was really a belief effect all along. In a 2010 series of studies, Job, Carol Dweck, and Walton found that people who thought of willpower as limited showed the classic depletion pattern, while people who saw it as non-limited (even self-renewing) kept performing across demanding tasks. A follow-up in 2015 linked the non-limited belief to better everyday self-regulation and better grades, especially when demands piled up.
I'd love to hand you that as the tidy new truth. I can't, quite. A 2023 preregistered replication in PLOS ONE tried to reproduce the original mindset finding and failed. The willpower-mindset story might be real in some conditions and not others. It might be smaller than it first looked. It might be another phantom. The honest position is that we don't know yet.
What we do know is what's not true, and that's still worth a lot. Self-control is not a battery that drains with each use. You don't have a fixed daily ration of discipline that your commute and your inbox spend before you get home. Writers at Psyche have argued for a better model: what shifts over a hard day isn't your capacity but your motivation — attention drifting from "have to" tasks toward "want to" ones. You're not empty at 11pm. You just want different things than you wanted at 9am.
That reframe changes what a lapsed habit means.
Under the battery model, your dead journaling streak is a resource problem, and the fix is to somehow acquire more willpower. Nobody can tell you how to do that, so you just feel defective. Under the motivation model, the lapse is information. Maybe the habit costs too much at the hour you scheduled it. Maybe it feels like a "have to" when it started as a "want to." Those are solvable. You can move the habit to a moment when you actually have appetite for it. You can shrink it. You can change its format entirely — this is roughly why we built Joice, because for a lot of people the blank page is the "have to" and talking out loud is the "want to," and swapping one for the other costs nothing.
The cruelest part of the willpower myth was never the science. It was the self-talk it licensed. I don't have the willpower sounds like humility, but it's a verdict: a claim that you're missing a substance other people have. The research says there's no such substance to be missing.
The streak will break again. When it does, skip the verdict and ask the boring question instead: what did this habit cost tonight, and who set the price? You did. Which means you can change it.