The Largest Test Yet of Whether Willpower Runs Out

This was, at the time, the most ambitious attempt yet to settle the ego-depletion question: 36 independent laboratories, over 3,500 participants, and two different standardised procedures for inducing and measuring depletion, all pre-registered before data collection began. The primary, pre-registered analysis found no significant effect, a vanishingly small d of 0.06, and a Bayesian analysis, which directly compares how likely the data are under competing hypotheses, found the results were about four times more likely if there genuinely was no depletion effect than if there was a real, moderate one. One interesting exploratory finding: the effect was somewhat larger in participants who reported feeling more fatigued generally, hinting that ordinary tiredness, rather than a specific 'self-control resource', might explain some of what earlier, smaller studies had picked up.

As with the 2016 replication, this scale is the entire point, a result this consistent across 36 independently run labs is difficult to attribute to any single lab's quirks or errors. Combined with the accumulated replication evidence before it, this study represents as close to a settled scientific verdict as psychology tends to produce: the specific, popular idea that everyday self-control draws down a single depletable reservoir, tested via this paradigm, has not held up.