Achieving the Same for Less: Improving Mood Depletes Blood Glucose for People with Poor (but not Good) Emotion Control.

Running head: EMOTION REGULATION AND GLUCOSE DEPLETION BRIEF REPORT Achieving the Same for Less: Improving Mood Depletes Blood Glucose for People with Poor (but not Good) Emotion Control Word count: 3425 Keywords. Emotion regulation, self-control, blood glucose, ego-depletion, resources Abstract Previous studies have found that acts of self-control like emotion regulation deplete blood glucose levels. The present experiment investigates the hypothesis that the extent to which people’s blood glucose levels decline during emotion regulation attempts is influenced by whether they believe themselves to be good or poor at emotion control. We found that although good and poor emotion regulators were equally able to achieve positive and negative moods, the blood glucose of poor emotion regulators was reduced after performing an affect-improving task, whereas the blood glucose of good emotion regulators remained unchanged. As evidence suggests that glucose is a limited energy resource upon which self-control relies, the implication is that good emotion regulators are able to achieve the same positive mood with less cost to their self-regulatory resource. Thus, depletion may not be an inevitable consequence of engaging in emotion regulation. Achieving the Same for Less: Improving Mood Depletes Blood Glucose for People with Poor (but not Good) Emotion Control Emotion regulation enables people to improve or worsen their feelings in a socially appropriate manner and facilitates the pursuit of hedonistic and instrumental goals (Tamir, Mitchell, & Gross, 2008). Like other forms of self-control, emotion regulation appears to depend upon a limited self-control resource (Baumeister, Bratslavsky, Muraven, & Tice, 1998), and sustained periods of emotion regulation are known to be psychologically draining (Grandey, 2003). The present research, however, tests the idea that regulating one’s feelings may not inevitably diminish the self-control resource of blood glucose. People who believe that they are generally successful at emotion control in their daily lives (‘good’ emotion regulators) are likely to have developed efficient mechanisms for managing their feelings and, as a consequence, should expend less glucose in garnering the same emotional outcome than people who believe that they are less successful at emotion control (‘poor’ emotion regulators). That is, being good at emotion control should enable people to achieve the same (feeling state) for less (cost to their blood glucose resource). The idea that emotion regulation is costly or ‘ego-depleting’ derives from the strength model of self-control (Baumeister, Vohs, & Tice, 2007). The strength model conceptualizes self-regulation as a conscious and effortful process that draws on a limited energy resource. In support of this idea, studies have shown that emotion regulation can lead to fatigue, memory impairment, and reduced self-control performance (Baumeister et al., 1998; Hagger, Wood, Stiff, & Chatzisarantis, 2010; Richards & Gross, 2000). More recent research has examined the physiological underpinnings of ego-depletion. Based on evidence that metabolization of glucose from the bloodstream enables the brain to carry out executive functions such as self-control, blood glucose has been implicated as the limited energy resource upon which self-control relies (Gailliot & Baumeister, 2007). Acts of self-control such as suppressing stereotypes have been shown to reduce blood glucose levels, and increasing glucose levels by drinking sugary lemonade appears to facilitate self-control (Gailliot, Baumeister, DeWall et al. 2007). Recent evidence suggests that emotion regulation, like other types of self-control, may also rely on glucose as a limited resource. For instance, Dvorak and Simons (2009) demonstrated that suppressing emotional responses to video clips reduced blood glucose more than merely watching the videos. Although controlling one’s emotions may often be a conscious and effortful process, recent research suggests that emotion regulation can also be instigated automatically (Williams, Bargh, Nocera, & Gray, 2009). Bargh’s (1990) automotive model posits that when a behavior has repeatedly and consistently been performed in the same situational context, action gradually comes under stimulus control such that merely encountering the relevant contextual cues can initiate the behavior automatically (i.e., swiftly and effortlessly; see also Wood & Neal’s, 2007, account of habit formation). Similarly, counteractive control theory (CCT; Trope & Fishbach, 2000) suggests that temptations (e.g., chocolate) can prime higher-order goals (e.g., to be slim) and initiate self-control behaviors (e.g., refusing the offer of a piece of chocolate) automatically. The idea that emotion control can become automatized has important implication for glucose consumption. Automatic behaviors are less reliant on regulatory resources than are conscious, effortful behaviors (Schmeichel & Baumeister, 2004) and so automatic emotion regulation should require less glucose (self-control resource) compared to controlled emotion regulation. However, not everyone exhibits automatic self-control. Studies suggest that only people who are generally successful at self-regulation will have developed efficient responses to goal-relevant cues and thus perform self-control in an automatic manner. For example, in CCT studies of self-control over dieting and academic success, Fishbach, Friedman, and Kruglanski (2003, Studies 3-4) only found a link between temptations and the automatic activation of higher-order goals among participants who rated themselves as generally good at self-regulation in the relevant domain; participants who rated themselves as poor self-regulators did not show this priming effect, and presumably had to rely on conscious, effortful self-control in order to handle temptations. If good self-regulators are able to perform self-control in a more efficient manner, then they should experience less blood glucose depletion during self-control tasks like emotion regulation. Only one study so far has examined this idea, but Dvorak and Simons (2009) found that the caliber of participants’ self-control over behavior and cognitions did not affect blood glucose depletion arising from emotion regulation. However, given that a person who is good at one type of self-control is not necessarily good at other types (Wills, Walker, Mendoza, & Ainette, 2006), we hypothesized that it may only be people with good emotion control who can regulate their emotions without depleting their glucose resources. The present study combines ideas from the strength model of self-control with evidence that good self-control automatizes behavioral performance to test the ‘same for less’ hypothesis in the domain of emotion regulation. Blood glucose measures were taken before and after participants undertook two emotion regulation tasks – an affect-improving task and an affect-worsening task. We focused on emotion regulation tasks that all participants should have enough resource to successfully perform (i.e., to regulate mood over a 15-minute period), in order to test whether participants with good emotion control expend less glucose in attaining the same outcome compared to participants with poor emotion control. Evidence indicates that worsening emotions is easier and less effortful than is improving one’s emotions (Ochsner, Ray, Cooper et al., 2004; Webb, Miles, & Sheeran, in press) and so worsening affect provides a useful control condition for the more resource-depleting affect-improving condition.1 We expected no change in blood glucose for participants with either good or poor emotion control in the affect-worsening (control) condition, as this is an easy task that should not expend blood glucose. For the affect-improving (experimental) task, we predicted that the blood glucose of participants with good emotion control would remain intact, whereas participants with poor emotion control would show a significant decline in their blood glucose levels. Method Participants Thirty-one staff and students at a UK University (M age = 34.40 years, 22 women) signed up to take part in a study that ostensibly investigated the emotional and physiological effects of listening to music. Participants were paid £20 (around $30). Volunteers with disorders relating to blood glucose (e.g., diabetes) were excluded. Participants refrained from eating or strenuously exercising during the three hours before participation, to allow glucose levels to stabilize (Gailliot et al., 2007). Procedure Participants attended two experimental sessions, scheduled one week apart but at the same time of day and on the same day of the week in order to restrict extraneous effects on blood glucose levels. Participants were assigned to different regulation instructions (affect-improving or affect-worsening) in each session. The order of sessions was counterbalanced, and no order effects were observed.2 In each session the experimenter first took participants’ baseline blood glucose reading (mmol/l) using an Accu-Chek compact meter. Next, a two-minute breathing exercise (Varley, Webb, & Sheeran, 2011) was used to reduce any anxiety arising from the blood glucose measurement. Participants then indicated the extent to which they were currently feeling ‘happy’, ‘energetic’, ‘calm’, ‘gloomy’, ‘sluggish’ and ‘anxious’ (items based on the UWIST checklist, Matthews, Jones, & Chamberlain, 1990) on a seven-point Likert-type scale ranging from ‘not at all’ to ‘a great extent’ (mean α across baseline and follow-up measures = .72). Participants were then given a set of emotion regulation instructions to follow during a ‘music task’ that was designed to give participants a context that would motivate emotion regulation. For the affect-improving condition, participants were told that the music they were about to hear was hard to feel happy to, but that they should try to make themselves feel happy and maintain the happy mood throughout the task. The affect-worsening instructions were identical, except participants were told that the music was hard to feel sad to, but that they should try to get into and maintain a sad mood throughout the task. Participants could use any strategies that they liked to regulate their feelings. In each session, participants listened to one of two collections of classical music lasting 15 minutes in total, with the collections counterbalanced across conditions. All of the music pieces received mean ratings of ‘neutral’ along three dimensions of affect (gloomy-happy; calm-anxious; sluggish-energetic) in a pilot study (N = 10). Following the emotion regulation task, participants completed a follow-up mood measure and a second blood glucose reading was taken. Participants’ perceptions of the caliber of their everyday emotion control were assessed using an adapted version of Fishbach et al.’s (2003) three-item measure of self-regulatory success. We chose this scale because it has been used as an indicator of perceived success at self-regulation in various domains including dieting and academics (Fishbach et al., 2003) and has been shown to predict actual success in such domains (e.g., negatively correlating with body mass index in Kroese, Adriaanse, Evers, & De Ridder, 2011). Our adapted items asked participants: In general… “how successful are you at controlling your emotions”, “how difficult do you find it to keep your feelings in check” (reverse-coded), and “how good are you at keeping your feelings under control”. Responses were given on a seven-point Likert-type scale from ‘not at all’ to ‘very’ (α = .66). Test-retest reliability over one week was satisfactory (r = .79, p < .01). Evidence for the validity of the adapted scale was sought in a pilot study with a separate sample (N = 105 students from the same university), where the scale was shown to predict success at emotion regulation in everyday life, as indicated by experiencing higher levels of positive emotions in general (measured using the PANAS scale, Watson, Clark, & Tellegen, 1988; r = .19, p < .05). In our main study, we also ruled out the possibility that scores on the adapted measure reflected perceptions of success in the specific tasks performed during the experiment (as opposed to perceptions of general success at emotion regulation), because scores did not correlate with participants’ self-rated success on the emotion regulation tasks, assessed using single-items (affect-improving, r = .04, p = .84; affect-worsening, r = .10, p = .59). Results Effects on Mood To assess participants’ success at regulating their emotions during the experiment, we conducted a 2-within (condition: affect-improving vs. affect-worsening) by 2-within (time of measurement: baseline vs. follow-up) ANCOVA with the caliber of participants’ emotion control as the covariate and mood as the dependent variable. The covariate was mean-centered prior to analysis as a repeated-measures correction (Delaney & Maxwell, 1981). There was no three-way interaction between condition, time of measurement, and caliber of emotion control, F(1, 28) = 0.19, p = .67, hp2 = .01, but there were significant main effects of condition, F(1, 28) = 50.64, p < .01, hp2 = .64, and time of measurement, F(1, 28) = 15.76, p < .01, hp2 = .36, and there was a significant two-way interaction between condition and time of measurement, F(1, 28) = 73.49, p < .01, hp2 = .72. Simple effects analyses (see Table 1) indicated that participants’ moods were more positive after, compared to before, the affect-improving task, F(1, 30) = 11.53, p < .01, hp2 = .28, and less positive after, compared to before, the affect-worsening task, F(1, 29) = 59.10, p < .01, hp2 = .67. A median split that divided participants into good versus poor emotion control groups confirmed that the two-way interaction between condition and time of measurement was significant for both the good emotion control group, F(1, 14) = 36.20, p < .01, hp2 = .72, and the poor emotion control group, F(1, 14) = 37.26, p < .01, hp2 = .73. Thus, the findings indicate that the key goal of our design was met – irrespective of the caliber of their emotion control participants were able to get themselves into better or worse moods. 3 (Table 1 about here) Effects on Blood Glucose Next, we examined the impact of achieving these mood changes on participants’ blood glucose. A 2-within (condition) by 2-within (time of measurement) ANCOVA with the caliber of participants’ emotion control as the covariate and blood glucose as the dependent variable revealed a significant three-way interaction between condition, time of measurement, and caliber of emotion control, F(1, 27) = 3.59, p < .05, hp2 = .12 (see Figure 1). All other effects were non-significant, Fs < 1.10, ps > .05, hp2 < .04. Our initial exploration of the interaction examined the equivalence of participants’ pre-task glucose levels. Findings showed that there were no differences in pre-task glucose levels between the two conditions, F(1, 29) = 0.12, p = .73, hp2 = .01, and that emotion control caliber was not correlated with pre-task glucose levels in either condition (affect-improving task, r = -.15, p = .44; affect-worsening-task, r = .10, p = .59), indicating that all participants had the same levels of glucose at the start of both tasks. (Figure 1 about here) To further decompose the interaction, we examined the two-way interaction between condition and time of measurement separately for participants with good versus poor emotion control, using the median split of this variable. There was a significant two-way interaction among participants with poor emotion control, F(1, 13) = 7.05, p < .05, hp2 = .35, but not among participants with good emotion control, F(1, 14) = 0.56, p = .47, hp2 = .04. Planned comparisons confirmed our expectations. Glucose levels did not change over the course of the affect-worsening task for participants with good emotion control (Ms = 5.21 and 5.17, SDs = 0.46 and 0.68, respectively), or for participants with poor emotion control (Ms = 5.05 and 5.17, SDs = 0.43 and 0.41, respectively), Fs < 3.09, ps > .05, hp2 < .18. Thus, affect-worsening was not depleting for either group. In the case of the affect-improving task, blood glucose levels were significantly lower after the affect-improving task (M = 5.00, SD = 0.53) than before this task (M = 5.21, SD = 0.50) for participants with poor emotion control, F(1, 13) = 6.68, p < .05, hp2 = .34. However, as predicted, participants with good emotion control showed no such decline over the course of the task (Ms = 5.12 and 5.19, SDs = 0.46 and 0.36, respectively), F(1, 14) = 1.59, p = .23, hp2 = .10. Thus, participants with poor emotion control attained the same improvement in mood as those with good emotion control but expended more glucose in the process. Discussion Emotion regulation is an important form of self-control, but can lead to fatigue and reduced self-control performance. Our findings suggest that this depletion is not inevitable, however. We found that attempting to improve feelings only diminished blood glucose levels among people who perceived themselves as having a lower level of emotion control in their everyday lives. People who believed they were generally successful at emotion regulation were able to achieve the same positive feeling state for less cost to their blood glucose. Given that blood glucose is thought to represent an energy resource upon which self-control relies (Gailliot & Baumeister, 2007), this finding can be explained with reference to implicit self-control theories, which suggest that people with good emotion control are able to evoke self-control in an automatic and thus more resource-efficient manner (e.g., Fishbach et al., 2003). In the present context, we gave participants a ‘negative’ stimulus and asked them to improve their emotions. Good emotion regulators may have learned associations between the use of strategies to improve their feelings and the contextual features that usually co-occur with them (e.g., being exposed to negative stimuli) such that perceiving the relevant contextual feature automatically triggered the associated response (Bargh, 1990; Wood & Neal, 2007). It is also possible that, through experience, good emotion regulators may have learned to use affect-improving strategies that are more efficient (e.g., reappraisal) as opposed to strategies that are relatively more effortful to enact (e.g., suppression) (Richards & Gross, 2000). In contrast, poor emotion regulators may not have developed such efficient mechanisms or strategies for control, meaning that they had to devote conscious effort to improving their emotions and therefore depleted their blood glucose levels. The implication is that in situations where poor emotion regulators are required to improve their feelings (e.g., as part of a job role; Grandey, 2003), the concomitant depletion in blood glucose may undermine other activities that require self-control (e.g., tasks involving memory, decision-making, and error correction). But are these poor regulators necessarily depleted, or might there be a way of turning poor regulators into efficient regulators? A recent study suggests one promising intervention. Kroese and colleagues (2011) showed that poor dietary regulators improved their resistance to food temptations when they formed an implementation intention (an if-then plan) linking those temptations to their dietary goals. Implementation intentions might, in a similar fashion, help to automatize poor regulators’ emotion control in the face of negative stimuli. A potential alternative explanation for the results we have reported is that rather than the good emotion regulators actually being more successful at emotion regulation in their everyday lives and so becoming less depleted because they had developed more efficient regulatory mechanisms and strategies, they simply believed that they were better at emotion control and this belief protected them from depletion. Certainly, the measure we used assessed perceptions of regulatory success rather than actual success, and although scores on the measure did correlate with an indicator of successful emotion control (feeling happy in everyday life), they did not predict performance differences in our experimental emotion regulation tasks. In line with this alternative explanation, evidence has begun to accumulate that beliefs about self-control may predict ego-depletion over and above actual expenditure of self-control resources (e.g., Clarkson, Hirt, Jia, & Alexander, 2010; Job, Dweck, & Walton, 2010). Although people sometimes engage in emotion regulation with the goal of worsening their feelings (e.g., when negative emotions are believed to be functional for the task at hand, Tamir et al., 2008), our results indicate that affect-worsening did not deplete blood glucose levels, irrespective of the caliber of participants’ emotion control. This finding is consistent with evidence that worsening emotions is easier than improving emotions (Webb et al., in press) and with studies where participants reported expending less effort when worsening their emotions compared to improving emotions (e.g., Ochsner et al., 2004). Kim and Hamann (2007) have suggested that positive emotions may be more malleable than negative emotions and so a fruitful avenue for future research may be to explore whether the actual or perceived malleability of affect also influences glucose expenditure. In conclusion, our study extends research on self-control by suggesting that the caliber of people’s self-control does not just predict the accessibility of goals and rates of goal attainment, but also the ability to conserve resources for future goal striving. In particular, our findings suggest that people who believe they are good self-regulators do not just enact self-control in a cognitively efficient manner (e.g., Fishbach et al., 2003); they also exert self-control in a more physiologically efficient manner compared with poor self-regulators. The present research also contributes to the strength model of self-control in showing that depletion is not an inevitable consequence of self-regulation; both person characteristics (perceived emotion control caliber) and the difficulty of the task (affect-improving vs. affect-worsening) influence the extent to which blood glucose is depleted by emotion regulation. References Bargh, J. A. (1990). Auto-motives: Preconscious determinants of social interaction. In E. T. Higgins & R. M. Sorrentino (Eds.), Handbook of motivation and cognition (Vol. 2, pp. 93-130). New York, NY: Guilford. 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, 1252-1265. Baumeister, R. F., Vohs, K. D., & Tice, D. M. (2007). The strength model of self-control. Current Directions in Psychological Science, 16, 396-403. Clarkson, J. J., Hirt, E. R., Jia, L., & Alexander, M. B. (2010). When perception is more than reality: The effect of perceived versus actual resource depletion on self-regulatory behavior. Journal of Personality and Social Psychology, 98, 29-46. Cousineau, D. (2005). Confidence intervals in within-subject designs: A simpler solution to Loftus and Masson’s method. Tutorial in Quantitative Methods for Psychology, 1, 42-45. Delaney, H. D., & Maxwell, S. E. (1981). On using analysis of covariance in repeated measures designs. Multivariate Behavioral Research, 16, 105-123. Dvorak, R. D., & Simons, J. S. (2009). Moderation of resource depletion in the self-control strength model: Differing effects of two modes of self-control. Personality and Social Psychology Bulletin, 35, 572-583. Fishbach, A., Friedman, R. S., & Kruglanski, A. W. (2003). Leading us not unto temptation: Momentary allurements elicit overriding goal activation. Journal of Personality and Social Psychology, 84, 296-309. Gailliot, M. T., & Baumeister, R. F. (2007). The physiology of willpower: Linking blood glucose to self-control. Personality and Social Psychology Review, 11, 303-327. Gailliot, M. T., Baumeister, R. F., DeWall, C. N., Maner, J. K., Plant, E. A., Tice, D. M., Brewer, L. E., & Schmeichel, B. J. (2007). Self-control relies on glucose as a limited energy source: Willpower is more than a metaphor. Journal of Personality and Social Psychology, 92, 325-336. Grandey, A. A. (2003). When “the show must go on”: Surface acting and deep acting as determinants of emotional exhaustion and peer-rated service delivery. Academy of Management Journal, 46, 86-96. 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, 496-525. Job, V., Dweck, C .S., & Walton, G. W. (2010). Ego depletion - Is it all in your head? Implicit theories about willpower affect self-regulation. Psychological Science, 21, 1686-1693. Kim, S. H., & Hamann, S. (2007). Neural correlates of positive and negative emotion regulation. Journal of Cognitive Neuroscience, 19, 776-798. Kroese, F. M., Adriaanse, M. A., Evers, C., & De Ridder, D. T. D. (2011). ‘Instant success’: Turning temptations into cues for goal-directed behavior. Personality and Social Psychology Bulletin, 37, 1389-1397. Matthews, G., Jones, D. M., & Chamberlain, A. G. (1990). Refining the measurement of mood: The UWIST Mood Adjective Checklist. British Journal of Psychology, 81, 17-42. Ochsner, K. N., Ray, R. D., Cooper, J. C., Robertson, E. R., Chopra, S., Gabrieli, J. D. E., & Gross, J. J. (2004). For better or for worse: Neural systems supporting the cognitive down- and up-regulation of negative emotion. Neuroimage, 23, 483-499. Richards, J. M., & Gross, J. J. (2000). Emotion regulation and memory: The cognitive costs of keeping one’s cool. Journal of Personality and Social Psychology, 79, 410-424. Schmeichel, B. J., & Baumeister, R. F. (2004). Self-regulatory strength. In R. F. Baumeister & K. D. Vohs (Eds.), Handbook of self-regulation (pp. 84-98). New York: Guilford Press. Tamir, M., Mitchell, C., & Gross, J. J. (2008). Hedonic and instrumental motives in anger regulation. Psychological Science, 19, 324-328. Trope, Y., & Fishbach, A. (2000). Counteractive self-control in overcoming temptation. Journal of Personality and Social Psychology, 79, 493-506. Varley, R., Webb, T. L., & Sheeran, P. (2011). Making self-help more helpful: A randomized controlled trial of the impact of augmenting self-help materials with implementation intentions on promoting the effective self-management of anxiety symptoms. Journal of Consulting and Clinical Psychology, 79, 123-128. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS Scales. Journal of Personality and Social Psychology, 47, 1063-1070. Webb, T. L., Miles, E., & Sheeran, P. (in press). Dealing with feeling: A meta-analysis of the effectiveness of strategies derived from the process model of emotion regulation. Psychological Bulletin. Williams, L. E., Bargh, J. A., Nocera, C. C., & Gray, J. R. (2009). The unconscious regulation of emotion: Nonconscious reappraisal goals modulate emotional reactivity. Emotion, 9, 847-854. Wills, T. A., Walker, C., Mendoza, D., & Ainette, M. G. (2006). Behavioral and emotional self-control: Relations to substance use in samples of middle and high school students. Psychology of Addictive Behaviors, 20, 265-278. Wood, W., & Neal, D. T. (2007). A new look at habits and the habit-goal interface. Psychological Review, 14, 843-863. Footnotes 1. A pilot experiment using emotion regulation scenarios in a between-subjects design (affect-improving N = 203; affect-worsening N = 196) confirmed that people perceive that affect-worsening is easier than affect-improving, verifying the validity of our comparison condition. In the study, participants reported that it would be easier to get into a sad mood from a happy mood (M = 13.85, SD = 5.07) than to get into a happy mood from a sad mood (M = 11.17, SD = 4.83), F(1, 396) = 29.27, p < .01, hp2 = .07. A full account of this study can be obtained from the first author. 2. Fifteen participants completed the affect-improving task first and sixteen completed the affect-worsening task first. There were no significant differences between these two groups in terms of their pre- or post-task moods and blood glucose levels, Fs < 3.34, ps > .05, hp2 < .11. 3. A pilot sample (N = 31) who were played the same collections of music but were instructed merely to listen to it verified that the music was not responsible for the changes that we observed in participants’ moods. For these participants, there was no difference in mood before (M = 5.33, SD = 0.73) versus after (M = 5.44, SD = 0.65) the music, F(1, 30) = 0.68, p = .42, hp2 = .02.  Table 1. Mood by Condition and Caliber of Emotion Control at Baseline and Follow-up All participants (N = 31) Poor emotion control (N = 15) Good emotion control (N = 16) Mean SD Mean SD Mean SD Affect-improving condition Baseline mood 5.02 0.83 4.91 0.79 5.13 0.82 Follow-up mood 5.57 0.72 5.43 0.65 5.70 0.78 Affect-worsening condition Baseline mood 5.13 0.71 4.94 0.67 5.32 0.73 Follow-up mood 3.82 0.89 3.80 0.94 3.83 0.87 Note. Higher scores represent more positive mood.  Figure Captions Figure 1. Three-way Interaction between Condition, Time of Measurement and Emotion Control Caliber on Blood Glucose Note. The panels show the interaction between condition and time of measurement for people with poor emotion control (left panel) and good emotion control (right panel) as defined by a median split of the variable. Error bars represent standard errors, calculated using Cousineau’s (2005) within-subjects method. motion Regulation and Glucose Depletion 19 motion Regulation and Glucose Depletion 18