Willpower ran out in the lab and then the effect ran out too
The most influential self-control study of the 1990s inspired a decade of advice — and then a 23-lab test couldn't find the effect at all.
In 1998, Roy Baumeister and colleagues put hungry volunteers in a room with a plate of chocolates and a bowl of radishes, told half of them to eat only the radishes, and then gave everyone a frustrating, unsolvable puzzle. The radish-eaters gave up on the puzzle much faster than the people who'd been allowed the chocolate. Baumeister's explanation became one of psychology's most quoted ideas: self-control works like a muscle. Use it resisting temptation, and you have less left over for the next demanding task.
The "ego depletion" theory spread fast, in part because it matched a feeling most people already had — that saying no all day leaves you frayed by evening. A 2010 meta-analysis pooling over 80 studies seemed to confirm it was real and reasonably large.
Then came the correction. In 2016, Martin Hagger — one of the authors of that supportive 2010 meta-analysis — ran a pre-registered replication across 23 independent labs, with the methodology locked in before any data came in. Across all 23, the ego depletion effect essentially disappeared.
The theory isn't dead; researchers are still arguing about which version of the experiment, if any, holds up. But the clean, muscle-like story that launched a thousand productivity articles no longer has the evidence it once had. The uncomfortable finding here isn't about willpower. It's about how confidently a field can be wrong for eighteen years running.