Martin Hagger

Who they are

The 2010 meta-analysis, with Wood, Stiff and Chatzisarantis, is the document that consolidated ego depletion. It pooled close to 200 experiments and reported an effect of d = 0.62, medium to large, large enough to plan a working day around. That estimate is what licensed the willpower advice of the following decade: the sugary drink, the willpower budget, the instruction to make hard decisions before the tank runs down. Almost every trade book treatment of self-control published after 2010 is standing on it, whether or not it names the paper.

Six years later Hagger organised the test. The 2016 study in Perspectives on Psychological Science was a preregistered registered replication across 23 laboratories with a single standardised protocol and 2,141 participants. It returned d = 0.04, 95% CI [-0.07, 0.15]. A confidence interval that wide, sitting on zero, is not a small effect confirmed. It is an effect that cannot be distinguished from no effect at all.

The gap between the two numbers has a diagnosis, and Evan Carter had supplied it two years earlier by reanalysing Hagger's own 2010 dataset for publication bias. Small studies with large reported effects were far more numerous than a real effect of that size would produce. The 2010 estimate was not wrong arithmetic. It was an accurate summary of a literature that had been filtered before it reached him.

Roy Baumeister's published reply to the 2016 result is specific enough to state rather than skip. In their 2022 defence, Baumeister and Dianne Tice argue that Hagger's e-crossing protocol was new and untested, and that it failed its own manipulation check on fatigue, which would mean the depletion manipulation never took hold and the study therefore tested nothing. That objection is on the record and has not been resolved here. What is also on the record is that two further multi-laboratory tests followed, one led by Kathleen Vohs returning d = 0.06 across 36 laboratories, and one led by Dang and colleagues returning a small but significant d = 0.10 across 12. The range those three produce, zero to 0.16 across more than 7,000 participants, is a long way below 0.62 whichever protocol you prefer.

There is a second reason his name matters on this site. Nir Eyal's Indistractable credits the 2010 meta-analysis of nearly 200 experiments to Evan Carter rather than to Hagger and colleagues. The attribution is wrong, and it is worth correcting, because the two papers make opposite claims about the same body of evidence.

Key papers

Hagger, M. S., Wood, C., Stiff, C., & Chatzisarantis, N. L. D. (2010), "Ego depletion and the strength model of self-control: A meta-analysis," Psychological Bulletin, 136(4), 495-525. The meta-analysis of close to 200 experiments that put the depletion effect at d = 0.62 and made the strength model the default account of self-control.

Hagger, M. S., Chatzisarantis, N. L. D., Alberts, H., et al. (2016), "A multilab preregistered replication of the ego-depletion effect," Perspectives on Psychological Science, 11(4), 546-573. Twenty-three laboratories, a single preregistered protocol, N = 2,141, d = 0.04, 95% CI [-0.07, 0.15]. The first large preregistered test of the effect, and the point at which the field's central estimate collapsed.

Key ideas

Meta-analysis inherits the biases of the literature it summarises. The 2010 estimate was a faithful summary of a filtered evidence base.

A preregistered, standardised, multi-laboratory protocol is a different instrument from a pool of independently published experiments, and the two can disagree by an order of magnitude on the same question.

A confidence interval spanning zero, as in the 2016 result, is not a small effect. It is an effect that cannot be separated from none.

Running the decisive test on your own headline finding, and publishing what it returns, is the mechanism by which self-correction actually happens.

The 2016 protocol is contested. Baumeister and Tice published a methodological objection in 2022 concerning the protocol's fatigue check; the replication report sets out the method in full, and both positions are on the record.

Concepts on this site

Ego Depletion

The Replication Crisis

Find them elsewhere

The Bluesky account lists the same three affiliations, Merced, Jyvaskyla and Griffith, as his Scholar profile, which is verified at ucmerced.edu, and the X account states the same Merced professorship

Checked 2026-09-10. Handles and affiliations go out of date, so treat this block as a snapshot rather than a standing fact.

---