Implicit Theories About Willpower Predict Self-Regulation and Grades in Everyday Life

Abstract

Laboratory research shows that when people believe that willpower is an abundant (rather than highly limited) resource they exhibit better self-control after demanding tasks. However, some have questioned whether this "nonlimited" theory leads to squandering of resources and worse outcomes in everyday life when demands on self-regulation are high. To examine this, we conducted a longitudinal study, assessing students' theories about willpower and tracking their self-regulation and academic performance. As hypothesized, a nonlimited theory predicted better self-regulation (better time management and less procrastination, unhealthy eating, and impulsive spending) for students who faced high self-regulatory demands. Moreover, among students taking a heavy course load, those with a nonlimited theory earned higher grades, which was mediated by less procrastination. These findings contradict the idea that a limited theory helps people allocate their resources more effectively; instead, it is people with the nonlimited theory who self-regulate well in the face of high demands. (PsycINFO Database Record
(c) 2015 APA, all rights reserved).

Full text

1
RUNNING HEAD: WILLPOWER-THEORIES PREDICT SELF-REGULATION AND
GRADES
Implicit Theories About Willpower Predict Self-Regulation and Grades in Everyday Life
Veronika Job1, Gregory M. Walton2, Katharina Bernecker1, Carol S. Dweck2
1 University of Zurich
2 Stanford University
Veronika Job (corresponding author)
Department of Psychology, University of Zurich, Binzmuehlestrasse 14/6, 8050 Zurich,
Switzerland
Phone: +41 (0)44 635 72 83 or +41(0)78 633 56 61, Fax: +41 (0) 44 635 75 19
Email: v.job@psychologie.uzh.ch
Gregory M. Walton
Department of Psychology, Stanford University, 450 Serra Mall, Stanford, CA 94305
Katharina Bernecker
Department of Psychology, University of Zurich, Binzmuehlestrasse 14/6, 8050 Zurich,
Switzerland
Carol S. Dweck
Department of Psychology, Stanford University, 450 Serra Mall, Stanford, CA 94305

2
Abstract
Laboratory research shows that when people believe that willpower is an abundant (rather than
highly limited) resource they exhibit better self-control after demanding tasks. However, some
have questioned whether this “nonlimited” theory leads to squandering of resources and worse
outcomes in everyday life when demands on self-regulation are high. To examine this, we
conducted a longitudinal study, assessing students’ theories about willpower and tracking their
self-regulation and academic performance. As hypothesized, a nonlimited theory predicted better
self-regulation (better time management and less procrastination, unhealthy eating, and
impulsive spending) for students who faced high self-regulatory demands. Moreover, among
students taking a heavy course load, those with a nonlimited theory earned higher grades, which
was mediated by less procrastination. These findings contradict the idea that a limited theory
helps people allocate their resources more effectively; instead, it is people with the nonlimited
theory who self-regulate well in the face of high demands.

3
Implicit Theories About Willpower Predict Self-Regulation and Grades in Everyday Life
Do people’s beliefs about the nature of self-control affect their ability to exert self-control
in everyday life settings? If so, what beliefs are most functional? One hypothesis is suggested by
the strength model of self-control (Baumeister, Bratslavsky, Muraven, & Tice, 1998; Baumeister,
Vohs, & Tice, 2007). This model proposes that self-control relies on a limited resource and that
understanding this limit helps people use this resource judiciously, improving self-regulation
especially when demands on self-control are high (Vohs, Baumeister, & Schmeichel, 2013). In
the present research, however, we propose the opposite, namely, that this limited-resource theory
undermines people’s self-regulatory efforts and, moreover, does so especially when demands on
self-control are high. The belief that willpower relies on a limited resource, we suggest, leads
people to act as though their self-regulatory resources are depleted long before they reach any
actual limit in their self-regulatory capacity. As a consequence, we expect that people with a
limited-resource theory will reduce their effort and engage in various overindulgent behaviors
when they face high demands on self-control. Instead, we propose that an alternative belief—the
belief that willpower is not easily used up and can even be fueled by the exertion of self-control
(a nonlimited theory)—promotes more successful self-regulation and performance when people
encounter challenging self-regulatory demands.
The Strength Model of Self-Control
Much contemporary research on self-regulation draws on the strength model of self-
control, which, as noted, posits that self-control relies on a limited energy resource (Baumeister
et al., 1998, 2007). According to this model, every act of self-control depletes this resource,
directly reducing the capacity to exert further self-control—a phenomenon termed ego depletion.
Empirical support for this model comes from numerous laboratory experiments, which show

4
that, after an initial task requiring self-control, people exhibit worse self-control on subsequent
tasks than do people who engaged in an initial undemanding task (for a meta-analysis, see
Hagger, Wood, Stiff, & Chatzisarantis, 2010).
Recent field studies extend these findings to everyday self-regulation, and appear to
suggest that the same principle applies. For instance, in an experience sampling study Hofmann,
Vohs, and Baumeister (2012) found that the more participants controlled themselves by resisting
desires the more likely they were to show self-control failures later in the day. Similarly,
research on stress and self-regulation confirms that when people contend with stressful
circumstances, such as daily hassles or academic examinations, they tend to engage in more
problematic, potentially harmful behaviors like eating unhealthy food, consuming alcohol,
smoking, and spending excessively (Ng & Jeffery, 2003; O’Connor, Jones, Conner, Mcmillan, &
Ferguson, 2008; Oaten & Cheng, 2005; Steptoe, Lipsey, & Wardle, 1998). It is thus well
documented that self-control often suffers when self-control demands are high, both in
laboratory and in everyday life settings.
Given the importance of self-regulation for successful goal-striving, health, and overall
functioning (e.g., Moffitt et al., 2011), identifying factors that predict better self-regulation,
especially in the face of high demands, is critical. Recent research has identified several variables
that moderate the ego-depletion effect. For instance, motivational variables like incentives,
expectations, and perceptions of a task can diminish or eliminate ego depletion in laboratory
settings (Clarkson, Hirt, Jia, & Alexander, 2010; Martijn, Tenbult, Merckelbach, Dreezens, & De
Vries, 2002; Muraven & Slessareva, 2003). Most pertinent to the present research, Job, Dweck,
and Walton (2010) found that people’s lay believes about willpower, so called implicit theories1,
determined whether people showed ego depletion at all.

5
Implicit Theories About Willpower
Challenging the strength model of self-control, Job and colleagues (2010) demonstrated
in a series of laboratory experiments that only people who believe that willpower is limited and
easily depleted (a limited theory of willpower, assessed with questions like “After a strenuous
mental activity, your energy is depleted and you must rest to get it refueled again”) show ego
depletion, that is, perform worse after an initial self-control task. People who reject the view that
willpower is highly constrained and who believe, instead, that willpower can even be self-
generating (e.g., “After a strenuous mental activity, you feel energized for further challenging
activities”) showed no impairment over a series of demanding self-control tasks (see also Miller
et al., 2012). We call the latter belief a nonlimited theory of willpower. We intentionally do not
use the term “unlimited.” People with a nonlimited theory may not believe that willpower is
limitless or that they are immune to depletion from highly strenuous tasks of long duration.
However, they reject the view that willpower is easily depleted by acts of self-control.
Job and colleagues found that implicit theories about willpower predict ego-depletion
both measuring theories about willpower as an individual-difference and manipulating them
experimentally, suggesting their causal role. These findings imply that self-regulatory failure
following the brief exertion of self-control results from people’s beliefs about their available
resources rather than from a true lack of resources (see also Job, Walton, Bernecker, & Dweck,
2013).
How do implicit theories about willpower affect ego depletion? Research suggests that
the belief that willpower is limited sensitizes people to cues that may signal the availability or
unavailability of mental resources. For example, finding an initial self-control task exhausting
predicted worse subsequent self-control performance for people with a limited theory but was

6
unpredictive for people with a nonlimited theory (Job et al., 2010, Study 3). In another series of
studies, ingesting glucose restored self-control for those with a limited theory but had no effect
on those with a nonlimited theory, who continued to perform well on self-control tasks whether
they had ingested glucose or not (Job et al., 2013). Previous research suggests that glucose
signals the restoration of self-control resources (Chambers, Bridge, & Jones, 2009; Molden et al.,
2012). Our results suggest that only people who believe that willpower is highly limited carefully
monitor for cues to the availability of self-control resources.
Extending this research, Vohs and colleagues (2013) replicated the effects of implicit
theories about willpower on ego depletion and raised an important question: Will the same
effects hold when self-control demands are especially high? Vohs and colleagues hypothesized
that implicit theories about willpower lead people to temporarily compensate for a lack of
resources. They suggest that people can do so effectively in the face of mild or moderate self-
control demands but not in the face of high demands, where “severe” depletion eventually takes
its toll. In a laboratory experiment, they examined how a manipulation of theories about
willpower interacted with three ego depletion conditions: a “no depletion” condition in which
participants completed no initial self-control tasks; a “mild depletion” condition in which
participants completed two initial self-control tasks; and a “severe depletion” condition in which
participants completed four initial self-control tasks. Vohs and colleagues replicated the finding
that a nonlimited theory of willpower improves self-control in the face of “mild depletion.” But
in the “severe depletion” condition, there was no positive effect of a nonlimited theory and on
one of two measures of self-control performance the effect even reversed: Participants led to
adopt a limited-resource theory performed better. Vohs and colleagues concluded that a
nonlimited theory can be counterproductive. Thinking that willpower is nonlimited, they write,

7
“might undermine the normal tendency to conserve resources (Muraven, Shmueli, & Burkley,
2006) so that people find themselves severely depleted after multiple tasks” (p.186).
Laboratory tasks, however, are not ideal for drawing conclusions about the limits of
willpower. There are many reasons people might display less effort after a long series of
demanding but potentially unmeaningful tasks. The critical test of how beliefs about willpower
affect self-regulation must be conducted in real-world settings in which people contend with
accumulating demands on their self-control as they strive to accomplish personal goals. Indeed,
in contrast to Vohs and colleagues’ conclusion, an earlier longitudinal study found that college
students who endorsed a nonlimited theory of willpower exhibited superior everyday self-
regulation during the week before final exams, when demands on self-control were assumed to
be high (Job et al., 2010, Study 4). They ate less unhealthy food, procrastinated less, and pursued
personal achievement goals more effectively than students with a limited theory.
The present research extends this prior study to provide a more detailed examination of
how implicit theories about willpower predict everyday self-regulation. The study does so in
several ways. First, the prior study simply assumed that self-regulatory demands were high for
all students at a particular time, that is, as final exams approached. In the present research, we
assessed the level of self-regulatory demands for each student on a week-by-week basis across an
academic term so we could identify the students who faced consistently high demands and those
who faced lower demands. We hypothesized that a nonlimited theory of willpower would predict
better self-regulatory outcomes among students who contended with high self-regulatory
demands but not necessarily among students who faced low self-regulatory demands. Second, we
assessed a broader range of self-regulatory outcomes than in past research, including not only
procrastination and unhealthy eating but also ineffective time management, impulsive spending,

8
and emotion-regulation failure. Finally, we examined a further important outcome that is
determined in part by self-regulation, and that is not self-reported: participants’ end-of-term
grade-point-average (GPA) (see Duckworth & Seligman, 2005). We examined whether a
nonlimited theory of willpower would predict a higher GPA that term (controlling for past GPA)
especially among students who took a heavy course load. We further expected that any
improvement in GPA would be mediated by better self-regulation, especially lower levels of
procrastination.
In summary, we tested the hypothesis that students holding a nonlimited theory of
willpower facing consistently high demands would display better self-regulation and
consequently would reach higher grades than would students holding a limited theory facing
similar circumstances. If this is the case, it would suggest that thinking of willpower as a
nonlimited resource, rather than harming people by leading them to waste their resources, helps
people stay focused on their goals when a heavy workload and accumulating demands make self-
regulation challenging.
Method
Participants and Procedure
Participants were 176 students at a selective university in the Western United States (101
women; Mage=21.21, SD=2.62). They were asked to complete an online questionnaire at five time
points, once each week during the second half of a 10-week term (T1-T5). Students received $3
for completing each questionnaire and a $10 bonus for completing all five questionnaires. A total
of 113 participants completed all five questionnaires; 26 completed four, 13 completed three, 10
completed two, and 14 completed one. Data from all participants who completed at least two
consecutive questionnaires were included in hierarchical linear modeling (HLM) analyses of

9
self-regulation (N=156). [Participants completed the relevant measures versus those who did not
did not differ in theories about willpower, t(174)=-.84, p=.40.] Each week on Monday morning
participants received a link to the online questionnaire with a request to respond that day. On
average, 65% of participants completed the questionnaire that day (ranging from 83% at T1 to
54% at T4). Non-responding participants were sent a reminder on Tuesday and could respond
until Wednesday night.
At the beginning of the first questionnaire participants provided informed consent and
were asked to authorize the release of their college academic records from the term of their
participation and the prior term. One hundred fifty-three participants (87%) authorized the
release. [Participants who authorized the release of their academic record did not differ in their
theories about willpower from those who did not, t(174)=0.23, p=.82.]
Measure of Implicit Theories About Willpower
At Time 1 participants completed a 6-item measure assessing theories about willpower
with respect to strenuous mental exertion (Job et al., 2010). Sample items include “After a
strenuous mental activity your energy is depleted and you must rest to get it refueled again”
(limited-resource theory) and “Your mental stamina fuels itself; even after strenuous mental
exertion you can continue doing more of it” (nonlimited-resource theory) (1=strongly agree,
6=strongly disagree; α=.85). Items referring to the limited resource-theory were reverse-scored
so that higher values represent greater agreement with the limited resource-theory (Mgrand=3.88,
SD=.88).
Measures of Everyday Self-Regulation Failures and Self-Regulatory Demands
Our questionnaire allowed us to determine which students faced high demands over a
several week period. Each questionnaire assessed, first, indices of everyday self-regulation

10
failures during the previous week and, second, self-regulatory demands anticipated in the current
week. This approach separates the assessment of the two critical variables for each week so as to
prevent reports of one from biasing reports of the other. Since we had no measure of self-
regulation failures during the last week of the study, we had complete information about
anticipated self-regulatory demands and self-regulation failures for four weeks.
Everyday self-regulation. Each questionnaire (T1-T5) assessed self-regulatory failures
during the prior week by asking participants to report the frequency of (1) procrastination,
defined as engaging in nonacademic activities rather than studying (e.g., “How often did you
meet friends instead of studying?”), (2) consumption of unhealthy (high fat/high sugar) foods
and drinks, like chocolate bars or salty snacks, (3) poor time management (e.g., “How often did
you come late to an appointment?”), (4) excess spending (e.g., “How often did you buy
something knowing that it’s actually too expensive for you?”), and (5) failure to control
emotions, (e.g. “How often did you have trouble controlling your temper?”) during the prior
week (1=never, 2=1 time per week, 3=2 times per week, 4=3-4 times per week, 5=5-6 times per
week, 6=1 time per day, 7=two or more times per day). Descriptive statistics and reliability
information are presented in Table 1.
To assess whether the five indicators of self-regulation failure converged as indicators of
a single latent variable, we conducted a confirmatory factor analysis with full information
maximum likelihood estimation on the five measures assessed at T1. A one-factor model fit the
data: χ2 (df=5, N=176)=3.03, p=.70, comparative fit index (CFI)=1.00, root mean square error of
approximation (RMSEA)=.00. All loadings were significant (standardized values:
procrastination=.67, consumption of unhealthy foods=.22, poor time management=.59, excess
spending=.46, emotion-regulation failure=.52). Therefore, in addition to examining each

11
measure separately, we also created a composite index of self-regulation failure by averaging
scores for the five variables at each time point.
Anticipated self-regulatory demands. We created a list of 13 demands undergraduate
students commonly face over an academic term. These included academic tasks (e.g., “class
presentations to deliver,” “tests to take”) and social stressors (e.g., “conflicts with one’s
professor or TA,” “experience of social exclusion or rejection”). For each demand, participants
were asked to “indicate how much you will have to deal with [this] task or experience during the
next seven days” (1=not at all, 2=a little, 3=somewhat, 4=very much). These ratings thus
allowed us to summarize diverse anticipated self-regulatory demands to create a single index for
each student.2
Academic performance and course load. Students’ college academic records provided
measures of (a) their grade-point-average (GPA) during the term the study was conducted and
the previous term and (b) their course load, i.e., the number of units students enrolled in each
term.
Trait Self-Control
If we find the hypothesized relationship between a nonlimited theory and fewer self-
regulatory failures, a potential alternative explanation involves trait self-control: Perhaps people
with a nonlimited theory about willpower show better self-regulatory outcomes simply because
they have greater self-control to begin with. To examine this possibility, we administered the
brief Trait Self-Control Scale (Tangney, Baumeister, & Boone, 2004) at the end of the T2
questionnaire. Participants indicated on a 5-point scale (1=not at all, 5=very much) how well
each of 13 statements about self-control described their typical behavior (e.g., “I’m good at
resisting temptation,” “I am lazy”) (α=.88).

12
Results
After reporting descriptive statistics and zero-order correlations, the results are divided
into two main sections. First, we analyzed everyday self-regulation with a multilevel approach.
Our primary focus was on between-participant differences in self-regulatory demands—whether
students who faced high demands during the term self-regulated better or more poorly as a
function of their implicit theory of willpower. We focused on this question because we expected
self-regulatory demands and behavior to vary more between- than within-participants over this
relatively short time period (i.e., five consecutive weeks within a term). However, we also
examined within-participant (i.e., week-to-week) changes in self-regulatory demands to
determine whether students showed differences in self-regulation, as a function of their theories
about willpower, on weeks they had previously predicted would pose high versus low demands.
Second, we examined students’ end-of-term GPA. A series of regression analyses tested
the hypothesis that theories about willpower would predict GPA, that this relationship would be
moderated by academic work-load, and that it would be mediated by procrastination.
Descriptive Statistics
Means and standard deviations of self-regulatory behaviors and self-regulatory demands
for each week are presented in Table 1. Over the course of the five measurement times, levels of
procrastination and time-management failure dropped: linear within-participant contrasts F(1,
93)=16.19, p=.001 and F(1, 93)=25.32, p<.001, respectively. As the end of the term and final
exams approached, students procrastinated less and managed their time better.
Table 2 reports the means, standard deviations, and zero-order correlations of theories
about willpower, course load in the current term, GPA in the current and previous term, as well
as trait self-control. There were no significant zero-order correlations among willpower theories,

13
course load, and GPA. As expected current- and prior-term GPA were highly correlated and
there was no mean difference between the two, t<1. There was also a significant correlation
between theories about willpower and trait self-control. Participants low in trait self-control
agreed more with a limited resource theory. As will be seen, however, this difference did not
account for the effect of theories about willpower on self-regulation and performance.
Everyday Self-Regulation Failure and Self-Regulatory Demands
Our data on students’ weekly self-regulation failure and self-regulatory demands conform
to a two-level hierarchical structure (repeated measures nested within individuals). Therefore, we
used hierarchical linear modeling (HLM 6.2 statistical software package, Raudenbush, Bryk,
Cheong, Congdon, & du Toit, 2004) to analyze between- and within-person differences in self-
regulation as our dependent variable. We first estimated an unconditional model with no
predictors at either level of the hierarchy, to see how much variation in weekly self-regulation
lay between- versus within-participants. This analysis revealed that, as might be expected, most
of the variance in self-regulation was between-participants (81%) (τ00=0.35); only 19% was
within-participants (σ2=0.08). Similarly, an unconditional model predicting forecasted demands
showed that, more of the variance was between-participants (62%) than within-participants
(38%) (σ2=0.12, τ00=0.08).
To estimate effects of self-regulatory demands at both the between-participants and
within-participants levels of analysis we ran a compositional model including the aggregated
score for mean demands across weeks as a participant-level predictor, as well as weekly scores
of demands (group-centered) as the week-level predictor (Enders & Tofighi, 2007; Raudenbush
& Bryk, 2002). This is the full model:
Self-Regulationij=B0j + B1jDemands + Rij

14
B0j=G00 + G01WillpowerTheory + G02MeanDemands + G03Theory×Demands + u0j
B1j=G10 + G11WillpowerTheory + u1j
G00 is the intercept. G01 and G02 represent the main effects of willpower theories and mean
demands on participants’ mean self-regulation failure reports. G03 represents the interaction
between willpower theory and demands, which tests our hypothesis at the between-participant
level. G10 represents the main within-participant effect of week-to-week changes in predicted
self-regulatory demands on week-to-week self-regulation reports. Finally, G11 represents the
cross-level interaction between willpower theory and self-regulatory demands, which tests
effects at the within-participant level.
Table 3 contains the coefficients for this model predicting the composite self-regulation
failure index as well as each measure of self-regulation failure. There was a highly significant
main effect of mean demands (G02): High demands were associated with more self-regulation
failure. The main effect of willpower theories (G01) (irrespective of self-regulatory demands) was
significant for procrastination: The more students endorsed a limited-resource theory the more
they procastinated. Most important, the interaction between willpower theories and mean
demands (G03) was significant for both the composite index and each measure of self-regulation
failure except poor time management. These results support our hypothesis: A limited theory
about willpower predicted more self-regulation failure among students who experienced greater
self-regulatory demands than their peers.
Figure 1 depicts predicted values of the composite index for students with a limited (+1
SD) and a nonlimited resource-theory (-1 SD) who had to deal with high (+1 SD) or low (-1 SD)
mean self-regulatory demands. Among students who generally faced high demands, those with a

15
limited resource-theory reported significantly more self-regulatory failures than those with a
nonlimited theory, b=0.27, se=0.07, t(152)=3.98, p<.001.
Among students who faced lower demands, self-regulatory failures were far less frequent.
Interestingly, however, among these students those with a nonlimited theory showed marginally
worse self-regulation than those with a limited resource-theory, b=-0.12, se=0.07, t(152)=-1.81,
p=.072. We return to this interesting point later.
Next, we tested the key contrast for each index of self-regulation failure. Among students
who faced high demands, those with a limited resource-theory, relative to those with a
nonlimited theory, procrastinated more (b=0.61, se=0.14, t(152)=4.20, p<.001), consumed more
unhealthy foods (b=0.24, se=0.12, t(152)=1.99, p=.048), managed their time marginally more
poorly (b=0.16, se=0.09, t(152)=1.91, p=.058), and spent more excessively (b=0.16, se=0.08,
t(152)=2.06, p=.041). For emotion regulation there was a weak trend in the same direction
(b=0.14, se=0.10, t(152)=1.39, p=.17). To illustrate the interactions, Figure 2 depicts
procrastination and unhealthy eating behavior for participants with a limited or nonlimited theory
who faced high versus low demands.
For students who faced low demands, willpower theories were not significantly related to
procrastination (b=-0.15, se=0.14, t(152)=-1.06, p=.29), unhealthy eating (b=-0.16, se=0.12,
t(152)=-1.35, p=.18), bad time management (b=0.03, se=0.08, t(152)=0.38, p=.70), or excess
spending (b=-0.11, se=0.07, t(152)=-1.52, p=.13). Students with a nonlimited theory, however,
reported marginally worse emotion regulation when demands were low (b=-0.18, se=0.10,
t(152)=-1.86, p=.064).
We have suggested that a limited-resource theory undermines self-regulation when self-
regulatory demands are high. A potential alternative explanation is that students with a limited

16
theory anticipate and have more self-regulatory demands than students with a nonlimited theory.
This was not the case. We ran an HLM analyses on students’ anticipated self-regulatory demands
with theories about willpower as a participant-level predictor. This analysis showed no
relationship between implicit theories and anticipated demands, b=-0.01, se=0.02, t(552)=-0.40,
p=.69. Further, the correlation between theories about willpower and mean self-regulatory
demands was not significant, (r=-.026, p=.75). Students with a limited and a nonlimited theory
anticipated similar self-regulatory demands. But only students with a limited theory responded to
high demands with more self-regulation failures.
A second potential alternative explanation, as noted earlier, involves trait self-control:
Perhaps students who endorse a limited willpower theory self-regulate less effectively in the face
of self-regulatory challenges simply because they have less trait self-control. This was not the
case. We ran an HLM-model predicting the composite index of self-regulation failure from
theories about willpower, mean demands, and their interaction, as well as trait self-control and its
interaction with mean demands as participant-level predictors. There was a significant main
effect for trait self-control, b=-0.39, se=0.05, t(146)=-7.92, p<.001: Students lower in trait self-
control reported more self-regulatory failures. The interaction between trait self-control and
demands was also significant, b=-0.07, se=0.03 t(146)=-2.20, p=.030: When demands were high,
students low in trait self-control reported the most self-regulatory failures. However, importantly,
in this analysis the interaction between theories about willpower and mean demands remained
significant, b=0.09, se=0.04, t(146)=2.43, p=.017. The effect of theories of willpower is not
accounted for by differences in trait self-control.
In contrast to these between-participant results, within-participant results were not
significant (see Table 3). First, within-participant (week-to-week) changes in demands (G10)

17
predicted only procrastination and time management failure. They were not related to the
composite index of self-regulation failure. Second, there was no cross-level interaction between
willpower theories and within-participant changes in demands (G11). That is, there was no
conjoint effect of participants’ week-to week changes in demands, corrected for their mean level
of demands, and theories about willpower on weekly reports of self-regulation. We suspect that a
longitudinal study with more intervals over a longer period of time might be able to better detect
conjoint effects of willpower theories and within-person changes in demands on participants’
self-regulation. Another reason for the lack of within–person effects could be that the cumulative
estimate of mean demands provides a more reliable measure of the demands a student faces than
week-by-week predictions. For example, on the Monday of one week a student might predict low
demands for the upcoming week, but then might accomplish less over the next few days and end
up having high demands the rest of the week. Or a student might predict that the week will be a
high-demand week but then an instructor grants the class an extension on an assignment and the
week becomes lower in demands. If so, the anticipation of demands on a week-by week basis
may be less accurate than cumulative estimates of demands over several weeks.
Academic Performance and Course Load
The self-report measures provide nuanced insight into how willpower theories predict
everyday self-regulation in the face of self-regulatory demands. Next, we tested whether
willpower theories also predict an objective (non-self-reported) and cumulative index of
successful self-regulation over time: students’ end-of-term GPA calculated from official school
records. If this shows the same pattern, it would extend the self-reported indices of self-
regulation and further illustrate the implications of theories about willpower for students’ lives.

18
Willpower theories and academic performance. To test whether willpower theories
predicted students’ end-of-term GPA, we conducted a hierarchical regression analyses. First, we
controlled for standardized GPA in the prior term, R2=.47, F(1,146)=128.01, p<.001. Second, we
added standardized willpower theories, which was significant,
β
=-.12, b=-.06, se=.03, ΔR2=.02,
F(1,145)=4.18, p=.04. Thus, even controlling for prior GPA, the more students endorsed a
limited resource-theory the lower was their end-of-term GPA.
Next, we tested whether this was especially the case among students taking a heavy
course load. We added course load and the willpower theories by course load interaction
(willpower theories and course load were independent, r=-.06, ns). In the final model (
Δ
R2=.04,
F(1,143)=10.74, p=.001), the main effect of willpower theories was marginally significant,
β
=-
.11, b=-.05, se=.03, t(143)=-1.80, p=.08. There was no main effect of course load,
β
=-.04, b=-
.02, se=.03, t<1. However, the willpower theories by course load interaction was significant,
β
=-
.19, b=-.09, se=.03, t(143)=-3.28, p=.001. As shown in Figure 3, willpower theories did not
predict GPA among students taking a light course load (1 SD below the mean, or 10.96 units out
of a possible 20),
β
=.12, b=.04, se=.12, t(144)=1.01, p=.31. But among students taking a heavy
course load (1 SD above the mean, or 18.14 units), those with a limited resource-theory earned
lower GPAs (Mest=3.41) than those with a nonlimited theory about willpower (Mest=3.69),
β
=-
.41, b=-.14, se=.11, t(144)=-3.77, p<.001. In addition, a heavy course load was associated with
worse performance only for students with a limited-resource theory,
β
=-.32, b=-.11, se=.12,
t(144)=-2.81, p=.006. Students with a nonlimited theory actually performed slightly better when
they had a heavy course load,
β
=.21, b=.07, se=.11, t(144)=1.81, p=.07.
We also tested whether a limited theory predicted worse grades on a longitudinal basis--
that is, when students’ course load increased from one term to the next. We examined students’

19
current-term GPA, with prior-term GPA controlled, and tested the effects of willpower theories,
change in course load (difference between current term course load and previous term course
load, Mchange =-0.76, SD=3.26) from the prior term to the current term, and the willpower theories
by change-in-course-load interaction. The interaction was significant,
β
=-.13, b=-.06, se=.03,
t(143)=-2.17, p=.03. Among students whose course load decreased (Mchange - 1 SD = -0.76 – 3.26
= decrease of 4.02 units), there was no effect of willpower theories on grades, t<1. But among
students whose course load increased (Mchange + 1 SD = -0.76 + 3.26 = increase of 2.5 units), those
with a limited theory earned worse grades than those with a nonlimited theory,
β
=-.35, b=-.12,
se=.12, t(144)=-2.95, p=.004. These findings confirm that GPA varies with students’ implicit
theories about willpower and their changing course load.
As was the case for everyday self-regulation, the effect of willpower theories on GPA
was not accounted for by trait self-control. We conducted a hierarchical regression predicting
GPA from previous-term GPA (block 1), theories about willpower, course load and trait self-
control (block 2), as well as the interactions between course load and trait self-control and
between course load and willpower theories (block 3). In the final model (ΔR2=.54,
F(6,124)=24.04, p<.01) the main effects for trait self-control and willpower theories were both
nonsignificant,
β
=.10, b=.04, se=.03, t(124)=1.49, p=.14, and,
β
=-.08, b=-.04, se=.03, t(124)=-
1.31, p=.19, respectively. The course load by trait self-control interaction was also not
significant,
β
=-.07, b=-.03, se=.03, t(124)=-1.08, p=.28. However, the course load by willpower
theory interaction remained significant,
β
=-.23, b=-.10, se=.03, t(124)=-3.71, p<.01. In the face
of a heavy course load, willpower theories predict GPA above and beyond trait self-control.
Mediation. Finally, we tested whether greater procrastination explained the relationship
between a limited willpower theory and GPA. Past research shows that procrastination leads to

20
lower grades (Steel, 2007). Likewise, in the present research mean procrastination (averaged
over five weeks) predicted GPA controlling for previous-term GPA (partial r(145)=-.32,
p<.001). Since the previous analysis showed that theories about willpower affect students’
procrastination, we tested whether students’ mean procrastination mediated the effect of
willpower theories on GPA (see Figure 4). We did so using the INDIRECT macro (Preacher &
Hayes, 2008), which uses bootstrapping to estimate the indirect effect of an independent variable
(i.e., willpower theories) on a dependent variable (i.e., GPA) through a mediator (i.e.,
procrastination). Because the macro provides only unstandardized path coefficients all variables
were z-standardized prior to using the macro to generate standardized coefficients. The
mediational model was significant, R2=.53, F(3,144)=52.71, p<.001. As reported earlier, there
was a main effect for willpower theories predicting procrastination; a limited-theory predicted
greater procrastination,
β
=.18, se=.07, t(144)=2.60, p=.01. The direct effect of procrastination on
GPA was also significant,
β
=-.11, se =.03, t(144)=-3.75, p<.001. The more students
procrastinated the lower was their end-of-term GPA. The bootstrapped indirect effect was
different from zero, 95% CI [-.005, -.047], and the direct effect of willpower theories on GPA
was no longer significant,
β
=-.03, se=.03, t(144)=-1.30, p=.20. The other indicators of self-
regulatory failure did not show this meditational pattern. The results suggest that the more
students endorsed a limited-resource theory of willpower the more they procrastinated and the
lower the GPA they earned.
General Discussion
The present research shows that students who think that willpower is limited and easily
depleted—the view of willpower suggested by the strength model of self-control—self-regulate
less well in their everyday lives when they face high self-regulatory demands. Far from

21
conserving their resources and showing strong self-regulation when needed, students who
endorsed the limited theory and who faced high demands over the term, procrastinated more
(e.g., watching TV instead of studying), ate more junk food, and reported more excessive
spending as compared to students with a nonlimited theory about willpower. This was the case
even though students with a limited and a nonlimited theory faced similar self-regulatory
demands. By measuring students’ self-regulatory demands, the present study provides the first
direct evidence that it is precisely in the face of consistently high demands that a nonlimited
theory of willpower predicts better everyday self-regulation.
Importantly, we found the same pattern for students’ term grades, an objective and
inherently important variable resulting from successful self-regulation. Among students who
took a heavy course load, students with the limited theory earned lower grades than students with
the nonlimited-resource theory. They did so, a mediation analysis suggested, because they were
more likely to procrastinate in completing their work. By contrast, the nonlimited-resource
theory led people to deploy their resources more effectively when they were needed most.
Notably, the effects of willpower theories on everyday self-regulation and on GPA did not arise
because students with a limited theory had lower trait self-control. The patterns remained
significant even controlling for trait self-control.
Our findings contradict the hypothesis that a nonlimited theory about willpower
undermines self-regulation and does so especially when demands are high. Relying on a
laboratory experiment, Vohs and colleagues (2013) suggested that the belief that willpower is
nonlimited might counteract ego depletion only in cases of mild or moderate demands but not
when self-control demands are “severe.” They speculated that this belief could even amplify ego
depletion and worsen self-regulation by undermining people’s “normal tendency” to conserve

22
their resources and then deploy them when they were needed. To the contrary, examining
students’ success and failure self-regulating in a highly demanding academic environment, we
found that the nonlimited theory was most predictive of better outcomes among students who
faced the greatest demands on their self-regulation.
In fact, it was only among students who faced low demands—when self-regulation lapses
may be less costly--that students with a nonlimited theory “wasted” their self-regulatory
resources relative to those with a limited theory. Under these conditions, they reported giving in
to impulses and pursuing non-academic activities at least as much as, if not more than, students
with a limited resource-theory. Thus, students with a nonlimited theory are not self-control super
heroes who never give in to temptations; nor are they unwilling to admit self-regulatory failure.
This view is further supported by the rather low correlation between theories about willpower
and trait self-control (-.17), suggesting that those with a nonlimited theory are not simply natural
self-regulators or people with outsized self-control abilities. Rather, those with a nonlimited
theory are people who lean in when demands on self-regulation are high.
This pattern was replicated in the analysis of students’ academic performance, where a
nonlimited vs. limited willpower theory predicted higher GPA among students who took heavy
course loads. It was also interesting to find that participants with a nonlimited theory tended to
earn even higher grades when they were dealing with a heavy course load than when they were
taking a light course load. These latter findings suggest that people with a nonlimited theory may
even profit from challenging circumstances. Indeed it is possible that in situations where they are
not sufficiently challenged (e.g., in a boring job), people with a nonlimited theory might be the
ones to show lower performance. For them, boredom or lack of challenge may be depleting! In

23
short, people with a nonlimited theory about willpower look strong when high demands require
effective self-regulation but do not perform better when demands are low.
Why did Vohs and colleagues (2013) find a different pattern in a laboratory study—that
the benefits of a nonlimited theory for self-control performance disappeared as the number of
self-control tasks increased? As a laboratory session wears on, many other factors beyond
participants’ self-control capacity may affect their willingness to exert further self-control on
laboratory tasks. For instance, participants may simply decide that they have done enough and/or
that the tasks are no longer interesting or consequential. A nonlimited theory about willpower
would not be functional if it led people to engage on a high level with every task that came along
regardless of its value or purpose. Future laboratory research may decompose the capacity to
exert self-control from the value or meaning of a task to the self.
In Vohs’ and colleagues (2013) research as well as in other recent theorizing on ego
depletion (Inzlicht & Schmeichel, 2012; Inzlicht, Schmeichel, & Macrea, 2014) the effects of
willpower theories are grouped together with those of other “motivational” factors such as
monetary incentives. It is often assumed that both counteract ego depletion through enhanced
motivation. Research on theories about willpower, however, proposes a different perspective. A
nonlimited theory does not just motivate people to regulate themselves better; instead, it removes
a process that undermines self-regulation. Our previous research suggests that a limited-resource
theory makes people more sensitive to or more vigilant for cues that signal the availability of
mental resources, like perceived exhaustion or ingested glucose (Job et al., 2010, 2013). From
this perspective people with a limited-resource theory perform worse under high demands
because as soon as they experience even low-level signs of strain or exhaustion (as soon as they
perceive any “depletion”) they begin to reduce effort on the task at hand. Instead of staying

24
focused on a demanding task they turn toward saving and/or replenishing their presumably
limited resources. Thus, willpower theories are not simply another variable that changes the
incentive value of a task. Rather, by affecting the fundamental assumptions people make about
the nature and workings of willpower, they can change how people approach and enact self-
regulation itself.
The present results suggest that a nonlimited theory of willpower is functional in a
student sample facing high demands and likely in other populations facing self-regulation
challenges. But if so why do many people believe that willpower is limited? Vohs and colleagues
(2013) argued that if a nonlimited theory were beneficial, these benefits would have led the
theory to spread across individuals and cultures. But a belief need not be functional to spread. It
just has to be simple and appealing (Bangerter & Heath, 2004; Dawkins, 2006; Wagner,
Kronberger, & Seifert, 2002). A fixed mindset about intelligence (the belief that intelligence is
fixed not malleable) is a simple and widespread idea that can have clear negative effects, for
instance in undermining students’ resilience and academic achievement (Blackwell,
Trzesniewski, & Dweck, 2007; Dweck & Leggett, 1988; Hong, Chiu, Dweck, Lin, & Wan,
1999). Further, both a fixed mindset about intelligence and a limited-resource theory about
willpower can have psychological benefits; for instance, as a justification for putting forth low
effort in the face of challenging tasks or temptations (see, e.g., Job et al., 2010; Robins & Pals,
2002).
Although a limited-resource theory might serve some psychological functions, the present
research documents its costs. The ability to self-regulate successfully is one of the most robust
predictors of major life outcomes, including health, wealth, and well-being (Moffitt, et al., 2011).
A critical question for future research involves better understanding the causal effects of

25
willpower beliefs in everyday settings and, if causal, how to change these beliefs to increase self-
regulatory success. Laboratory studies show that implicit theories about willpower can be
manipulated and that their effects when manipulated parallel their effects when measured (Job et
al., 2010, 2013; Miller et al., 2012). Thus it seems probable that willpower beliefs have causal
effects in everyday life settings and, in these settings too, they may be changed. Nevertheless,
randomized field-experiments that manipulate willpower beliefs and examine everyday self-
regulatory outcomes are necessary to establish causality. Such field experiments would also test
a novel means to improve people’s self-regulatory outcomes, a pressing issue (Diamond, 2012;
Duckworth, Grant, Loew, Oettingen, & Gollwitzer, 2011). Previous field-experiments show that
it is possible to change people’s implicit theories about intelligence and personality in field
settings, with beneficial consequences including for academic performance (Aronson, Fried, &
Good, 2002; Blackwell et al., 2007) and social outcomes (Yeager, Trzesniewski, & Dweck,
2013). Such interventions give people information (e.g., scientific reports) about the nature of
human qualities and help them internalize this information using powerful persuasive techniques
(e.g., “saying-is-believing” exercises, Aronson, 1999; see also Yeager & Walton, 2011). Could
this approach change people’s beliefs about the nature of willpower in a relatively enduring way?
Importantly, it may be essential to pair such learning opportunities with information
about effective strategies that can help people avoid self-regulatory failures. Although the
present research did not examine self-regulation strategies, these strategies may be an important
consideration in the development of interventions to promote a non-limited theory of willpower.
Ironically, simply learning that willpower is stronger than one might have supposed could
backfire if this encourages people to put themselves in situations they are ill-equipped to deal
with (e.g., keeping temptations close at hand in the belief that they will be able to resist them

26
indefinitely). Effective self-regulation strategies may involve formulating plans to cope with
temptations (e.g., implementation intentions) or structuring one’s environment to avoid
temptations (e.g., putting junk food in a high cabinet, blocking Facebook while trying to study)
(Duckworth et al., 2011; Magen & Gross, 2007; Neal, Wood, & Drolet, 2013; Stadler, Oettingen,
& Gollwitzer, 2010; Webb & Sheeran, 2003). Both for broad populations in demanding
environments (e.g., students) and for clinical populations (e.g., diabetics), it would be exciting if
exercises to teach people a nonlimited-resource theory plus effective self-regulatory strategies
could increase their success as they face stressful demands and strive to accomplish their goals.

27
References
Aronson, E. (1999). The power of self-persuasion. American Psychologist, 54(11), 584–588.
doi:10.1037/h0088188
Aronson, J., Fried, C. B., & Good, C. (2002). Reducing the effects of stereotype threat on
African American college students by shaping theories of intelligence. Journal of
Experimental Social Psychology, 38(2), 113–125. doi:10.1006/jesp.2001.1491
Bangerter, A., & Heath, C. (2004). The Mozart effect: Tracking the evolution of a scientific
legend. The British Journal of Social Psychology, 43(4), 605–23.
doi:10.1348/0144666042565353
Baumeister, R. E., Bratslavsky, E., Muraven, M., & Tice, D. M. (1998). Ego depletion: Is the
active self a limited resource? Journal of Personality and Social Psychology, 74(5), 1252–
1265. doi:10.1037/0022-3514.74.5.1252
Baumeister, R. F., Vohs, K. D., & Tice, D. M. (2007). The strength model of self-control.
Current Directions in Psychological Science, 16(6), 351–355. doi:10.1111/j.1467-
8721.2007.00534.x
Blackwell, L. S., Trzesniewski, K. H., & Dweck, C. S. (2007). Implicit theories of intelligence
predict achievement across an adolescent transition: A longitudinal study and an
intervention. Child Development, 78(1), 246–63. doi:10.1111/j.1467-8624.2007.00995.x
Chambers, E. S., Bridge, M. W., & Jones, D. A. (2009). Carbohydrate sensing in the human
mouth: effects on exercise performance and brain activity. The Journal of Physiology,
587(8), 1779-1794. doi: 10.1113/jphysiol.2008.164285
Clarkson, J. J., Hirt, E. R., Jia, L., & Alexander, M. B. (2010). When perception is more than
reality: The effects of perceived versus actual resource depletion on self-regulatory

28
behavior. Journal of Personality and Social Psychology, 98(1), 29–46.
doi:10.1037/a0017539
Dawkins, R. (2006). The Selfish Gene. New York: Oxford University Press.
Diamond, A. (2012). Activities and programs that improve children’s executive functions.
Current Directions in Psychological Science, 21(5), 335–341.
doi:10.1177/0963721412453722
Duckworth, A. L., Grant, H., Loew, B., Oettingen, G., & Gollwitzer, P. M. (2011). Self-
regulation strategies improve self-discipline in adolescents: Benefits of mental contrasting
and implementation intentions. Educational Psychology, 31(1), 17–26.
doi:10.1080/01443410.2010.506003
Duckworth, A. L. & Seligman, M. E. P. (2005). Self-discipline outdoes IQ in predicting
academic performance of adolescents. Psychological Science, 16, 939-944.
doi:10.1111/j.1467-9280.2005.01641.x
Dweck, C. S., & Leggett, E. L. (1988). A social-cognitive approach to motivation and
personality. Psychological Review, 95(2), 256–273. doi:10.1037//0033-295X.95.2.256
Enders, C. K., & Tofighi, D. (2007). Centering predictor variables in cross-sectional multilevel
models: a new look at an old issue. Psychological Methods, 12(2), 121–38.
doi:10.1037/1082-989X.12.2.121
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(4), 495–525.
doi:10.1037/a0019486

29
Hofmann, W., Vohs, K. D., & Baumeister, R. F. (2012). What people desire, feel conflicted
about, and try to resist in everyday life. Psychological Science, 23(6), 582–588.
doi:10.1177/0956797612437426
Hong, Y., Chiu, C., Dweck, C. S., Lin, D. M.-S., & Wan, W. (1999). Implicit theories,
attributions, and coping: A meaning system approach. Journal of Personality and Social
Psychology, 77(3), 588–599. doi:10.1037//0022-3514.77.3.588
Inzlicht, M., & Schmeichel, B. J. (2012). What is ego depletion? Toward a mechanistic revision
of the resource model of self-control. Perspectives on Psychological Science, 7, 450-463.
doi: 10.1177/1745691612454134
Inzlicht, M., Schmeichel, B. J., & Macrae, C. N. (2014). Why self-control seems (but may not
be) limited. Trends in Cognitive Sciences, 18, 127-133. doi: 10.1016/j.tics.2013.12009
Job, V., Dweck, C. S., & Walton, G. M. (2010). Ego depletion—is it all in your head? Implicit
theories about willpower affect self-regulation. Psychological Science, 21(11), 1686–93.
doi:10.1177/0956797610384745
Job, V., Walton, G. M., Bernecker, K., & Dweck, C. S. (2013). Beliefs about willpower
determine the impact of glucose on self-control. Proceedings of the National Academy of
Sciences, 110(37), 14837–14842. doi:10.1073/pnas.1313475110
Magen, E., & Gross, J. J. (2007). Harnessing the need for immediate gratification: cognitive
reconstrual modulates the reward value of temptations. Emotion, 7(2), 415–28.
doi:10.1037/1528-3542.7.2.415
Martijn, C., Tenbult, P., Merckelbach, H., Dreezens, E., & De Vries, N. K. (2002). Getting a grip
on ourselves: Challenging expectancies about loss of energy after self-control. Social
Cognition, 20(6), 441–460. doi:10.1521/soco.20.6.441.22978

30
Miller, E. M., Walton, G. M., Dweck, C. S., Job, V., Trzesniewski, K. H., & McClure, S. M.
(2012). Theories of willpower affect sustained learning. PLoS ONE, 7(6), e38680.
doi:10.1371/journal.pone.0038680
Molden, D. C., Hui, C. M., Scholer, A. A., Meier, B. P., Noreen, E. E., D’ Agostino, P. R., &
Martin, V. (2012). Motivational versus metabolic effects of carbohydrates on self-control.
Psychological Science, 23(10), 1137-1144. doi: 10.1177/0956797612439069
Moffitt, T. E., Arseneault, L., Belsky, D., Dickson, N., Hancox, R. J., & Harrington, H. (2011).
A gradient of childhood self-control predicts health, wealth, and public safety. Proceedings
of the National Academy of Sciences, 108(7), 2693–2698. doi:10.1073/pnas.1010076108
Muraven, M., Shmueli, D., & Burkley, E. (2006). Conserving self-control strength. Journal of
Personality and Social Psychology, 91(3), 524–537. doi:10.1037/0022-3514.91.3.524
Muraven, M., & Slessareva, E. (2003). Mechanisms of Self-Control Failure: Motivation and
Limited Resources. Personality and Social Psychology Bulletin, 29(7), 894–906.
doi:10.1177/0146167203253209
Neal, D. T., Wood, W., & Drolet, A. (2013). How do people adhere to goals when willpower is
low? The profits (and pitfalls) of strong habits. Journal of Personality and Social
Psychology, 104(6), 959–75. doi:10.1037/a0032626
Ng, D. M., & Jeffery, R. W. (2003). Relationships between perceived stress and health behaviors
in a sample of working adults. Health Psychology, 22(6), 638–642. doi:10.1037/0278-
6133.22.6.638
O’Connor, D. B., Jones, F., Conner, M., Mcmillan, B., & Ferguson, E. (2008). Effects of daily
hassles and eating style on eating behavior. Health Psychology, 27(1), 20–31.
doi:10.1037/0278-6133.27.1.S20

31
Oaten, M., & Cheng, K. (2005). Academic examination stress impairs self-control. Journal of
Social and Clinical Psychology, 24(2), 254–279. doi:10.1521/jscp.24.2.254.62276
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behavior Research Methods,
40(3), 879–891. doi:10.3758/BRM.40.3.879
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data
analysis methods (2nd ed.). Thousand Oaks, CA: Sage Publications.
Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., Congdon, R. T., & du Toit, M. (2004). HLM 6:
Hierarchical linear and nonlinear modeling. Chicago, IL: Scientific Software
International.
Robins, R. W., & Pals, J. L. (2002). Implicit self-theories in the academic domain: Implications
for goal orientation, attributions, affect, and self-esteem change. Self and Identity, 1(4),
313–336. doi:10.1080/15298860290106805
Stadler, G., Oettingen, G., & Gollwitzer, P. M. (2010). Intervention effects of information and
self-regulation on eating fruits and vegetables over two years. Health Psychology, 29(3),
274–83. doi:10.1037/a0018644
Steel, P. (2007). The nature of procrastination: a meta-analytic and theoretical review of
quintessential self-regulatory failure. Psychological Bulletin, 133(1), 65–94.
doi:10.1037/0033-2909.133.1.65
Steptoe, A., Lipsey, Z., & Wardle, J. (1998). Stress, hassles and variations in alcohol
consumption, food choice and physical exercise: A diary study. British Journal of Health
Psychology, 3(1), 51–63. doi:10.1111/j.2044-8287.1998.tb00555.x

32
Tangney, J. P., Baumeister, R. F., & Boone, A. L. (2004). High self-control predicts good
adjustment, less pathology, better grades, and interpersonal success. Journal of
Personality, 72, 271–324. doi: 10.1111/j.0022-3506.2004.00263.x
Tice, D. M., & Baumeister, R. F. (1997). Longitudinal study of procrastination, performance,
stress, and health: The costs and benefits of dawdling. Psychological Science, 8(6), 454–
458. doi:10.1111/j.1467-9280.1997.tb00460.x
Vohs, K. D., Baumeister, R. F., & Schmeichel, B. J. (2013). Motivation, personal beliefs, and
limited resources all contribute to self-control. Journal of Experimental Social Psychology,
48(4), 943–947. doi:10.1016/j.jesp.2012.08.007
Wagner, W., Kronberger, N., & Seifert, F. (2002). Collective symbolic coping with new
technology: Knowledge, images and public discourse. The British Journal of Social
Psychology, 41(3), 323–43. doi:10.1348/014466602760344241
Webb, T. L., & Sheeran, P. (2003). Can implementation intentions help to overcome ego-
depletion? Journal of Experimental Social Psychology, 39(3), 279–286.
doi:10.1016/S0022-1031(02)00527-9
Yeager, D. S, Trzesniewski, K. H., & Dweck, C. S. (2013). An implicit theories of personality
intervention reduces adolescent aggression in response to victimization and exclusion.
Child Development, 84(3), 970–88. doi:10.1111/cdev.12003
Yeager, D. S., & Walton, G. M. (2011). Social-psychological interventions in dducation: They’re
not magic. Review of Educational Research, 81(2), 267–301.
doi:10.3102/0034654311405999

33
Footnotes
1 Here the term “implicit” means that theories about willpower are most often not
articulated. However, we assume that people are able to recognize their beliefs when asked to
respond to items that make them explicit.
2 An important question concerns how self-regulatory demands change over the course of
an academic term. In past research, we theorized that demands increase as finals week
approaches (Job et al. 2010). The present data allow a test of this assumption. This was the case
for academic demands. Specifically, participants’ ratings of how much they would have to deal
with tests increased over the four weeks as indicated by a significant linear within-participant
contrast, F(1, 128)=59.20, p<.001. There was no such increase for non-academic demands;
indeed, social stressors (e.g., social obligations) declined as the end of the term approached, F(1,
128)=5.67, p=.04. A strength of the present study is that, rather than assuming that all students
are facing high or low demands at certain times, we assessed the level of demands each student
anticipated week by week over the second half of the term.

34
Table 1.
Means and standard deviations of self-regulation failure (1=never, 7=two or more times per day), negative affect (1=never, 5=very
often), and forecasted self-control demands (1=not at all, 4=very much).
Scale
Sample Item
Range
α
Week 1
Week 2
Week 3
Week 4
N
158
149
143
119
Procrastination
“How often did you meet friends
instead of studying?” (6 items)
1 - 7
.78-.87
3.37
(1.14)
3.30
(1.26)
3.29
(1.28)
2.92
(1.31)
Unhealthy Food
“How often did you eat chocolate or
candy bars?” (6 items)
1 - 7
.67-.74
2.99
(1.00)
2.96
(0.95)
2.98
(1.00)
2.95
(1.03)
Poor Time
Management
“How often did you come late to an
appointment?” (4 items)
1 - 7
.62-.76
1.77
(0.73)
1.81
(0.81)
1.73
(0.91)
1.41
(0.62)
Excess Spending
“How often did you buy something
knowing that it’s actually too
expensive for you?” (6 items)
1 - 7
.80-.91
1.58
(0.68)
1.53
(0.65)
1.64
(0.88)
1.45
(0.66)
Emotion Reg. Failure
““How often did you have trouble
controlling your temper?” (3 items)
1 - 7
.76-.87
2.06
(0.92)
2.03
(1.08)
2.00
(1.03)
2.00
(0.92)
Forecasted Self-
Regulatory Demands
“How much will you have to deal
with papers/essays due?” (13 items)
1 - 4
.66-.72
25.74
(5.70)
24.99
(6.14)
25.13
(5.79)
24.32
(5.92)

35
Table 2.
Descriptive statistics and correlations for participants-level variables.
Variable
2
3
4
5
6
M
SD
1.
Willpower Theory
-.07
-.13
-.01
-.17*
.10
3.88
.88
2.
Course Loada
-.01
.02
-.16+
-.39**
14.60
3.64
3.
Current Term GPA
.68**
.23**
.25**
3.57
.46
4.
Previous Term GPA
.17*
.28**
3.60
.42
5.
Trait Self-Control
-.10
3.07
.72
6.
Age
20.58
1.74
Note. N=149. a Units taken in the current term.
* p < .05 ** p < .001 (two-tailed).

36
Table 3.
Unstandardized coefficients from a multilevel linear model of self-regulation and affect
Predictor
Composite
Self-Reg.
Failure
Procrastination
Unhealthy
Food
Poor Time
Management
Excess
Spending
Emotion-
Regulation
Failure
G00 Intercept
2.31 (0.04)***
3.26 (0.08)***
3.00 (0.07)***
1.71 (0.05)***
1.56 (0.04)***
2.03 (0.06)***
G01 Willpower Theory
(WT)
0.08 (0.04)+
0.23 (0.09)*
0.04 (0.08)
0.11 (0.06)+
0.02 (0.05)
-0.02 (0.06)
G02 Mean Demands
0.95 (0.10)***
1.18 (0.22)***
0.55 (0.19)**
0.92 (0.13)***
0.73 (0.12)***
1.27 (0.15)***
G03 WT ×
Mean Demands
0.17 (0.04)***
0.33 (0.09)**
0.18 (0.08)*
0.05 (0.05)
0.12 (0.06)*
0.13 (0.06)*
G10 Weekly Demands
0.05 (0.06)
0.28 (0.14)*
-0.03 (0.10)
0.32 (0.10)**
-0.04 (0.10)
-0.23 (0.14)
G11 WT ×
Weekly Demands
-0.07 (0.7)
-0.01 (0.18)
-0.17 (0.12)
0.09 (0.12)
-0.00 (0.12)
-0.25 (0.17)
Note. WT = Willpower Theory; Standard errors are given in parentheses.
+p < .10. *p < .05. **p < .01. ***p < .001.

37
Figure Captions
Figure 1. Mean self-regulation failure (composite index) as a function of forecasted self-
regulatory demands and willpower theory. The limited and nonlimited-resource theory groups
represent participants 1 SD above and below the mean on the willpower-theories measure. High
vs. low demand represents ±1 SD from the grand mean.
Figure 2. Mean procrastination and unhealthy eating (1=never, 2=1 time per week, 3=2
times per week, 4=3-4 times per week, 5=5-6 times per week, 6=1 time per day, 7=two or more
times per day) as a function of self-regulatory demands and willpower theory. The limited and
nonlimited-resource theory groups represent participants 1 SD above and below the mean on the
willpower-theories measure. High vs. low demand represents ±1 SD from the grand mean.
Figure 3. Mean grade point average (GPA) as a function of academic demands (units
taken) and willpower theory (±1 SD on the willpower theory measure). High vs. low demands
represent ±1 SD from the mean. Means are adjusted for prior-term GPA, age, and sex.
Figure 4. Mediation model testing the indirect effect of willpower theory on GPA through
procrastination. * p < .05. ** p < .001.

38
Figure 1

39
Figure 2

40
Figure 3

41
Figure 4