Dianne Tice

Who they are

Tice appears on this site because of two papers, separated by nearly a quarter of a century, that bracket the whole history of a research programme.

The first is Baumeister, Bratslavsky, Muraven and Tice (1998), published in the Journal of Personality and Social Psychology. Sixty-seven students were seated at a table holding freshly baked chocolate chip cookies and a bowl of radishes, and a third of them were told they could eat only the radishes. Everyone was then handed geometry problems that had no solutions. The radish group gave up after 8.35 minutes; the group allowed the cookies lasted 18.90. What the radish group had spent, the authors argued, was a general-purpose resource, and the strength model of self-control was built on that inference. It became one of the most influential studies in modern social psychology, and the source of a great deal of practical advice: make hard decisions early, budget your willpower across the day, eat before a difficult conversation.

The second is Baumeister and Tice (2022), "Ego Depletion is the Best Replicated Finding in All of Social Psychology", in the Scholarly Journal of Psychology and Behavioral Sciences. By then the model had been through a decade of testing. Evan Carter and Michael McCullough's 2014 bias-correction reanalysis found a corrected estimate that could be zero. Martin Hagger's 23-laboratory registered replication returned d = 0.04 in 2016. Kathleen Vohs led a 36-laboratory preregistered test that returned d = 0.06 in 2021, with a confirmatory Bayesian analysis finding the data four times more likely under the null.

The defence sets out four criteria for calling an effect well replicated: many supportive findings across laboratories and methods with few contrary results, at least one preregistered success, a significant multi-site replication, and evidence from outside the laboratory. On the third criterion it points to Dang and colleagues' 12-laboratory replication, also 2021, which did find a small but statistically significant effect at d = 0.10, rising to 0.16 once apparently random responders were excluded. It raises methodological objections to both of the famous null replications, concerning the fatigue check in the Hagger protocol and the sample used in the Vohs analysis; the replication teams' published reports set out their methods in full, and this site records the two positions without adjudicating.

That objection is on the record and belongs on this page rather than being waved through. The piece is worth reading properly rather than dismissed on the strength of the journal it appeared in, and Dang et al. (2021) is a real result that the flat debunking version of this story tends to skip. Where the argument runs out is at the size of what it rescues. Grant all four criteria and what remains is an effect of roughly a tenth of a standard deviation under some conditions. That is not nothing. It is also not what the advice was sold on, and none of the practical guidance built on the model survives at that size. The glucose claim, that a sugary drink restores the resource, has fared worse still: Vadillo, Gold and Osman found the literature's evidential value weak by p-curve in 2016.

This site's view is that the finding did not survive at anything like its original size, and says so. That is a verdict on the finding rather than on the people who produced it. Tice was doing 1990s social psychology under 1990s norms, in a field that had not yet built preregistration, multi-site designs or routine bias correction, and the reason the effect could be measured down so precisely is that the original paper reported its method clearly enough to be replicated at all. Sitting on both the founding paper and the defence is not stubbornness. It is what a consistent position looks like when the evidence around it moves.

Books

The vault records no books by Tice. She is a working scientist whose output here is journal articles, listed below. Nothing further is claimed about the rest of her bibliography, because nothing further has been verified.

Key papers

Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. (1998), "Ego depletion: Is the active self a limited resource?," Journal of Personality and Social Psychology, 74(5), 1252-1265. The radish study, N = 67. The paper that introduced ego depletion and the strength model of self-control.

Baumeister, R. F., & Tice, D. M. (2022), "Ego depletion is the best replicated finding in all of social psychology," Scholarly Journal of Psychology and Behavioral Sciences, 6(2), 686-688. Four criteria for a well replicated effect, the case built on Dang et al. (2021), and methodological objections to the Hagger and Vohs replications, recorded here as one side of an open question.

Key ideas

The strength model of self-control: acts of restraint draw on a limited common resource, so exerting control on one task leaves less available for the next.

Ego depletion as the observable consequence, operationalised in 1998 as persistence on an unsolvable task.

Four criteria for calling an effect well replicated, proposed in the 2022 defence: convergent findings across laboratories and methods, a preregistered success, a significant multi-site replication, and evidence from the field.

Two methodological objections to the replications, published in 2022, concerning the fatigue check in the Hagger protocol and the sample used in the Vohs analysis; both positions are on the record.

The unresolved remainder. Granting the entire defence still leaves an effect near d = 0.10, which cannot carry the advice that was built on it.

Concepts on this site

Ego Depletion

The Replication Crisis

---