Does Sugar Really Fuel Willpower?

By Neuroskeptic | July 14, 2016 10:15 am

Another prominent psychology theory has come under scrutiny by researchers who say the published results look unrealistic.

In a new paper, Miguel A. Vadillo et al. take aim at the idea that the body’s reserves of willpower rely on glucose.

The background here is the ‘ego depletion‘ model, a psychological theory which holds that self-control is effortful and draws on a limited resource, which can eventually be depleted if it’s overused. Many researchers have proposed that glucose is this resource. Numerous studies have shown that a sugary drink restores the depleted ego, boosting performance on self-control measures.

But Vadillo et al. say that publication bias and/or p-hacking is prevalent in this literature, calling the results into question. The authors present a p-curve analysis of the published glucose-boosts-willpower literature (including 19 papers in all) showing that it is highly unlikely to be free of bias:

vadillo_pcurveThe p-curve is a technique which plots the density of p-values below the significance threshold of 0.05. If there is a real effect (i.e. if glucose does boost the ego), the p-values should be clustered around 0, i.e. there should be a peak on the left side of the p-curve. However, in the case of the glucose literature, the peak is at 0.04, just barely significant, which is unlikely to occur when there is a real effect, even a small one. The observed p-curve is consistent with the theory that the positive results are just chance findings which were selectively reported in the published literature.

Such selective reporting is not unique to this literature, mind you, but is rather a near universal problem in all science which uses p-values, as I’ve decried many times.

Overall, this doesn’t look good for the glucose-ego theory. That said, the p-curve is a relatively new technique, and has not yet gained universal acceptance as a way of detecting bias. I wonder what the creators of the p-curve, Simonsohn, Simmons and Nielson, think of this paper?

ResearchBlogging.orgVadillo, M., Gold, N., & Osman, M. (2016). The Bitter Truth About Sugar and Willpower: The Limited Evidential Value of the Glucose Model of Ego Depletion Psychological Science DOI: 10.1177/0956797616654911

CATEGORIZED UNDER: FixingScience, papers, select, Top Posts
  • Priit Kallas

    “just barely significant”? You don’t really understand p-value do you? It’s not black and white blow and above 0.05 it’s a gradient.

    • Neuroskeptic

      Within the conventional framework that p 0.05 is not significant, it is fair to say that p = 0.04 is barely significant because while it is significant, it’s close to being nonsignificant.

      I agree that significance is not black and white “really” however the convention is to impose a dichotomy drawing the line at p=0.05.

      • Priit Kallas

        p=0.04 and p=0.06 mean that in 100 experiments 4 or 6 might come up with different result, this is not “just barely”.

        • RogerSweeny

          True. But if the convention is that anything more than p=0.05 is not significant and anything less than p=0.05 is significant, then *under that convention*, p=0.04 is “just barely” significant.

    • Anonymouse

      Your comment is inconsistent. Even ignoring that this phrasing is standard in many fields, Neuroskeptic’s very use of “(just) barely” suggests a non-dichotomous, gradual understanding of p-values, without ignoring the popular use of the threshold of .05.

      Without understanding the nature of p-values, reporting on a p-curve wouldn’t make much sense either.

  • Uncle Al

    I provided several examples to sugar being a straw man. Man-up and post CENSORED where my post resided.

  • Anonymouse

    Willpower Challenge: Try cutting glucose from your diet.

  • Pingback: What I learned from 100 habit change listicles – The Habit Digest()

  • Arayma

    I don’t remember, did you address the paper “Could a Neuroscientist Understand a Microprocessor?” I’m afraid that sugar and willpower don’t matter at all if the state of the field is at such a primitive level.

  • Pingback: The End of Ego-Depletion Theory? - Neuroskeptic()

  • lizarde

    I wish I’d known this before I started grad school in social psych. This was a pretty good rationalization for grazing and sweet drinks when writing papers…

  • Pingback: Signal Conduction -19 August 2016 | Cognitales()

  • Christine Sarah Pike

    As every failed dieter will attest, sugar does not enhance willpower! Also, i am bothered that an ostensibly serious scientific study assumes unquestioningly the reality of such a dubious and unproven notion as ‘ego’.

  • Pingback: Homepage()

  • Pingback: Why A Morning Workout Sets You Up For Daily Success()



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar