Academia.eduAcademia.edu
Psychological Science http://pss.sagepub.com/ Motivational Versus Metabolic Effects of Carbohydrates on Self-Control Daniel C. Molden, Chin Ming Hui, Abigail A. Scholer, Brian P. Meier, Eric E. Noreen, Paul R. D'Agostino and Valerie Martin Psychological Science published online 12 September 2012 DOI: 10.1177/0956797612439069 The online version of this article can be found at: http://pss.sagepub.com/content/early/2012/09/12/0956797612439069 Published by: http://www.sagepublications.com On behalf of: Association for Psychological Science Additional services and information for Psychological Science can be found at: Email Alerts: http://pss.sagepub.com/cgi/alerts Subscriptions: http://pss.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav >> OnlineFirst Version of Record - Sep 12, 2012 What is This? Downloaded from pss.sagepub.com by guest on September 12, 2012 Psychological Science OnlineFirst, published on September 12, 2012 as doi:10.1177/0956797612439069 Research Article Psychological Science Motivational Versus Metabolic Effects XX(X) 1­–8 © The Author(s) 2012 Reprints and permission: of Carbohydrates on Self-Control sagepub.com/journalsPermissions.nav DOI: 10.1177/0956797612439069 http://pss.sagepub.com Daniel C. Molden1, Chin Ming Hui1, Abigail A. Scholer2 , Brian P. Meier3, Eric E. Noreen3, Paul R. D’Agostino3, and Valerie Martin3 1 Northwestern University, 2University of Waterloo, and 3Gettysburg College Abstract Self-control is critical for achievement and well-being. However, people’s capacity for self-control is limited and becomes depleted through use. One prominent explanation for this depletion posits that self-control consumes energy through carbohydrate metabolization, which further suggests that ingesting carbohydrates improves self-control. Some evidence has supported this energy model, but because of its broad implications for efforts to improve self-control, we reevaluated the role of carbohydrates in self-control processes. In four experiments, we found that (a) exerting self-control did not increase carbohydrate metabolization, as assessed with highly precise measurements of blood glucose levels under carefully standardized conditions; (b) rinsing one’s mouth with, but not ingesting, carbohydrate solutions immediately bolstered self- control; and (c) carbohydrate rinsing did not increase blood glucose. These findings challenge metabolic explanations for the role of carbohydrates in self-control depletion; we therefore propose an alternative motivational model for these and other previously observed effects of carbohydrates on self-control. Keywords self control, motivation, strength model of self-regulation, ego depletion, glucose, mental performance Received 12/9/11; Accepted 1/22/12 People’s effortful regulation of their thoughts and actions— on subsequent tasks that also require self-control declines, that is, their use of self-control—has an enormous impact on even when the initial and subsequent tasks involve different their lives. Studies have repeatedly demonstrated that the more modalities (e.g., perceptual monitoring vs. logical reasoning; self-control individuals typically exercise, the happier, health- Hagger, Wood, Stiff, & Chatzisarantis, 2010). ier, and more successful they are in their professions and rela- tionships from adolescence to old age (Mischel, Shoda, & Rodriguez, 1989; Moffitt et al., 2011; Tangney, Baumeister, & The Energy Model of Self-Control Boone, 2004). Accordingly, researchers have long worked to The metaphorical conceptualization of self-control as a muscle understand who engages in self-control and how people’s self- whose effectiveness lessens with continued use has been enor- control abilities can be improved (Baumeister, Gailliot, mously influential and has greatly advanced thinking and DeWall, & Oaten, 2006; Diamond, Barnett, Thomas, & Munro, research on this topic. However, some researchers have recently 2007). argued that similarities between self-control processes and One primary challenge to improving self-control is that muscle function are not simply metaphorical (Gailliot & whenever people exert self-control, they appear to draw on a Baumeister, 2007): Studies have suggested that just as muscles global pool of self-regulatory resources that is limited and metabolize simple carbohydrates as their primary fuel, self- therefore can be depleted. Using the metaphor of a muscle, control may also rely on carbohydrate metabolization (Gailliot researchers have likened the depletion of self-regulatory et al., 2007; Gailliot, Peruche, Plant, & Baumeister, 2009; resources to the expenditure of energy during physical activ- Masicampo & Baumeister, 2008). That is, these studies have ity, which produces temporary fatigue and difficulty in sus- taining attention and effort (Muraven & Baumeister, 2000). Corresponding Author: Consistent with this metaphor, results from a recent meta- Daniel C. Molden, Northwestern University, 2029 Sheridan Rd., Evanston, IL analysis of 83 studies clearly demonstrated that following 60208 tasks requiring self-control or executive function, performance E-mail: molden@northwestern.edu Downloaded from pss.sagepub.com by guest on September 12, 2012 2 Molden et al. implied that just as physical exertion consumes carbohydrates, Baumeister, Gailliot, & Maner, 2008; Gailliot et al., 2007; exerting self-control depletes the body’s carbohydrate stores, Gailliot et al., 2009; Masicampo & Baumeister, 2008). Com- and just as physical activity can be bolstered when carbohy- pared with findings that self-control reduces blood glucose, drates are ingested, self-control is also enhanced by carbohy- findings that carbohydrate consumption improves self-control drate intake. have been more consistent, and these latter findings match This energy model of limited self-control has several wide- results from other experiments demonstrating that carbohy- ranging implications. For example, individuals with inherited drate ingestion boosts performance on mental and physical or acquired deficiencies in carbohydrate metabolization would tasks (Messier, 2004; Riby, 2004). However, it is unclear presumably face great challenges in achieving success and whether these effects are truly due to connections between car- well-being. Also, despite the current trend of eliminating bohydrate metabolization and brain function (Beedie & Lane, carbohydrate-laden foods from school and office cafeterias, 2012; Gibson, 2007; Kurzban, 2010; Messier, 2004). such foods could actually help sustain learning and productiv- ity. Carefully evaluating this energy model of self-control is therefore broadly important for both public policy and public An Alternative to the Energy Model: health and was the primary objective of the research reported Motivational Effects of Carbohydrates here. Recently, several researchers have discovered an alternative mechanism by which carbohydrates influence physical effort. Multiple experiments have demonstrated that participants who Challenges to the Energy Model briefly rinse their mouths with, but do not ingest, carbohydrate of Self-Control solutions during intense physical activity (e.g., cycling or run- The idea that carbohydrate metabolization plays a role in ning time trials) show significant increases in performance as self-control has not gone unchallenged (Beedie & Lane, 2012; compared with participants who rinse with placebo solutions Kurzban, 2010). In addition to questions about the methodol- containing noncarbohydrate sweeteners (Chambers, Bridge, ogy used to assess such metabolization, criticisms have & Jones, 2009; see Painelli, Nicastro, & Lancha, 2010). focused on the physiological mechanisms proposed to explain Indeed, one study demonstrated that carbohydrate rinsing had how carbohydrate metabolization fuels self-control (also see a greater effect on performance than carbohydrate ingestion Gibson, 2007; Messier, 2004). The specific goals of the pres- did (Pottier, Bouckaert, Gilis, Roels, & Derave, 2010). These ent research were therefore (a) to test whether the behavioral studies indicate that carbohydrates affect persistence and per- effects of exerting self-control are linked to the metabolization formance in nonenergetic ways. of carbohydrates using a more precise methodology than has Furthermore, neuroimaging studies have suggested a spe- been used in previous work, and (b) to evaluate an alternative cific origin for these nonenergetic effects: Carbohydrate mouth motivational model for explaining the link between the inges- rinses activate dopaminergic pathways in the striatum—a region tion of carbohydrates and improved self-control. of the brain associated with responses to rewards (Kringelbach, Although the relationship between the availability of sim- 2004)—whereas artificially sweetened noncarbohydrate mouth ple carbohydrates, such as glucose, in the blood and the utili- rinses do not (Chambers et al., 2009). Thus, the mere sensing of zation of these carbohydrates in the brain is not straightforward carbohydrates in the mouth, whether or not they are ingested, (Gibson, 2007; Messier, 2004), one piece of evidence for the may signal the possibility of reward (i.e., the future availability energy model of limited self-control is that engaging in self- of additional energy), which could motivate, rather than fuel, control appears to lower blood glucose levels (Dvorak & physical effort. In addition, because prolonged physical exer- Simons, 2009; Gailliot et al., 2007). However, other studies tion requires self-control (Morsella, 2005), existing findings have shown contradictory results (Kurzban, 2010). One reason that carbohydrate ingestion boosts mental activity requiring for this inconsistency could be that studies linking carbohy- self-control could also be explained by increases in people’s drate metabolization to self-control have employed commer- motivation to perseverate rather than in their level of energy. cially available Accu-Chek blood glucose monitors (Roche To test whether merely sensing carbohydrates in the mouth Diagnostics, Basel, Switzerland), whose results are less pre- can motivate self-regulation, in Experiments 2 and 3, we cise than those of formal laboratory assessments (Beedie & examined how rinsing with a carbohydrate solution influences Lane, 2012; Khan, Vasquez, Gray, Wians, & Kroll, 2006). self-control. Whereas previous studies have included a 10- Therefore, in Experiment 1, we evaluated how exerting self- to 12-min delay between the ingestion of carbohydrates and control affects both subsequent self-control and blood glucose secondary self-control tasks, to allow time for metabolization levels using best-practice laboratory methods. (Gailliot et al., 2007; Gailliot et al., 2009; Masicampo & A second piece of evidence for the energy model of self- Baumeister, 2008), in Experiments 2 and 3, the secondary self- control is that ingesting carbohydrates after performing a task control task was performed immediately after rinsing, before that requires self-control improves performance on subsequent metabolization could possibly occur. In addition, to ensure the tasks that also require self-control. That is, carbohydrate con- generalizability of the results, we used different, but equally sumption appears to replenish the resources depleted by initial well-validated, measures and manipulations of executive self-control efforts and to sustain future efforts (DeWall, function and self-control in each experiment. Downloaded from pss.sagepub.com by guest on September 12, 2012 Carbohydrates’ Effects on Self-Control 3 Finally, although Experiments 2 and 3 tested whether car- effects of the perceptual-vigilance task on their subsequent bohydrate rinsing improved self-control, they did not examine self-control. In the anagram task, participants were instructed whether simply rinsing one’s mouth with carbohydrates to generate as many words as they could from a set of seven increases blood glucose by inducing the release of endogenous letters. The amount of time participants persisted on the task stores. This question was therefore our focus in Experiment 4. was our primary measure of their continued self-control. Experiment 1: Does Self-Control Results and discussion Consume Carbohydrates? Because time spent on the anagram task was skewed (skew- Method ness = 1.21), analyses were conducted on log-transformed times (Judd & McClelland, 1989), but for ease of exposition, Participants. Eighty-five college students (52 females, 33 raw means are reported here. Our results replicated those of males; mean age = 19.28 years, SD = 1.25) participated in previous research (Hagger et al., 2010): Participants in the return for course credit or payment. To control for initial blood high-depletion condition (M = 3.80 min, SD = 2.74) persisted glucose levels and potential glycemic responses, we (a) re- less than participants in the low-depletion condition did (M = quired that participants weigh at least 110 lb, (b) instructed 4.60 min, SD = 2.41), t(83) = 2.09, p = .04, d = 0.47. participants to abstain from eating for 4 hr and from vigorous However, a 2 (condition: low depletion vs. high depletion) exercise for 24 hr before the experiment, and (c) ran the exper- × 2 (glucose assessment: pretask vs. posttask) mixed analysis iment between 4 p.m. and 7 p.m. of variance (ANOVA) on blood glucose levels with repeated measures on the second factor revealed no main effect of con- Procedure and materials. Participants first received a base- dition and no Condition × Glucose Assessment interaction, line assessment of their blood glucose and then performed a Fs(1, 83) < 2.05, ps > .16. Thus, despite affecting subsequent perceptual-vigilance task requiring either a low or a high level persistence, initially exerting greater self-control did not lead of self-control. After participants’ blood glucose was reas- to greater carbohydrate metabolization. Indeed, overall, there sessed, we had them complete an anagram task to measure was a marginally significant increase in blood glucose levels their continued exertion of self-control. between the pretask assessment (M = 81.27 mg/dl, SD = 7.49) To assess blood glucose both before and after the and the posttask assessment (M = 82.39 mg/dl, SD = 8.68), perceptual-vigilance task, we followed the laboratory refer- F(1, 83) = 2.94, p = .09, d = 0.19. This increase could perhaps ence procedures against which all commercial glucose moni- be explained by a task-related release of cortisol, which tors are tested for accuracy (Khan et al., 2006). Participants increases blood glucose (Miller & Tyrell, 1995), but it cannot placed their hands on a heating pad for 2 min, after which a be explained by the energy model, which predicts decreases in small sample of blood was taken from their finger via capillary blood glucose following the exertion of self-control. puncture. These samples were analyzed in duplicate with a Finally, in contrast to the findings of Gailliot et al. YSI 2700 Glucose/Lactate Analyzer (Yellow Springs Instru- (2007), neither absolute levels of blood glucose following the ments, Yellow Springs, OH), and glucose levels were calcu- perceptual-vigilance task nor changes in blood glucose from lated in milligrams per deciliter. before to after the task were significantly correlated with per- The perceptual-vigilance task was identical to tasks used in sistence on the anagram task in either condition, rs < |.20|, many previous studies of self-control depletion (Baumeister, ps > .17. Thus, our results from Experiment 1, in which we Bratslavsky, Muraven, & Tice, 1998; Hagger et al., 2010). All employed the most sensitive measures of blood glucose avail- participants first received one page of text and were instructed able, add to existing questions about whether the depletion of to cross out every “e.” Those participants assigned to the low- self-control resources involves carbohydrate metabolization depletion condition next received another, similar page of text (Beedie & Lane, 2012; Kurzban, 2010), as proposed by the with the same instructions. Those participants assigned to the energy model. high-depletion condition received another page of text with a different set of instructions asking them to cross out every “e” that was neither adjacent to nor one letter removed from Experiment 2: Must Carbohydrates Be another vowel. Following these new rules required partici- Metabolized to Bolster Self-Control? pants in the high-depletion condition to monitor the text more Our results from Experiment 1 failed to support the predic- vigilantly than participants in the low-depletion condition did tions of the energy model of self-control concerning the and to inhibit the practiced response instantiated by the initial metabolization of carbohydrates during the exertion of self- rules, which increased their expenditure of self-control control. In Experiment 2, we evaluated a second prediction of resources (Baumeister et al., 1998). the energy model: that the ingestion of carbohydrates provides In keeping with many previous studies on self-control additional “fuel” for self-control and reduces the depletion of depletion (Muraven, Tice, & Baumeister, 1998; see Hagger self-control resources. After performing a task that required et al., 2010), following the second blood glucose assessment, either a low or a high level of self-control, participants rinsed we had participants perform an anagram task to assess the their mouths with, but did not ingest, a solution flavored with Downloaded from pss.sagepub.com by guest on September 12, 2012 4 Molden et al. either a carbohydrate-based sweetener or a noncarbohydrate- High Depletion based sweetener. They then immediately performed a second Low Depletion task that required self-control. Any results showing that merely 90 rinsing with carbohydrate solutions bolstered self-regulation 80 would further challenge the energy model’s account of carbo- hydrates’ effects on self-control. 70 60 Persistence (s) Method 50 Participants. Forty-five university students participated in 40 return for course credit. Data from 1 participant were elimi- 30 nated because the participant’s responses on a baseline self- control measure were extreme (4.59 SD above the grand 20 mean). The final sample therefore consisted of 44 participants 10 (28 females, 16 males; mean age = 18.84 years, SD = 0.57). All participants abstained from eating for at least 4 hr prior to 0 the experiment and took part between 9 a.m. and 12 p.m. or Noncarbohydrate Rinse Carbohydrate Rinse between 4 p.m. and 7 p.m. The time at which participants took Fig. 1. Results from Experiment 2: participants’ persistence in squeezing part in the experiment had no moderating effects on any of the a handgrip immediately after rinsing their mouths for 5 s with a solution analyses reported (all ps > .28). containing a noncarbohydrate-based sweetener or a carbohydrate-based sweetener. The rinse was administered after an initial perceptual-vigilance Procedure and materials. To control for baseline differences task that resulted in either a low or a high depletion of self-control resources. Error bars represent 95% confidence intervals. in self-control resources, we had participants begin the experi- ment by squeezing together the handles of a high-tension hand- grip to suspend a wad of paper in the air for as long as possible. between-participants analysis of covariance (ANCOVA) was Persistence on this task despite growing discomfort has been performed on persistence times for the second handgrip task, widely used in previous studies as an index of self-control with persistence times for the first handgrip task included as a (Muraven & Slessareva, 2003; see Hagger et al., 2010). covariate. Results revealed only a significant interaction, F(1, Participants next completed the same perceptual-vigilance 39) = 4.54, p = .04. task used in Experiment 1, in either the low- or the high- Figure 1 shows that among participants who rinsed with the depletion condition. Then, following previously established noncarbohydrate solution, those in the high-depletion condi- procedures (Chambers et al., 2009; see Painelli et al., 2010), tion displayed significantly reduced persistence compared we gave participants in the carbohydrate-rinse condition a cup with those in the low-depletion condition, F(1, 39) = 6.60, containing 25 ml of a 6.4% table-sugar solution and gave par- p = .01, d = 0.77, a result replicating the typically observed ticipants in the noncarbohydrate-rinse condition a cup con- depletion of self-control after self-regulatory exertion (Hagger taining 25 ml of a 3.2% solution of Equal, a noncarbohydrate, et al., 2010). However, there were no differences between the aspartame-based sweetener approximately twice as sweet as high- and low-depletion conditions among participants who sugar. Participants rinsed their mouths with all 25 ml of the rinsed with the carbohydrate solution, F(1, 39) = 0.22, p = .64, solution for 5 s and then spit it back into the empty cup. d = 0.17. Furthermore, in the high-depletion condition, rinsing Immediately after rinsing, participants completed a second with the carbohydrate solution produced significantly greater trial of the handgrip task; the duration of each participant’s persistence than did rinsing with the noncarbohydrate solu- persistence on this task served as our measure of his or her tion, F(1, 39) = 4.12, p = .05, d = 0.63, whereas the rinsing self-control following the depletion and rinsing manipula- manipulation had no effect in the low-depletion condition, tions. Finally, participants completed the Brief Mood Intro- F(1, 39) = 0.90, p = .35, d = 0.28. Thus, previous findings spection Scale (Mayer & Gaschke, 1988); rated the rinsing concerning the effects of carbohydrates on depleted self- solution’s sweetness, refreshingness, and tastiness, using control (Gailliot et al., 2007; Gailliot et al., 2009; Masicampo scales from 1 (very low) to 7 (very high; α = .58); and reported & Baumeister, 2008) were replicated in our experiment, even what they thought was used to flavor the solution. when carbohydrates were not ingested and when there was not sufficient time for them to be metabolized (cf. Chambers et al., 2009). Our results from Experiment 2 thus indicated that car- Results and discussion bohydrates’ effects on self-control are not necessarily related Because persistence times were not highly skewed in this to metabolic consumption, as proposed by the energy model, experiment (skewness = 0.75), untransformed data were used and can operate through nonenergetic mechanisms. for all analyses (Judd & McClelland, 1989). A 2 (depletion: Participants’ mood and ratings of the taste of the solution low vs. high) × 2 (rinse: carbohydrate vs. noncarbohydrate) did not differ across conditions, Fs(1, 40) < 1.43, ps > .24, Downloaded from pss.sagepub.com by guest on September 12, 2012 Carbohydrates’ Effects on Self-Control 5 ds < 0.36. Overall, when guessing what had been used to fla- Schmeichel, Vohs, & Baumeister, 2003). In this task, partici- vor the solution with which they had rinsed, 65% of partici- pants watched a 6-min video that showed a woman speaking pants said sugar, 23% said some other type of natural flavoring while a series of words appeared at the bottom of the screen (e.g., fruit or tea), 7% said artificial sweetener, and 5% said for 10 s each. All participants were instructed to keep their they did not know. Guesses did not differ between participants attention focused on the woman’s face instead of the words who had rinsed with the sugar solution and those who had (i.e., to exert self-control to inhibit their natural orienting rinsed with the aspartame solution, χ2(3, N = 44) = 4.83, p = response). Previous research has demonstrated that perform- .18. ing this task induces a relatively high depletion of self-control resources (Hagger et al., 2010). Next, participants rinsed with either a carbohydrate-sweet- Experiment 3: Must Carbohydrates Be ened or a noncarbohydrate-sweetened solution, following the Metabolized to Bolster Cognitive as Well as same procedures used in Experiment 2. Immediately after- Physical Self-Control? ward, they completed 96 more trials of the Stroop task. Stroop- Our results from Experiment 2 demonstrated that carbohy- interference scores on these latter trials served as our primary drate metabolism is not necessary for persistence on a physical measure of participants’ self-control following the depletion task. In Experiment 3, we tested whether the same was true for and rinsing manipulations. Finally, participants completed the persistence on a cognitive task. Brief Mood Introspection Scale and rated the taste of the solu- tion with which they had rinsed, following the same proce- dures used in Experiment 2. Method Participants. Thirty-one college students (22 females, 9 males; mean age = 18.58 years, SD = 0.85) participated in Results and discussion return for course credit. All participants abstained from eating Results are shown in Figure 2. Because Stroop-interference for at least 4 hr prior to the experiment and took part between scores were not highly skewed (skewness = 0.69), untrans- 9 a.m. and 12 p.m. or between 4 p.m. and 7 p.m. The time at formed data were used for all analyses. A one-way ANCOVA which participants took part in the experiment had no moder- on participants’ Stroop-interference scores from the second ating effects on any of the analyses reported (all ps > .16). Stroop task, with scores from the first Stroop task included as a covariate, revealed that interference was significantly Procedure and materials. To control for baseline differences lower following carbohydrate rinsing than following noncar- in self-control resources, we had participants begin the experi- bohydrate rinsing, F(1, 28) = 5.02, p = .03, d = 0.73. Again, ment by completing the color-word Stroop task (Stroop, 1935). there were no differences between conditions in the valence or This task involves identifying the color of letter strings as quickly and accurately as possible; on some trials, the letter 140 strings form words for colors that are incongruent with the color in which the letters are displayed (e.g., the word “blue” 120 displayed in red type), which creates response interference. Response Latency (ms) The speed with which people overcome this interference has 100 been widely used as a measure of executive function and self- control (Gailliot et al., 2007; Richeson, Baird, Gordon, Heath- 80 erton, & Wyland, 2003; see Hagger et al., 2010). The task consisted of 24 trials. Participants were instructed to press a 60 response key whose color matched the color of the letters in each trial as quickly as possible without sacrificing accuracy. 40 We calculated participants’ Stroop-interference scores by sub- tracting each participant’s mean response latency for trials on 20 which letter strings did not form words (i.e., “xxxxxx”) from his or her mean response latency for trials on which letter 0 strings formed color words incongruent with the display color Noncarbohydrate Rinse Carbohydrate Rinse (Job, Dweck, & Walton, 2010; Richeson et al., 2003). Higher Fig. 2.  Results from Experiment 3: mean response latencies for incongruent interference scores indicated greater failure to overcome trials of the color-word Stroop task; the task was completed immediately response interference and thus less self-control. after participants had rinsed their mouths for 5 s with a solution containing a noncarbohydrate-based or a carbohydrate-based sweetener. The rinse was Participants next performed a perceptual-vigilance task administered after an initial perceptual-vigilance task that resulted in a high that has been frequently used in previous studies examining depletion of self-control resources. Error bars represent 95% confidence the depletion of self-control resources (Hagger et al., 2010; intervals. Downloaded from pss.sagepub.com by guest on September 12, 2012 6 Molden et al. arousal of participants’ mood or in ratings of the taste of the Results and discussion solution, Fs(1, 29) < 1.55, ps > .22, ds < 0.45. Our findings in A 2 (condition: ingest vs. rinse) × 2 (glucose assessment: pre- Experiment 3 thus replicated those from Experiment 2; fur- task vs. posttask) mixed ANOVA on blood glucose levels with thermore, given that the effects of carbohydrates on self- repeated measures on the second factor revealed an interac- control emerged immediately and in the absence of ingestion, tion, F(1, 18) = 52.06, p < .001. Although, as expected, blood our results once again indicated that metabolization is not nec- glucose increased substantially among participants who had essary for carbohydrates to bolster self-control, which runs ingested carbohydrates (pretask: M = 84.10 mg/dl, SD = 6.76; contrary to the energy model of self-control. posttask: M = 113.97 mg/dl, SD = 15.69), F(1, 18) = 119.57, p < .001, d = 2.58, it did not increase among participants who Experiment 4: Does Rinsing With had rinsed with the carbohydrate solution (pretask: M = 81.78 Carbohydrates Release Endogenous mg/dl, SD = 6.01; posttask: M = 83.78 mg/dl, SD = 9.23), F(1, 18) = 0.53, p = .47, d = 0.17. Energy Stores? In Experiments 2 and 3, carbohydrates were never ingested and could not have been metabolized in the short time between General Discussion rinsing and subsequent exertions of self-control; it is therefore Altogether, the results of the four experiments presented here likely that the observed effects of carbohydrates on self- challenge the energy model of self-control. Participants who control were due to nonenergetic mechanisms. However, rins- exerted greater self-control persisted less on subsequent tasks ing one’s mouth with carbohydrates could perhaps increase but did not show any evidence of increased carbohydrate blood glucose by immediately prompting the release of endog- metabolization, as assessed using the most sensitive measures enous carbohydrate stores. If this type of mechanism were available. Furthermore, participants who rinsed their mouths responsible for the results of our experiments, our findings with, but did not ingest, carbohydrate solutions showed imme- would still be broadly consistent with the energy model of diate boosts in self-control, and rinsing itself did not increase self-control. Therefore, in Experiment 4, we directly tested the blood glucose levels. These findings are all consistent with a effects of rinsing with a carbohydrate solution on blood glu- motivational, rather than metabolic, role of carbohydrates in cose levels. self-control (Painelli et al., 2010). Although our results demonstrate that carbohydrate metab- olization is not necessary to sustain self-control, they do not Method rule out the possibility that carbohydrate metabolization could Participants. Twenty college students (11 females, 9 males; benefit effort and performance in certain circumstances. It is mean age = 19.30 years, SD = 1.26) participated in return for also possible that sustained rinsing with carbohydrates has course credit or payment. As in Experiment 1, to control for diminishing effects over time. However, given the inconsistent initial blood glucose levels and potential glycemic responses, findings concerning the effect of mental effort on blood glu- we (a) required that participants weigh at least 110 lb, cose levels, and the clear evidence for alternative mechanisms (b) instructed participants to abstain from eating for 4 hr and provided by our results and the results of prior work (Beedie & from vigorous exercise for 24 hr before the experiment, and Lane, 2012; Kurzban, 2010; Painelli et al., 2010), we believe (c) ran the experiment between 4 p.m. and 7 p.m. that a motivational rather than metabolic model of carbohy- drates’ effects on self-control offers a superior explanation for Procedure and materials. All participants received 350 ml of the available data. a 21.4% glucose (maltodextrin) solution. This solution was An additional reason to favor motivational models is that substantially greater in quantity and had a substantially greater many other findings concerning the depletion of self-control concentration of glucose than the carbohydrate solution used are also better explained by motivational models than by in the previous experiments, and was designed to provide a energy models. For example, studies have shown that the strong test of the endogenous-release hypothesis. In the ingest depleting effects of initial efforts at self-control are reversed condition, participants drank the entire volume of the solution. by increased incentives for performance on subsequent tasks In the rinse condition, participants took a mouthful of the solu- and by positive affect (Muraven & Slessareva, 2003; Tice, tion, swished it around for 5 s, and then spit it into an empty Baumeister, Shmueli, & Muraven, 2007). Depletion is also cup, repeating these steps until they had rinsed with the entire produced by mere perceptions of effortful expenditures of volume of the solution. Following ingestion or rinsing, partici- self-control (Clarkson, Hirt, Jia, & Alexander, 2010; Wan & pants completed unrelated filler tasks for 12 min. We assessed Sternthal, 2008) and by expectations of future needs for self- participants’ blood glucose levels prior to rinsing or ingestion control (Muraven, Shmueli, & Burkley, 2006), and it occurs and after the 12-min delay using the same laboratory proce- only when people believe that their self-control abilities dures used in Experiment 1. are limited (Job et al., 2010). Thus, although engaging in Downloaded from pss.sagepub.com by guest on September 12, 2012 Carbohydrates’ Effects on Self-Control 7 self-control often leaves people less willing to subsequently Gailliot, M. T., & Baumeister, R. F. (2007). The physiology of will- exert further self-control (or more conscious about conserving power: Linking blood glucose to self-control. Personality and what they believe is a limited resource), it does not leave them Social Psychology Review, 11, 303–327. without the energy to do so when properly motivated. Gailliot, M. T., Baumeister, R. F., DeWall, C. N., Maner, J. K., Plant, In conclusion, our results help to clarify the psychological E. A., Tice, D. M., . . . Schmeichel, B. J. (2007). Self-control and physiological mechanisms responsible for declining self- relies on glucose as a limited energy source: Willpower is more control following continued exertion, even when exertion is than a metaphor. Journal of Personality and Social Psychology, split between different tasks or goals. Findings indicating that 92, 325–336. this decline is better explained by deficits in motivation than Gailliot, M. T., Peruche, B. M., Plant, E. A., & Baumeister, R. F. by decreased energy or ability suggest the possibility of new, (2009). Stereotypes and prejudice in the blood: Sucrose drinks promising interventions. Indeed, interventions focused on sus- reduce prejudice and stereotyping. Journal of Experimental taining or altering self-control motivations could be developed Social Psychology, 45, 288–290. to increase people’s self-control and thus improve outcomes Gibson, E. L. (2007). Carbohydrates and mental function: Feeding or throughout their lives. impeding the brain? Nutrition Bulletin, 32, 71–83. Hagger, M. S., Wood, C., Stiff, C., & Chatzisarantis, N. L. (2010). Acknowledgments Ego depletion and the strength model of self-control: A meta- The first two authors contributed equally to this manuscript. We analysis. Psychological Bulletin, 136, 495–525. thank the organizers of and contributors to the 23rd Annual Duck Job, V., Dweck, C., & Walton, G. M. (2010). Ego depletion—is it Conference on Social Cognition (June 2011) in Duck, NC, for their all in your head? Implicit theories about willpower affect self- feedback on the experiments described here. regulation. Psychological Science, 21, 1686–1693. Judd, C. M., & McClelland, G. H. (1989). Data analysis: A model Declaration of Conflicting Interests comparison approach. New York, NY: Harcourt Brace Jovano­ The authors declared that they had no conflicts of interest with vich. respect to their authorship or the publication of this article. Khan, A. I., Vasquez, Y., Gray, J., Wians, F. H., Jr., & Kroll, M. H. (2006). The variability of results between point-of-care testing References glucose meters and the central laboratory analyzer. Archives of Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. Pathology & Laboratory Medicine, 130, 1527–1532. (1998). Ego depletion: Is the active self a limited resource? Jour- Kringelbach, M. L. (2004). Food for thought: Hedonic experience nal of Personality and Social Psychology, 74, 1252–1265. beyond homeostasis in the human brain. Neuroscience, 126, Baumeister, R. F., Gailliot, M., DeWall, C. N., & Oaten, M. (2006). 807–819. Self-regulation and personality: How interventions increase regu- Kurzban, R. (2010). Does the brain consume additional glucose dur- latory success, and how depletion moderates the effects of traits ing self-control tasks? Evolutionary Psychology, 8, 244–259. on behavior. Journal of Personality, 74, 1773–1802. Masicampo, E. J., & Baumeister, R. F. (2008). Toward a physiology Beedie, C. J., & Lane, A. M. (2012). The role of glucose in self-control: of dual-process reasoning and judgment: Lemonade, willpower, Another look at the evidence and an alternative conceptualization. and expensive rule-based analysis. Psychological Science, 19, Personality and Social Psychology Review, 16, 143–153. 255–260. Chambers, E. S., Bridge, M. W., & Jones, D. A. (2009). Carbohydrate Mayer, J. D., & Gaschke, Y. N. (1988). The experience and meta- sensing in the human mouth: Effects on exercise performance and experience of mood. Journal of Personality and Social Psychol- brain activity. Journal of Physiology, 587, 1779–1794. ogy, 55, 102–111. Clarkson, J. J., Hirt, E. R., Jia, L., & Alexander, M. B. (2010). When Messier, C. (2004). Glucose improvement of memory: A review. perception is more than reality: The effects of perceived versus European Journal of Pharmacology, 490, 33–57. actual resource depletion on self-regulatory behavior. Journal of Miller, W., & Tyrell, J. (1995). Endocrinology and metabolism. New Personality and Social Psychology, 98, 29–46. York, NY: McGraw-Hill. DeWall, C. N., Baumeister, R. F., Gailliot, M. T., & Maner, J. K. Mischel, W., Shoda, Y., & Rodriguez, M. I. (1989). Delay of gratifi- (2008). Depletion makes the heart grow less helpful: Helping as cation in children. Science, 244, 933–938. a function of self-regulatory energy and genetic relatedness. Per- Moffitt, T. E., Arseneault, L., Belsky, D., Dickson, N., Hancox, R. J., sonality and Social Psychology Bulletin, 34, 1653–1662. Harrington, H., . . . Caspi, A. (2011). A gradient of childhood self- Diamond, A., Barnett, W. S., Thomas, J., & Munro, S. (2007). Pre- control predicts health, wealth, and public safety. Proceedings of school program improves cognitive control. Science, 318, 1387– the National Academy of Sciences, USA, 108, 2693–2698. 1388. Morsella, E. (2005). The function of phenomenal states: Supramodu- Dvorak, R. D., & Simons, J. S. (2009). Moderation of resource deple- lar interaction theory. Psychological Review, 112, 1000–1021. tion in the self-control strength model: Differing effects of two Muraven, M., & Baumeister, R. F. (2000). Self-regulation and deple- modes of self-control. Personality and Social Psychology Bul- tion of limited resources: Does self-control resemble a muscle? letin, 35, 572–583. Psychological Bulletin, 126, 247–259. Downloaded from pss.sagepub.com by guest on September 12, 2012 8 Molden et al. Muraven, M., Shmueli, D., & Burkley, E. (2006). Conserving self- Richeson, J. A., Baird, A. A., Gordon, H. L., Heatherton, T. F., & Wyland, control strength. Journal of Personality and Social Psychology, C. L. (2003). An fMRI investigation of the impact of interracial 91, 524–537. contact on executive function. Nature Neuroscience, 6, 1323–1328. Muraven, M., & Slessareva, E. (2003). Mechanisms of self-control Schmeichel, B. J., Vohs, K. D., & Baumeister, R. F. (2003). Intel- failure: Motivation and limited resources. Personality and Social lectual performance and ego depletion: Role of the self in logical Psychology Bulletin, 29, 894–906. reasoning and other information processing. Journal of Personal- Muraven, M., Tice, D. M., & Baumeister, R. F. (1998). Self-control as ity and Social Psychology, 85, 33–46. limited resource: Regulatory depletion patterns. Journal of Per- Stroop, J. R. (1935). Studies of interference in serial verbal reactions. sonality and Social Psychology, 74, 774–789. Journal of Experimental Psychology, 18, 643–662. Painelli, V. S., Nicastro, H., & Lancha, A. H., Jr. (2010). Carbohy- Tangney, J. P., Baumeister, R. F., & Boone, A. L. (2004). High self- drate mouth rinse: Does it improve endurance exercise perfor- control predicts good adjustment, less pathology, better grades, mance? Nutrition Journal, 9, 33. and interpersonal success. Journal of Personality, 72, 271–324. Pottier, A., Bouckaert, J., Gilis, W., Roels, T., & Derave, W. (2010). Tice, D. M., Baumeister, R. F., Shmueli, D., & Muraven, M. (2007). Mouth rinse but not ingestion of a carbohydrate solution improves Restoring the self: Positive affect helps improve self-regulation 1-h cycle time trial performance. Scandinavian Journal of Medi- following ego depletion. Journal of Experimental Social Psy- cine & Science in Sports, 20, 105–111. chology, 43, 379–384. Riby, L. M. (2004). The impact of age and task domain on cognitive Wan, E. W., & Sternthal, B. (2008). Regulating the effects of deple- performance: A meta-analytic review of the glucose facilitation tion through monitoring. Personality and Social Psychology Bul- effect. Brain Impairment, 5, 145–165. letin, 34, 32–46. Downloaded from pss.sagepub.com by guest on September 12, 2012