How strong is the evidence for the argument made in Willpower by Baumeister and Tierney?

Basically willpower (or better: "self-control") is supposed to be proportional to the amount of glucose in your blood and if you use up the willpower, it takes time to replenish the resource.

Robert Kurzban argues in Psychology Today that it's false.

Is there an academic consensus about the issue? If so what's the current consensus?

  • Good question! Thanks for the link to the Kurzban critique of the idea, it's all too rare to see critique get voiced clearly in psychology. There is a follow-up too. His peeve seems to be mostly the resource model, not the idea that glucose has some causative influence (which is already pretty big...) – Ruben Jun 19 '12 at 17:29
  • Oh wow, he basically calls Gaillot & co. conmen, because they don't share data. I think you won't necessarily find a consensus as I'd subjectively say there's this field of "self" psychology, which this idea belongs too, where most probably believe this (or don't dare contradict Baumeister), and then there's psychologists who have a sensible metatheory, they likely don't read much of this ;-). I'm not an expert, but that may explain it. Oh that and the probably pervasive publication bias in psychology. – Ruben Jun 19 '12 at 17:47
  • Willpower was the topic of NPR's Science Friday segment two weeks ago. Listen to it on Science Friday's site or on NPR's site. – Sam I Am Jul 2 '12 at 2:41
  • You need look no further than the Running community, to find that there is a huge variety of opinions on basically carbo loading. Some say that eating carbs (in short) "gives you energy". But others say that carb-based diets simply prevent your body from working as energetically as it normally does and should. (So you had the era where the Kenyans would crush everyone at the marathon, and laugh at the carb-loading idea.) Example article – Fattie Jul 26 '18 at 15:41
  • Yes; as it's the fuel for the brain, a lack of it will result in death. Dead people have no willpower (unless they wrote a will while alive on glucose). – Cees Timmerman Jan 17 '19 at 5:42

You're lucky, the intense debate about psychology's (and other sciences') false-positive problem has spawned an article (by Schimmack, hailing from your FU Berlin by the way) which uses Gaillot's initial study as a case study of incredible results. The author shows how dodgy methodology, analysis and reporting choices made the initial study incredible (actually some of Gaillot et al.'s reported studies didn't replicate the effect, but those tests weren't reported).

Cherries (picking and emphasis mine):

He [Kurzban] was informed that the data from Study 1 were corrupted and not available, but he obtained the data from studies 3 to 6, which also provided information about the influence of self-regulation on glucose consumption although these results were not reported in the original article. The main finding was that across the four studies, the results showed no significant decrease in glucose levels.
sample sizes were modest, ranging from N = 12 to 102. Four studies had sample sizes of N < 20, which Simmons et al. (2011) considered to require special justification.
Effect sizes of the various paradigms were converted into d-values. Effect sizes varied considerably from d = .23 to d = 1.53. This variation was strongly negatively related to sample sizes, r = -.79, without any explicit explanations for this correlation.
This indicates that from a statistical point of view, Bem’s (2011) evidence for ESP is more credible than Gailliot et al.’s (2007) evidence for a role of blood-glucose in self-regulation.
In conclusion, Gailliot et al. (2007) had limited resources to examine the role of blood glucose in self-regulation. By attempting replications in nine studies they did not provide strong evidence for their theory. Rather, the results are incredible and difficult to replicate, presumably because the original studies yielded inflated effect sizes. A better solution would have been to test the three hypotheses in a single study with a large sample. This approach also makes it possible to test additional hypotheses, such as mediation (Dvorak & Simons, 2009). Thus, Example 2 illustrates that a single powerful study is more informative than several small studies

I'm glad you asked this question and led me to Kurzban's blog. My first problem with Gaillot et al. (read it years ago) was that with effects this big, it seemed implausible that so many people eat more sweets than they "want".

I realize that this isn't a meta-analysis of the findings, but I think it helps a lot to see what is dodgy about it spelled out clearly by more skeptical scientists. I'd reckon (from experience, without having read them all) that some of the articles collected by Sklivvz suffer from similar ailments (clearly Gaillot et al. is not a solid, confirmatory study) and that those by Kurzban do not. There is a grand unified idea behind it, i.e. "the self" but I'm unsure whether this is a good metatheory that can generate claims to be disproved as well.

Also: Why was this moved from cogsci? Seems like a perfect fit, maybe we should move it back. Skepticism of cognitive science should be part of cogsci if you ask me.

  • 1
    Downvoter, what can I improve? – Ruben Jul 2 '12 at 13:48
  • Note that your comment that this should not have been migrated probably isn't doing much help here. It belongs on meta.cogsci. – Oddthinking Aug 28 '14 at 1:01
  • Did you mean "non-credible" results instead of "incredible" ("unglaubwürdig" statt "unglaublich") ? – Eike Pierstorff Aug 29 '14 at 7:05
  • @EikePierstorff I of course meant incredible as in not credible, as is clear from the context. I used that word because Schimmack called his index the Incredibility Index. However, he, you and me are all Germans. I know that the colloquial use of incredible is more common, but I think the literal use is still understood. You could ask on english.SE if you care. – Ruben Sep 15 '14 at 17:25

Yes, there's a correlation and a causation. Here are 9 studies that prove it:

The present work suggests that self-control relies on glucose as a limited energy source. Laboratory tests of self-control (i.e., the Stroop task, thought suppression, emotion regulation, attention control) and of social behaviors (i.e., helping behavior, coping with thoughts of death, stifling prejudice during an interracial interaction) showed that (a) acts of self-control reduced blood glucose levels, (b) low levels of blood glucose after an initial self-control task predicted poor performance on a subsequent self-control task, and (c) initial acts of self-control impaired performance on subsequent self-control tasks, but consuming a glucose drink eliminated these impairments. Self-control requires a certain amount of glucose to operate unimpaired. A single act of self-control causes glucose to drop below optimal levels, thereby impairing subsequent attempts at self-control.

Self-Control Relies on Glucose as a Limited Energy Source: Willpower Is More Than a Metaphor by Gailliot, et al., Journal of Pers and Soc Psycholy (2007), Vol. 92, No. 2, 325–336, DOI: 10.1037/0022-3514.92.2.325

Past research indicates that self-control relies on some sort of limited energy source. This review suggests that blood glucose is one important part of the energy source of self-control.

Unlocking the Energy Dynamics of Executive Functioning: Linking Executive Functioning to Brain Glycogen doi: 10.1111/j.1745-6924.2008.00077.x - Persp on Psychol Sc, July 2008 vol. 3 no. 4 245-263

This provides a different spin on it:

The strength model suggests that self-control relies on a limited resource. One candidate for this resource is glucose. Counter to the proposals of the glucose hypothesis, this study argues that the resource issue is one of allocation, not of limited supply.

The Role of Glucose in Self-Control: Another Look at the Evidence and an Alternative Conceptualization - Pers Soc Psychol Rev May 1, 2012 16: 143-153

Interpersonal provocation is a common and robust antecedent to aggression. Four studies identified angry rumination and reduced self-control as mechanisms underlying the provocation—aggression relationship. Following provocation, participants demonstrated decreased self-control on an unpleasant task relative to a control condition (Study 1). When provoked, rumination reduced self-control and increased aggression. This effect was mediated by reduced self-control capacity (Study 2). State rumination following provocation, but not anger per se, mediated the effect of trait rumination on aggression (Study 3). Bolstering self-regulatory resources by consuming a glucose beverage improved performance on a measure of inhibitory control following rumination (Study 4). These findings suggest that rumination following an anger-inducing provocation reduces self-control and increases aggression. Bolstering self-regulatory resources may reduce this adverse effect.

Understanding Impulsive Aggression: Angry Rumination and Reduced Self-Control Capacity Are Mechanisms Underlying the Provocation-Aggression Relationship - Pers Soc Psychol Bull June 1, 2011 37: 850-862

Worthy of the Annals of Improbable Research:

We test the common caricature of realism that justice is “what the judge ate for breakfast” in sequential parole decisions made by experienced judges. We record the judges’two daily food breaks, which result in segmenting the deliberations of the day into three distinct “decision sessions.” We find that the percentage of favorable rulings drops gradually from ≈65% to nearly zero within each decision session and returns abruptly to ≈65% after a break. Our findings suggest that judicial rulings can be swayed by extraneous variables that should have no bearing on legal decisions.

Extraneous factors in judicial decisions - Proc. Natl. Acad. Sci. USA April 26, 2011 108: 6889-6892

These researchers prove reality is not so simple, after all:

Much recent research suggests that willpower—the capacity to exert self-control—is a limited resource that is depleted after exertion. We propose that whether depletion takes place or not depends on a person’s belief about whether willpower is a limited resource. Study 1 found that individual differences in lay theories about willpower moderate ego-depletion effects.

Ego Depletion--Is It All in Your Head?: Implicit Theories About Willpower Affect Self-Regulation - Psychological Science November 1, 2010 21: 1686-1693

A group of participants who drank a soft drink that contained sugar showed a reduced rate of future discounting afterward, when we controlled for sex, age, body mass index, and the taste of the drink. In contrast, a group of participants who drank a soft drink that contained artificial sweetener showed an increased rate of future discounting. Blood glucose levels not only varied as a result of caloric intake but also regulated the rate of future discounting, according to participants’ dynamic body-energy budget.

Sweet Future: Fluctuating Blood Glucose Levels Affect Future Discounting - Psychological Science February 1, 2010 21: 183-188

Past work suggests that executive functioning relies on glucose as a depletable energy, such that executive functioning uses a relatively large amount of glucose and is impaired when glucose is low. Glucose from the bloodstream is one energy source for the brain, and glucose stored in the brain as glycogen is another. A review of the literature on glycogen suggests that executive functioning uses it in much the same way as glucose, such that executive functioning uses glycogen and is impaired when glycogen is low. Findings on stress, physical persistence, glucose tolerance, diabetes, sleep, heat, and other topics provide general support for this view.

Unlocking the Energy Dynamics of Executive Functioning: Linking Executive Functioning to Brain Glycogen - Perspectives on Psychological Science July 1, 2008 3: 245-263

This experiment used the attraction effect to test the hypothesis that ingestion of sugar can reduce reliance on intuitive, heuristic-based decision making. In the attraction effect, a difficult choice between two options is swayed by the presence of a seemingly irrelevant ‘‘decoy’’ option. We replicated this effect and the finding that the effect increases when people have depleted their mental resources performing a previous self-control task.

Toward a Physiology of Dual-Process Reasoning and Judgment: Lemonade, Willpower, and Expensive Rule-Based Analysis - Psychological Science March 1, 2008 19: 255-260

  • 1
    Those studies point to a lot of single pieces of evidence. The book itself cites 209 sources. Is there something like a general theory with academic acceptance that links all those pieces of evidence together? Initially I posted the question over at cognitive sciences to get a summary that goes beyond a single list where I don't really know about the quality of the single studies. – Christian Jun 17 '12 at 11:57
  • @Christian the top study seems to be of a good quality, obviously what the other studies show, is that reality is not as black and white as the claim suggests. On the other hand, I would say that the scientific consensus on the matter is univocal... – Sklivvz Jun 17 '12 at 19:13
  • 1
    Reality isn't black&white isn't really any improvement on the amount of knowledge that I had when I posed the question and pointed to different people holding different positions. I didn't check all your sources specifically but I think most of them are already cited by Baumeister&Tierney or Kurzban. When you discover that reality isn't black&white it helps to ask an expert about how reality actually is. I know that expert opinions aren't popular on this site and therefore I asked on cogsci. – Christian Jun 18 '12 at 10:58
  • However, correlation doesn't not imply causation per se and it still seems to be something that is up for debate. – rjzii Jun 18 '12 at 18:50
  • @RobZ this does examine causation. – Sklivvz Jun 18 '12 at 22:22

As of right now, it seems like the issue is actually still up for debate within the scientific community and there was an article published in Psychology Today in 2011 that attacks the general concept and the model of the experiments themselves. Another recent study 1 showed a correlation between the reward (i.e. the sweet drink) and the action being preformed drove performance as opposed to glucose itself.

Based upon my reading, the issue seems to be the current understanding of the brain indicates that there is a fixed consumption of energy that is independent of the tasks being performed and preventing yourself from doing something shouldn't cause more energy to be consumed.

Another recent (March 2012) article has pointed out that newer studies and models are being replaced by the concept that

Recent experiments indicate that cognition heavily influences willpower. Specifically, our beliefs about willpower – whether we conceive of it as biologically limited or not – immensely influence our self-control.

Which has been giving rise to the concept of ego depletion as a willpower model as opposed to glucose depletion so at the present time it seems like the best answer is that further research is needed and the jury is still out as to what the best model will end up being.

  1. See the articles cited in the article for more details.
  • Good job looking at the more recent articles. Do keep in mind that they are meant as incremental improvements over the previous results, though. – Sklivvz Jun 18 '12 at 22:23
  • @Sklivvz - True which is why I pointed out that the best answer is likely that more research is needed. It seems like there is some soft evidence that willpower might be a multivariate problem so there will not be just once factor that defines it. – rjzii Jun 18 '12 at 22:26

I can only quote The Willpower Instinct: How Self-Control Works, Why It Matters, and What You Can Do To Get More of It by Kelly McGonigal

Kelly McGonigal makes connection with glucose and willpower (claiming it is one of most energy demanding tasks) and also quotes the work of Baumeister from the book in question. She also notes that it's not direct relationship between glucose level in blood, but the direction of change of glucose level in blood.


Obviously we don’t have an actual self-control muscle hidden underneath our biceps, keeping our hands from reaching for dessert or our wallet. We do, however, have something like a self-control muscle in our brain. Even though the brain is an organ, not a muscle, it does get tired from repeated acts of self-control. Neuroscientists have found that with each use of willpower, the self-control system of the brain becomes less active. Just like a tired runner’s legs can give out, the brain seems to run out of the strength to keep going. Matthew Gailliot, a young psychologist working with Roy Baumeister, wondered whether a tired brain was essentially a problem of energy. Self-control is an energy-expensive task for the brain, and our internal energy supply is limited—after all, it’s not like we have an intravenous sugar drip into our prefrontal cortex. Gailliot asked himself: Could willpower exhaustion simply be the result of the brain running out of energy? To find out, he decided to test whether giving people energy—in the form of sugar—could restore exhausted willpower. He brought people into the laboratory to perform a wide range of self-control tasks, from ignoring distractions to controlling their emotions. Before and after each task, he measured their blood sugar levels. The more a person’s blood sugar dropped after a self-control task, the worse his performance on the next task. It appeared as if self-control was draining the body of energy, and this energy loss was weakening self-control. Gailliot then gave the willpower-drained participants a glass of lemonade. Half of them received sugar-sweetened lemonade to restore blood sugar; the other half received a placebo drink that was artificially sweetened and would not supply any usable energy. Amazingly, boosting blood sugar restored willpower. The participants who drank sugar-sweetened lemonade showed improved self-control, while the self-control of those who drank the placebo lemonade continued to deteriorate. Low blood sugar levels turn out to predict a wide range of willpower failures, from giving up on a difficult test to lashing out at others when you’re angry. Gailliot, now a professor at Zirve University in Turkey, has found that people with low blood sugar are also more likely to rely on stereotypes and less likely to donate money to charity or help a stranger. It is as if running low on energy biases us to be the worst versions of ourselves. In contrast, giving participants a sugar boost turns them back into the best versions of themselves: more persistent and less impulsive; more thoughtful and less selfish. Well, as you can imagine, this is just about the most best-received finding I’ve ever described in class. The implications are at once counterintuitive and delightful. Sugar is your new best friend. Eating a candy bar or drinking soda can be an act of self-control! (Or at least restoring self-control.) My students love these studies and are only too happy to test the hypothesis themselves. One student used a steady supply of Skittles to get through a difficult project. Another kept a tin of Altoids (one of the last breath mints to contain real sugar) in his pocket, popping them during long meetings to outlast his colleagues. I applaud their enthusiasm for translating science into action and empathize with their sweet tooth. And I even confess that for years, I brought candy to every Introduction to Psychology class, hoping to get the undergraduate students focused and off Facebook.7 If sugar were truly the secret to more willpower, I’m sure I’d have a runaway bestseller on my hands and a lot of eager corporate sponsors. But as my students and I were trying our own willpower-replenishing experiments, some scientists—including Gailliot—started to raise some smart questions. How much energy, exactly, was getting used up during acts of mental self-control? And did restoring that energy really require consuming a substantial amount of sugar? University of Pennsylvania psychologist Robert Kurzban has argued that the actual amount of energy your brain needs to exert self-control is less than half a Tic Tac per minute. This may be more than the brain uses for other mental tasks, but it is far less than your body uses when it exercises. So assuming you have the resources to walk around the block without collapsing, the absolute demands of self-control couldn’t possibly deplete your entire body’s store of energy. And surely it wouldn’t require refueling with a sugar-laden 100-calorie drink. Why, then, does the brain’s increased energy consumption during self-control seem to deplete willpower so quickly?


To answer this question, it may be helpful to recall the American banking crisis of 2009. After the 2008 financial meltdown, banks received an influx of money from the government. These funds were supposed to help the banks cover their own financial obligations so they could start lending again. But the banks refused to lend money to small businesses and individual borrowers. They weren’t confident in the money supply, so they hoarded the resources they had. Stingy bastards! It turns out that your brain can be a bit of a stingy bastard, too. The human brain has, at any given time, a very small supply of energy. It can store some energy in its cells, but it is mostly dependent on a steady stream of glucose circulating in the body’s bloodstream. Special glucose-detecting brain cells are constantly monitoring the availability of energy. When the brain detects a drop in available energy, it gets a little nervous. What if it runs out of energy? Like the banks, it may decide to stop spending and save what resources it has. It will keep itself on a tight energy budget, unwilling to spend its full supply of energy. The first expense to be cut? Self-control, one of the most energy-expensive tasks the brain performs. To conserve energy, the brain may become reluctant to give you the full mental resources you need to resist temptation, focus your attention, or control your emotions. University of South Dakota researchers X. T. Wang, a behavioral economist, and Robert Dvorak, a psychologist, have proposed an “energy budget” model of self-control. They argue that the brain treats energy like money. It will spend energy when resources are high, but save energy when resources are dropping. To test this idea, they invited sixty-five adults—ranging in age from nineteen to fifty-one—into the laboratory for a test of their willpower. Participants were given a series of choices between two rewards, such as $120 tomorrow or $450 in a month. One reward was always smaller, but participants would get it faster than the larger reward. Psychologists consider this a classic test of self-control, as it pits immediate gratification against more-favorable long-term consequences. At the end of the study, the participants had the opportunity to win one of their chosen rewards. This ensured that they were motivated to make real decisions based on what they wanted to win. Before the choosing began, the researchers measured participants’ blood sugar levels to determine the baseline status of available “funds” for self-control. After the first round of decisions, participants were given either a regular, sugary soda (to boost blood sugar levels) or a zero-calorie diet soda. The researchers then measured blood sugar levels again, and asked the participants to make another series of choices. The participants who drank the regular soda showed a sharp increase in blood sugar. They also became more likely to delay gratification for the bigger reward. In contrast, blood sugar dropped among the participants who drank the diet soda. These participants were now more likely to choose the immediate gratification of the quicker, smaller reward. Importantly, it wasn’t the absolute level of blood sugar that predicted a participant’s choices—it was the direction of change. The brain asked, “Is available energy increasing or decreasing?” It then made a strategic choice about whether to spend or save that energy.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .