A CRITICAL EXAMINATION OF THE EGO-DEPLETION EFFECT: CAN YOU VS. WILL
YOU ENGAGE IN EFFORTFUL SELF-REGULATION?
By
Emily Swensen Darowski

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Psychology
2011

ABSTRACT
A CRITICAL EXAMINATION OF THE EGO-DEPLETION EFFECT: CAN YOU VS. WILL
YOU ENGAGE IN EFFORTFUL SELF-REGULATION?
By
Emily Swensen Darowski
Self-regulation refers to people’s ability to regulate their own thoughts and
actions, and override them when necessary. Measures of self-regulation have been
observed to correlate with a number of important outcomes, including academic
achievement and job performance (Cleary, 2006; Gailliot, 2008; Mischel, Shoda, &
Peake, 1988). Nevertheless, there is still debate about how best to conceptualize selfregulation. Does self-regulation draw on ability or motivation?
This debate has been discussed, in recent years, in the context of a finding
referred to as the ego-depletion effect. This effect has typically been demonstrated
using a two-task procedure (e.g., Muraven, Tice, & Baumeister, 1998). In a depletion
condition, participants perform a task that presumably places a high demand on selfregulation, whereas in a control condition, participants perform a non-regulatory task. All
participants then complete a demanding regulatory task. In a number of studies,
participants in the depletion condition have performed worse on the second task than
participants in the control condition.
One prominent explanation of ego-depletion is the resource depletion account.
The basic idea of this account is that self-regulation draws on a limited resource that
becomes even more limited after engaging in self-regulation; this negatively affects
subsequent self-regulatory behavior. Thus, according to this account, ego-depletion
reflects an inability to continue regulating effectively.

There are several limitations of the resource account of ego-depletion and
alternative explanations are possible. Specifically, a motivation account suggests that
ego-depletion reflects an unwillingness to perform the second task rather than an
inability to do so. That is, it could be that individuals who have previously exerted more
effort may be less inclined to exert effort thereafter without sufficient motivation.
In this dissertation, I conducted five studies to investigate this possibility. In the
first study, I performed a meta-analysis and found a relationship between the depletion
effect and ratings of effort and difficulty on the first task. In the second study, I found a
depletion-like effect after two initial non-regulatory tasks that required different levels of
effort exertion. In the third and fourth studies, I found depletion-like effects using
motivation manipulations—even when, across conditions, participants’ resource was
presumably depleted to the same degree. In the fifth study, I investigated whether the
role of effort and motivation extended to a three-task conservation procedure (see
Muraven, Shmeuli, & Burkley, 2006). While I did not find evidence for conservation, I did
find some evidence that motivation continues to impact performance after two
demanding regulatory tasks. Together, these findings suggest that motivated effort
allocation has a significant impact on ego-depletion. Specifically, the results suggest
that individuals are less inclined to exert effort after prior exertion. If sufficiently
motivated, however, participants can and will exert effort.

Dedicated to Dad—the lesson you taught us most, the power of work, kept us all going
when you were gone—and to Mom—the lessons you teach us most, the power of love
and service, keep us living for the right reasons.

iv

ACKNOWLEDGEMENTS

I wish to thank my advisor Zach Hambrick for his guidance and continuous
support through the years and specifically for his tireless help in completing this
dissertation. I would like to thank Rick DeShon for the many meetings hashing and rehashing the depletion effect and for his wisdom about life in general. I would also like to
thank my other committee members Ryan Bowles and Tim Pleskac for their mentoring.
Thanks go to Rose T. Zacks, who was always an example of an outstanding researcher
and mentor. Many undergraduate research assistants helped me with data collection
and I would like to thank them here: Kyle Hines, Sarah Harfst, Lindsey Lewis, Bridget
Parler, Jooi Dave, Jason Maise, and David Fried.
In addition, I would like to thank the following people: my husband Joe for his
unconditional love and support, and for the many hours he helped me read and reread
this document; Lisa Helder Babcock, Nicole Moon, Thomas Wagner, and Nikole Patson
Huffman for their friendship through graduate school; my mom, Kate Darowski, and
Kelly Sherman, for watching Lizzie so I could work; and to many other friends and family
for believing in me.

v

TABLE OF CONTENTS
LIST OF TABLES...........................................................................................................viii
LIST OF FIGURES...........................................................................................................ix
CHAPTER 1: INTRODUCTION........................................................................................1
The Depletion Effect.........................................................................................................1
The Resource Depletion Model........................................................................................4
Limitations of the Resource Depletion Model...................................................................7
An Alternative Account of the Depletion Effect: Motivated Effort Allocation...................10
Cognitive Miser and Motivated Tactician.............................................................11
Sequential and Continuous Task Performance...................................................12
Overcoming Depletion Effects.............................................................................14
Overview of the Current Research.................................................................................15
CHAPTER 2: STUDY 1- META-ANALYSIS...................................................................17
Study 1 Rationale and Predictions.................................................................................17
Method...........................................................................................................................18
Results...........................................................................................................................19
Analysis Plan.......................................................................................................19
Data Preparation.................................................................................................20
Sample-Weighted Average Effect Sizes..............................................................22
Weighted Least Squares Analyses......................................................................22
Discussion......................................................................................................................23
CHAPTER 3: STUDY 2- NON-REGULATORY INITIAL TASKS....................................25
Study 2 Rationale and Predictions..................................................................................25
Method............................................................................................................................26
Participants..........................................................................................................26
Procedure and Materials......................................................................................27
Results............................................................................................................................29
Manipulation Checks............................................................................................29
Anagram Persistence and Number Correct.........................................................29
Discussion......................................................................................................................32
CHAPTER 4: STUDY 3- INFLUENCE OF GOAL SETTING..........................................33
Study 3 Rationale and Predictions..................................................................................33
Method............................................................................................................................34
Participants..........................................................................................................34
Procedure and Materials......................................................................................35
Results............................................................................................................................38
Pilot Study............................................................................................................38
Study 3 Manipulation Checks..............................................................................38
Study 3 Anagram Persistence and Number Correct............................................40
Discussion......................................................................................................................44

vi

CHAPTER 5: STUDY 4- INFLUENCE OF PERCEIVED EFFORT.................................46
Study 4 Rationale and Predictions..................................................................................46
Method............................................................................................................................47
Participants..........................................................................................................47
Procedure and Materials......................................................................................48
Results............................................................................................................................51
Manipulation Checks............................................................................................51
Anagram Persistence and Number Correct.........................................................53
Discussion......................................................................................................................57
CHAPTER 6: STUDY 5- CONSERVATION OF RESOURCE VS. EFFORT..................59
Study 5 Rationale and Predictions..................................................................................59
Method............................................................................................................................62
Participants..........................................................................................................62
Procedure and Materials......................................................................................62
Results............................................................................................................................68
Manipulation Checks............................................................................................68
Conservation........................................................................................................78
Anagram Persistence and Number Correct........................................................80
Anagram Performance Comparison....................................................................83
Discussion......................................................................................................................85
CHAPTER 7: GENERAL DISCUSSION.........................................................................90
Theoretical Implications..................................................................................................91
Further Issues, Limitations, and Future Directions.........................................................96
Conclusion....................................................................................................................100
APPENDIX A: META-ANALYSIS DETAILS.................................................................103
APPENDIX B: MANIPULATION CHECK QUESTIONS...............................................107
APPENDIX C: 48 ANAGRAMS FROM STUDY 2.........................................................110
APPENDIX D: DEMOGRAPHICS QUESTIONNAIRE..................................................111
APPENDIX E: TYPING TASK DETAILS......................................................................112
APPENDIX F: 80 ANAGRAMS FROM STUDIES 3 & 5...............................................114
APPENDIX G: STUDY 4 PUPIL DATA REPORT FOR
HIGH-PERCEIVED-EFFORT PARTICIPANTS............................................................115
APPENDIX H: STUDY 4 PUPIL DATA REPORT FOR
LOW-PERCEIVED-EFFORT PARTICIPANTS.............................................................117
APPENDIX I: 48 ANAGRAMS FROM STUDY 4..........................................................119
REFERENCES.............................................................................................................121

vii

LIST OF TABLES
Table 1: Descriptive and Inferential Statistics for Study 2 Questionnaires.....................30
Table 2: Descriptive and Inferential Statistics for Study 2 Anagram Performance.........31
Table 3: Descriptive Statistics for the Pilot Study and Study 3 Anagram
Performance...................................................................................................................41
Table 4: Descriptive and Inferential Statistics for Study 3 Questionnaires.....................42
Table 5: Descriptive and Inferential Statistics for Study 4 Questionnaires.....................54
Table 6: Descriptive and Inferential Statistics for Study 4 Anagram Performance.........56
Table 7: Design of Study 5.............................................................................................63
Table 8: Descriptive Statistics for Study 5 Manipulation Check Questions on the
Typing Task....................................................................................................................69
Table 9: Descriptive Statistics for Study 5 Manipulation Check Questions on the
Stroop Task....................................................................................................................72
Table 10: Descriptive Statistics for Study 5 Manipulation Check Questions on the
Anagram Task................................................................................................................75
Table 11: Descriptive Statistics for Study 5 BMIS, Vocabulary, and ACT Variables......77
Table 12: Descriptive Statistics for Study 5 Stroop Task Performance..........................79
Table 13: Descriptive Statistics for Study 5 Anagram Performance...............................82
Table A1: Condition ns and Sample Weights for the Meta-Analysis............................103

viii

LIST OF FIGURES
Figure 1: Number correct on the anagram task by goal type and prior task...................84
Figure 2: Persistence on the anagram task by goal type and prior task.........................85
Figure A1: False representation of pupil diameter data for participants in the highperceived-effort condition.............................................................................................114
Figure A2: False representation of pupil diameter data for participants in the lowperceived-effort condition.............................................................................................116

ix

CHAPTER 1: INTRODUCTION
Self-regulation has been defined as the ability to direct one’s thoughts and
actions (Muraven, Tice, & Baumeister, 1998) and can be used to control emotions,
drives, impulses, appetites, performance, and attention (Vohs & Baumeister, 2004).
Measures of self-regulation correlate with important outcomes such as relationship
satisfaction, academic achievement, and job performance (Cleary, 2006; Funder &
Block, 1989; Funder, Block, & Block, 1983; Gailliot, 2008; Mischel, Shoda, & Peake
1988). Specific examples of self-regulatory failures include binge eating, impulse buying,
unprotected sex, illicit drug use, gambling, violence, and procrastination (Baumeister &
Heatherton, 1996).
The Depletion Effect
The purpose of this dissertation was to investigate the nature of so-called egodepletion in self-regulation. This behavioral phenomenon, hereafter referred to as the
depletion effect, has typically been demonstrated using a two-task procedure. In a
depletion condition, participants perform a task presumed to place a high demand on
self-regulation, whereas in a control condition, participants complete a non-regulatory
task. All participants then complete a demanding regulatory task. The depletion effect
refers to the finding that performance on the second task is significantly worse in the
depletion condition than in the control condition (for a review, see Baumeister, Vohs, &
Tice, 2007). In general, the two tasks last 5 to 10 minutes each, indicating this effect
occurs over a relatively short time span (Hagger, Wood, Stiff, & Chatzisarantis, 2010).
One paradigm used in research on the depletion effect involves emotion
regulation. For example, in some studies, participants have watched an emotionally

1

evocative movie clip (e.g., a comedy act). In the depletion condition, participants are
instructed to suppress or exaggerate their emotional reaction to the movie clip, whereas
in the control condition, they simply watch the movie clip. In studies involving this
paradigm, performance decrements have occurred in depletion conditions relative to
control conditions on tasks designed to measure persistence (Baumeister, Bratslavsky,
Muraven, & Tice, 1998; Gailliot et al., 2007), physical endurance (Muraven et al., 1998),
reasoning (Schmeichel, Vohs, & Baumeister, 2003), fluency (Schmeichel, Demaree,
Robinson, & Pu, 2006), color-word interference (Inzlicht & Gutsell, 2007), and working
memory (Schmeichel, 2007).
Another paradigm involves attention regulation. For example, in some studies,
participants have examined a passage of text (Baumeister et al., 1998). In the depletion
condition, participants must find all of the Es within the text, but only cross off Es that
satisfy a given rule, whereas in a control condition, participants cross off all of the Es. In
studies involving this paradigm, performance decrements occurred in depletion
conditions relative to control conditions on tasks designed to measure persistence
(Baumeister et al., 1998; Tice, Baumeister, Shmueli, & Muraven, 2007) and attention
regulation (Muraven, Shmueli, & Burkley, 2006).
In paradigms such as these, the intended distinction between a self-regulatory
and non-regulatory initial task has been the involvement of response inhibition and
suppression (Muraven et al., 1998). The Stroop task, one of the most widely studied
inhibition tasks (MacLeod, 1991), has been utilized in depletion effect studies. In this
task, individuals identify what color a color-name appears in when the color and colorname match (low inhibition) or mismatch (high inhibition). When used as the initial task,

2

participants who completed high inhibition trials subsequently persisted less on
unsolvable puzzles relative to participants who solved low inhibition trials (Gailliot et al.,
2007; Webb & Sheeran, 2003). When used as the second task, performance was
reduced on high inhibition Stroop trials following tasks that involved regulating attention
(Gailliot et al., 2007), suppressing emotion (Inzlicht & Gutsell, 2007), and counting
backwards by 7s while standing on one leg (Webb & Sheeran, 2003).
To summarize, the depletion effect has occurred using a wide variety of tasks,
where researchers often manipulate one form of self-regulation with the first task (e.g.,
emotion regulation) and then test another form of self-regulation with the second task
(e.g., attention regulation). Tasks have involved factors ranging from suppressing
emotions and thoughts, to resisting temptation, to persisting at difficult problems, to
enduring a physical task, to reasoning and problem solving, and to working memory. In
a recent meta-analysis of depletion effect studies, the averaged corrected standardized
+

mean difference for performance on the second task was d = 0.62, representing a
medium-to-large effect size (Hagger et al., 2010).
Due to the impact of self-regulatory successes and failures on the individual and
society, it is important to understand why depletion effects occur (Baumeister &
Heatherton, 1996). The specific question motivating this research was whether
depletion effects occur because people are unable to engage in self-regulation, or
whether they are unwilling to do so. A resource account of depletion effects implies
individuals are incapable of self-regulation following a demanding self-regulatory task.
By contrast, a motivation account implies individuals choose how much effort to allocate
to self-regulation.

3

The Resource Depletion Model
The resource depletion model of self-regulation has three main assumptions. The
first assumption is that all self-regulatory thoughts or actions rely on a domain-general
mental resource (Muraven & Baumeister, 2000). In other words, any type of regulatory
action (e.g., emotional or attentional) will draw on the same resource. The second
assumption is that the resource is limited (Baumeister et al., 1998). This assumption
leads to the prediction that the ability to self-regulate at any given moment is determined
by the amount of the resource available, and whether that amount is sufficient to meet
regulatory demands. If demands on the resource are greater than the supply available,
then performance suffers (Just, Carpenter, & Miyake, 2003).
The final assumption is that the self-regulatory resource has limited strength.
That is, the resource is not only finite, but is also expendable. Baumeister and
colleagues described this assumption in terms of a muscle analogy: If a person uses a
muscle strenuously for some period, it becomes fatigued and is left in a state of reduced
capacity (see Muraven et al., 1998). If a person uses that same muscle immediately
afterward to do another task, performing that task is harder and less effective than if the
muscle had been at rest.
Stemming from the muscle analogy, there are two additional hypotheses of the
resource depletion model (Muraven et al., 1998). The recovery hypothesis is that the
resource returns to baseline capacity following a period of rest. The training hypothesis
is that practicing regulatory behaviors increases one’s ability to self-regulate effectively.
The benefit of training is not limited to the specific behavior practiced but applies to the
general use of the resource.

4

To test the recovery hypothesis, Tyler and Burns (2008) manipulated the amount
of time between the first task and the second task (i.e., 1, 3, or 10 minutes). The
dependent measure on the second task, handgrip duration, did not change significantly
for control participants across the time interval conditions. However, Tyler and Burns did
find that handgrip duration varied as a function of time interval in the depletion condition.
That is, relative to a baseline measure, handgrip duration was longest in the 10-minute
condition, shorter in the 3-minute condition, and still shorter in the 1-minute condition.
Tyler and Burns argued that as the interval between tasks increased, participants’
regulatory resource recovered to a greater extent, allowing them to hold the handgrip
longer. In a second study, Tyler and Burns found a reduced depletion effect for
participants who relaxed during a 3-minute interim period between the first and second
task, relative to control participants who simply waited during that period. Additionally,
Oaten, Williams, Jones, and Zadro (2008) found that the depletion effect was reduced
after a 45-minute break, although within the parameters of their study, the reduction was
limited to individuals low in social anxiety.
To test the training hypothesis, Muraven (2010) administered a single selfregulatory task, the stop signal task. In this task, participants pressed a key to indicate
whether a square appeared on the right- or left-hand side of a computer screen; when a
tone sounded, however, participants were instructed to withhold their response to the
stimuli. Then, during a two-week interim, participants assigned to training conditions
practiced self-regulation (e.g., avoiding sweets or practicing handgrip endurance), while
participants assigned to control conditions practiced non-regulatory activities (e.g.,
simple arithmetic). Finally, participants completed the stop signal task a second time.

5

From the first testing session to the second, Muraven found that participants’ stop signal
performance improved more in the training conditions than in the control conditions.
Using the typical two-task procedure in testing sessions, other researchers found a
reduced depletion effect in training participants relative to control participants (Gailliot,
Plant, Butz, & Baumeister, 2007; Muraven, Baumeister, & Tice, 1999). Oaten and
Cheng (2006a, 2006b, 2007) found similar results using training periods that lasted
several months. In these studies, participants engaged in study, exercise, or financial
monitoring programs as part of their regulatory training. Proponents of the resource
depletion model have argued that training reduces susceptibility to the depletion effect
by strengthening the capacity or stamina of the regulatory resource (Muraven et al.,
1999).
The resource depletion model offers one explanation of the depletion effect. That
is, in the typical two-task procedure, the initial regulatory task in the depletion condition
expends participants’ resources, whereas the initial non-regulatory task in the control
condition does not. Participants in depletion conditions perform worse on the second
task than participants in control conditions because, for a period, they cannot regulate
as effectively. According to this resource account, the ability to regulate falters in the
short-term, but this ability recovers with time and even increases through training.
Additionally, the depletion effect occurs across varying forms of self-regulation because
the regulatory resource is domain-general (Baumeister et al., 1998; Muraven et al.,
1998).

6

Limitations of the Resource Depletion Model
Though the resource depletion model provides an account of the depletion effect
and offers several additional hypotheses, this model has several limitations. First, as
others have noted, resource models are difficult to falsify and are based on circular
reasoning (e.g., Alberts, Martijn, Greb, Merckelbach, & de Vries, 2007; Hirst & Kalmar,
1987; Navon, 1984; Wickens, 1980). Navon (1984) suggested that resource models
may be useful, as long as there is an explanatory mechanism that provides testable
predictions about the resource. However, Baumeister et al. (1998) acknowledged that
the resource depletion model is limited because the nature of the resource is unknown.
Self-regulation presumably involves many processes (Wegner, 1994), and it is unclear
which process defines the resource and drives the depletion effect. When choosing
tasks that involve a high level of self-regulation, most researchers choose tasks that
require participants to override some response. However, there is no direct empirical
evidence showing that response inhibition is the driving mechanism underlying the
regulatory resource.
Second, while I have reported some support for the additional hypotheses of the
resource depletion model, this support is limited and deserves some qualifying
statements. A few studies showed that the depletion effect was reduced partially after a
3-minute break and more so after a 10-minute or 45-minute break (Oaten et al., 2008;
Tyler & Burns, 2008). However, most depletion studies have included a 2-5 minute
break between the first and second task when participants have filled out manipulation
check questionnaires (Baumeister et al., 1998; Muraven et al., 1998). Even with this
break, the depletion effect has occurred in many cases. In a study by Tice et al. (2007),

7

participants in a depletion condition had a 5-minute break without any intervening tasks
or manipulation checks, but still demonstrated performance decrements on the second
task relative to control participants. Muraven and Baumeister (2000) speculated that the
resource recovers more slowly than it is used. If this is the case, then the small break
between tasks may be insufficient to fully overcome the depletion effect. This would
mean that the brief interim period in most studies attenuates the depletion effect.
However, findings of a recent meta-analysis by Hagger et al. (2010) on the
depletion effect suggest this may not be the case. In this meta-analysis, the interim
period was included as a potential moderator of the depletion effect. Since very few
studies have indicated the exact length of the interim period, Hagger et al. coded the
interim period into the following categories: no activities reported, participants
completed questionnaires, or participants completed filler tasks/rested. Hagger et al.
predicted the depletion effect should be larger in studies that reported no activities since
the chance for recovery should be less in these cases. Instead, they found the opposite:
the depletion effect was significantly smaller in the studies that reported no activities
during the interim compared to studies that had questionnaires, fillers, or a break.
Hagger et al. indicated that crude coding and interrelationships among moderators may
have affected this finding. Overall, these mixed results suggest that it is inconclusive
whether recovery occurs in the manner proposed by the resource depletion model.
Evidence supporting the training hypothesis is more consistent. In their metaanalysis, Hagger et al. (2010) found that training has a large and significant effect, such
that practicing self-regulation has reduced the degree of depletion that participants
experience. One drawback of the training studies is that they do not address the

8

underlying mechanism driving the training effect. Training may improve the efficiency of
the resource or it may actually increase the capacity of the resource (Muraven et al.,
1999).
Alternatively, an interesting result from one of the training studies suggests that
individuals may benefit from training for another reason. Muraven et al. (1999)
examined the size of the depletion effect before and after a two-week period and found
mixed results. For several training conditions, or those who practiced self-regulation, the
size of the depletion effect was smaller after training than before, as predicted. For other
training conditions, the size of the depletion effect did not differ after training.
Unexpectedly, the control condition, or those who did not practice self-regulation,
showed a large increase in the size of the depletion effect after the interim period.
Muraven et al. (1999) decided to examine the influence of effort in these results
by assigning participants to high, low, and no effort groups based on their reports of
training activities. Those who exhibited high effort showed reduced effects of depletion
relative to the control condition. Those in the low effort and no effort groups did not differ
from the control condition. Thus, those who engaged more in the training regimens were
the ones who benefited the most and showed reduced effects of depletion (but see
Muraven, 2010).
To summarize, the resource account may be inadequate to fully explain the
depletion effect. The resource depletion model is limited because the nature of the
resource is ill defined, which makes it difficult to develop falsifiable hypotheses. Mixed
support for the additional hypotheses associated with the model also limits the resource

9

account. Furthermore, as I will discuss next, alternative accounts of the empirical results
are possible.
An Alternative Account of the Depletion Effect: Motivated Effort Allocation
According to the resource account, differences in the ability to regulate explain
the depletion effect. Alternatively, differences in emotion or motivation could explain the
effect. For example, participants in depletion conditions could experience negative
mood changes after completing the first task, which could then cause them to perform
worse on the second task relative to control participants. However, researchers have
successfully ruled out emotion as a driving factor in the depletion effect. Almost all
depletion studies have included a measure of mood and arousal in between the first and
second task and there have typically been no differences in positive or negative affect
or arousal after the first task (Baumeister et al., 1998; Muraven et al., 1998; Webb &
Sheeran, 2003). In cases where differences in mood or emotion emerged, including the
variable as a covariate rarely altered the results (e.g., Alberts et al., 2007).
While emotion does not seem to influence the depletion effect, a motivation
account has not been sufficiently explored. Such an account can be thought of in terms
of a general principle known as the law of least effort. As Eisenberger (1992) stated,
“performance evolves toward the minimal amount of energy expenditure required for
reinforcement” (p. 249). Rats demonstrated this by exerting the minimum effort required
to push levers for reinforcement and choosing the least effortful behavioral sequence
that led to a reward (Brener & Mitchell, 1989; Lewis, 1965; Solomon, 1948). When two
responses resulted in similar reinforcement, rats chose the less effortful response
(Lewis, 1965; Solomon, 1948).

10

Cognitive Miser and Motivated Tactician
The law of least effort extends to human behavior as well. Fiske and Taylor
(1984) introduced the concept of cognitive miserliness that suggested people often use
mental short cuts when processing information to exert minimal effort. Consistent with
this idea, research on judgmental biases showed that humans make decisions about
their environment using heuristics (Kahneman & Tversky, 1973). Other areas of
research indicate that people tend to use short cuts such as schemas, scripts,
stereotypes, and categorical judgments when perceiving the world (Operario & Fiske,
1999). According to the cognitive miser view, using these short cuts is automatic and
part of normal processing, even though they sometimes lead to errors.
Later, Fiske and Taylor refined the cognitive miser view and introduced the
concept of the motivated tactician (Fiske, 1992; Fiske & Taylor, 1991). The basic idea of
this concept is that people behave in miserly ways because they perform to a level that
is good enough for current goals and intentions. In other words, people will only give as
much effort as a given situation justifies (Brehm & Self, 1989). Fiske (1992) also
suggested that people do not engage in shortcuts if they are sufficiently motivated to
process information in other ways. Fiske (2004) cited research where activation of
stereotypes depended on several motivation variables such as temporary goals,
attitudes, values, and the immediate context. As an example, Ebenbach and Keltner
(1998) found that people in power were less motivated to judge people accurately (e.g.,
their judgments were stereotypical), whereas people who dislike conflicts were more
motivated to make accurate judgments.

11

Similarly, Hockey (1997) proposed that people have a range of effort they
typically exert. If circumstances call for additional effort, they can exert more effort or
reduce their performance standard; generally, people choose the latter. Meijman (1997)
suggested that choosing the path of least effort is beneficial for health; that is, making
high effort a habitual response has maladaptive effects similar to those of chronic stress
because both activate the same physiological systems. Ultimately, it appears efficient to
make decisions to exert the least amount of effort for the greatest amount of gain.
Lower effort processing can actually be quite useful and accurate. When faced
with making decisions with limited time and knowledge, people must often “take the
best” and easiest strategy. Gigerenzer and Goldstein (1996) used computer simulations
to test fast and frugal algorithms of inferential reasoning against classical rational
modes of reasoning and found that these two modes were comparable in terms of
speed and accuracy. Some results even favored the fast and frugal algorithms.
As cognitive misers and motivated tacticians, people are generally inclined to use
low effort strategies and processing. There are both psychological and physiological
benefits for doing so (Gigerenzer & Goldstein, 1996; Meijman, 1997). However, people
can increase effort when motivated (e.g., Ebenbach and Keltner, 1998).
Sequential and Continuous Task Performance
The cognitive miser and motivation tactician concepts are useful in explaining
behavior when there are sequential demands on effort. Kivetz and Zheng (2006) used
these concepts to explain how people balance hard work and rewards or relaxation.
They found that individuals who had invested greater effort on an initial task were more
likely to choose the easier or more gratifying of two options on a second task. For

12

example, after a more effortful first task, individuals were more likely to choose a
second task that involved watching movies than one that involved self-evaluation.
These authors argued that individuals use hard work as a way to justify or entitle selfgratification.
Similar trade-offs occur in the workplace. When people must perform cognitively
demanding tasks sequentially or continuously, performance suffers, a phenomenon
known as after-effects (Broadbent, 1979). In many after-effects studies, researchers
have manipulated cognitive demand between experimental and control conditions.
Participants in experimental conditions consistently reported higher fatigue compared to
controls (Lorist et al., 2000; Schellekens, Sijtsma, Vegter, & Meijman, 2000; van der
Linden, Frese, & Meijman 2003a; van der Linden, Frese, & Sonnentag, 2003b; van der
Linden, Massar, Schellekens, Ellenbroek, & Verkes, 2006). More important, self-report
measures indicated that experimental participants, relative to control participants, were
less willing to exert effort and even resisted exerting effort (van der Linden & Eling, 2006;
van der Linden et al., 2003a). Experimental participants also reported an aversion to
task continuation (Boksem, Meijman, & Lorist, 2005; Lorist et al., 2000).
Research also demonstrated that participants overcame after-effects if they were
motivated. Boksem, Meijman, and Lorist (2006) had participants complete a task
consisting of seven 20-minute intervals. Before the last interval of the task, participants
were told that their performance would be compared to others’ and that the best
performer would receive a cash reward. Whereas performance began to decline during
the previous intervals, performance improved significantly after the motivation
intervention. Boksem et al. explained their findings in terms of an effort/reward

13

imbalance. Specifically, they suggested that maintaining performance required a high
level of effort. When there was no reason or reward to maintain effort, performance
declined. When offered a reward for the last interval, participants were motivated to
increase their effort and improved performance. In other words, participants behaved as
motivated tacticians.
Overcoming Depletion Effects
Consistent with the motivated tactician concept, Schellekens et al. (2000)
suggested individuals avoid the depletion effect as long as they are willing to expend
more effort. In fact, several studies showed that individuals overcome the depletion
effect through motivation. As examples, Martijn, Tenbult, Merckelbach, Dreezens, and
de Vries (2002) demonstrated that when participants were told the depletion effect
should not occur, it did not. Muraven and Slessareva (2003) eliminated the depletion
effect by telling participants that performance on the second task was beneficial to
themselves and others. Furthermore, Alberts et al. (2007) primed persistence before the
second task and influenced participants to engage longer in a physical endurance task
despite an initial depleting task. Finally, in their meta-analysis, Hagger et al. (2010)
combined data across several studies that included a motivation manipulation and
found a large effect. That is, depleted participants given motivation strategies performed
better on the second task compared to depleted participants with no motivation
strategies.
The motivated tactician concept may explain the benefits of self-regulatory
training as well. Rather than strengthening the resource in some way, as assumed in
the resource depletion model (Muraven et al., 1999), an alternative view is that after

14

practicing self-regulation, individuals increase effort tolerance or endurance. That is,
individuals motivate themselves to engage in high effort tasks or strategies for longer
periods.
The effects of training may link to research on learned industriousness, which
refers to the finding that people adapt to expected levels of effort. When people receive
rewards or reinforcement for high effort, they perform accordingly (Eisenberger, 1992).
Converse and DeShon (2009) found that the depletion effect occurred after one effortful
regulatory task, but not after two. They suggested in a two-task procedure, individuals
did not have enough experience to learn the expected level of effort exertion and adapt
to that level. Individuals did adapt, however, when performing a three-task procedure.
Muraven et al. (2006) used a three-task procedure and found that people were
more likely to conserve (and perform worse) on a current task if they anticipated future
self-regulatory demands. Hagger et al. (2010) found a large effect of conservation
manipulations across several studies. Essentially, the concept of conservation is an
extension of the motivated tactician concept. That is, when people anticipate a need for
high effort in the future, they will adjust their current effort to a lower level (e.g., Muraven
et al., 2006) unless there is sufficient motivation to continue at the same level (e.g., via
reward, Converse & DeShon, 2009).
Overview of the Current Research
To summarize, research on decision making (e.g., heuristics) and workplace
performance (e.g., after-effects) provides evidence for miserliness and motivated effort
tactics. These concepts readily apply to self-regulation and the depletion effect. Existing
research has shown that people can overcome the depletion effect, not by rest as the

15

resource account suggests, but by increasing motivation (e.g., Muraven & Slessareva,
2003). That is, people adaptively adjust their effort in response to external (e.g., rewards
and future demands) and internal (e.g., justification) motivation cues. The effects of
training may even result from learned industriousness rather than greater resource
capacity or stamina (Converse & DeShon, 2009). Overall, this evidence supports a
motivation account rather than a resource account of the depletion effect.
The purpose of my dissertation was to further test whether changes in the
motivation to exert effort can provide an alternative explanation of the depletion effect. I
conducted five studies to investigate this account. In the first study, I performed a metaanalysis examining the relationship between the depletion effect and self-report ratings
of effort, difficulty, and fatigue following the first task. In the second study, I examined
whether a depletion-like effect could occur after an effortful task that did not require selfregulation. In the third and fourth studies, I examined whether I could create a depletionlike effect among participants objectively depleted to the same degree by using
motivation manipulations. In the fifth study, I examined whether the depletion effect is
attributable to resource conservation or effort conservation.

16

CHAPTER 2: STUDY 1- META-ANALYSIS
Study 1 Rationale and Predictions
When asked to rate effort after an initial task, participants in depletion conditions
have often indicated the task required more effort than participants in control conditions
(Baumeister et al., 1998; Muraven et al., 1998). Participants have also rated the first
task in depletion conditions as more difficult than in control conditions (Muraven et al.,
1998; Schmeichel, 2007; Vohs & Schmeichel, 2003; but see Schmeichel et al., 2006).
When participants reported a personal experience in which they successfully selfregulated, Muraven et al. found they often mentioned exerting significant effort.
Furthermore, self-reported fatigue or exhaustion has increased more after an initial
depleting task than after an initial control task (Alberts et al., 2007; Baumeister et al.,
1998; Muraven et al., 1998; Webb & Sheeran, 2003).
Proponents of the resource account have argued that these manipulation checks
corroborate the notion that differences in self-regulatory demands on the first task
primarily influence the size of the depletion effect. Alternatively, it seems possible that
differences in effort or difficulty demands on the first task primarily influence the size of
the depletion effect. That is, performance may drop after the first task not because the
first task required self-regulation per se, but rather because it required a high degree of
effort. Differences in effort demands may also lead to the proposed domain generality of
the resource. In other words, the type of task does not matter so long as the first and
second task demand a higher level of effort.
To investigate whether these self-reported differences on the first task related to
the size of the depletion effect, I performed a meta-analysis on a subset of published

17

depletion studies. If variables such as effort have a significant impact on the depletion
effect, there should be a strong positive relationship between the two. That is, greater
differences in effort ratings on the first task should be associated with greater
differences in performance on the second task, suggesting that the more effort
participants exert initially, the less effort they exert subsequently.
Fortunately, the meta-analysis by Hagger et al. (2010) provided the effect sizes
from a large number of depletion studies. These studies used the two-task procedure or
similar designs (see Hagger et al. for a complete discussion of search and inclusion
criteria). These researchers also provided the effect sizes of commonly used
manipulation check questions. Although they showed signs of heterogeneity, Hagger et
al. reported that the self-report effect sizes for effort, difficulty, and fatigue were
significantly different from zero and were medium to large in magnitude. Effect sizes for
other self-report items (i.e., positive affect, negative affect, and self-efficacy) were much
smaller and two of them were not significantly different from zero.
Using their supplemental materials, I selected all of the studies that provided the
depletion effect size as well as the effect sizes for self-report ratings of effort, difficulty,
or fatigue. I predicted the self-report effects would have a positive relationship with the
depletion effect. I also predicted that the effort effect would have the strongest
relationship with the depletion effect compared to the effects of difficulty and fatigue.
Method
In total, I identified 124 tests of the association between the depletion effect and
self-reported effort, difficulty, or fatigue. I removed eight tests where researchers did not

18

1

predict that a depletion effect would occur. For example, I only included one test from
Schmeichel’s second study (2007) where participants completed a digit span task for
the dependent measure. The task had forward digit span trials, where participants
recalled digits in the order given, and reverse digit span trials, where participants
recalled digits backwards. Schmeichel predicted a depletion effect on the reverse trials
because people must use complex executive functions to recall the digits backwards.
He predicted there would be no depletion effect on forward trials because people can
use basic memory processes to recall the digits in order. I included the test involving
reverse digit trials, but not the test involving forward digit trials. The final data set of 116
tests (effort, k = 31; difficulty, k = 59; and fatigue, k = 26) came from 87 studies reported
in 48 research articles.
Results
Analysis Plan
First, with a separate analysis for each, I calculated the sample-weighted
average effect size for the depletion effect and for the effort effect, difficulty effect, and
fatigue effect. Second, I performed a weighted least squares (WLS) analysis with the
depletion effect regressed on the self-report effect. Thus, I performed the WLS analysis
three times with effort, difficulty, or fatigue as the predictor variable. In each WLS
analysis, I included sample size as a weighting variable.
1

The following tests of the depletion effect and self-report effect were removed from the
analyses: Schmeichel (2007, Study 2: 1 test where the dependent variable was forward
digit span performance); Tyler & Burns (2009, Study 2: 1 test from the 20-minute task
duration condition); Wright et al. (2007, Study 1: 2 tests from the low and high relevance
conditions that had a low performance standard; Study 2: 1 test from the low relevance
condition); Wright, Martin, & Bland (2003, Study 1: 2 tests from the low standard
condition); Wright, Stewart, & Barnett (2008, Study 1: 1 test where the dependent task
was a non-regulatory math task).
19

Data Preparation
There are several points to make regarding the data set. First, if studies reported
multiple manipulation checks, such as effort and difficulty, I included the effort effect and
depletion effect together as one test and the difficulty effect and depletion effect
together as another test (e.g., Baumeister et al., 1998, Study 1). I used these separate
tests in their respective WLS analysis. However, I included the depletion effect, which
was the same for both of these tests because it came from the same study, only once
when calculating the sample-weighted average depletion effect.
Second, Hagger et al. (2010) reported a total sample size for each study, but did
not report condition ns separately. I went back to the original studies to determine the
sample size for each condition as precisely as possible. I noticed several cases where
the sample size used to compute the depletion effect differed from the sample size used
to compute the self-report effect. In most of these cases, the study involved multiple
conditions, but not all of them were a test of the typical depletion effect (e.g., in Muraven
& Slessareva (2003) some conditions received a motivation manipulation). It appeared
that researchers based the depletion effect on a simple main effect, whereas they
based the self-report effect on a main effect. When computing sample-weighted
average effect sizes, I used the simple main effect condition ns for the depletion effect
2

and the main effect condition ns for the self-report effect. When conducting the WLS

2

I have included a table of the ns I used in Appendix A as a reference. I did conduct the
WLS analyses using the ns reported by Hagger et al. (2010). The results differed
somewhat, with the r decreasing slightly from .263 to .241 in the effort WLS analysis,
and increasing moderately from .213 to .265 and from .048 to .111 in the difficulty and
fatigue WLS analyses, respectively. As I believe the weights I used were more accurate,
as well as more conservative, I discussed and interpreted the results based on those
weights.
20

analyses, I used the smallest combined sample size as the weight for each study. This
provided the most conservative estimate of the relationship under investigation.
Third, there were two cases where a single study had more than one betweensubjects test of the depletion effect. In Schmeichel’s first study (2007), half of the
participants received a sentence span task as the dependent measure and half of the
participants received an operation span task as the dependent measure. In Tyler and
Burns’ study (2009), there were three sets of control/depletion conditions based on the
length of the first task: 3-minutes, 10-minutes, or 20-minutes. I removed the 20-minute
condition from the analyses as specified earlier (see Note 1), but kept the 3-minute and
10-minute conditions. I included a depletion effect from each set of these conditions
when calculating the sample-weighted average depletion effect size. However, because
researchers reported the main effect of self-report ratings and not a simple main effect,
this meant the self-report effect size was the same for both sets of conditions coming
from one study. In these cases, I included the self-report effect size only once when
calculating the sample-weighted average self-report effect size. I included each set of
conditions in the WLS analyses.

3

3

I randomly removed one test from each of these two studies, thus eliminating the
repetition of a self-report effect size. This removal only affected the difficulty WLS
analysis; despite removing only two data points, it resulted in r decreasing from .213
to .163. By random selection, one of the data points I removed was in the upper right
quadrant of the data spread. If I keep this data point and remove the data point from the
other set of conditions in this study, the r decreased less dramatically from .213 to .201.
If anything, this may suggest that potential outliers exaggerated the relationship
between the depletion effect and difficulty ratings.

21

Sample-Weighted Average Effect Sizes
Effect sizes (d) for the second task (i.e., the depletion effect) ranged from 0.12 to
2.60 (k = 90). The sample-weighted average effect size across all of the tests was d+ =
0.63 with a 95% confidence interval from 0.57 to 0.69. This suggests that differences
between conditions on the first task have a medium effect on second task performance.
Effect sizes (d) for self-reported effort on the first task ranged from 0.06 to 1.89 (k = 31).
The sample-weighted average effect size across all of the tests was d+ = 0.69 with a
95% confidence interval from 0.60 to 0.79. The demands of the initial task had a
medium effect on effort ratings (Cohen, 1987). Effect sizes (d) for self-reported difficulty
on the first task ranged from -0.31 to 3.39 (k = 57). The sample-weighted average effect
size across all of the tests was d+ = 1.03 with a 95% confidence interval from 0.95 to
1.10. The demands of the initial task had a large effect on difficulty ratings. Effect sizes
(d) for self-reported fatigue on the first task ranged from -0.43 to 1.67 (k = 26). The
sample-weighted average effect size across all of the tests was d+ = 0.56 with a 95%
confidence interval from 0.46 to 0.66. The demands of the initial task had a medium
effect on fatigue ratings. As expected, since this is a subset of the larger meta-analysis,
these sample-weighted average effect sizes are similar to those reported by Hagger et
al. (2010) in magnitude.
Weighted Least Squares Analyses
Next, I performed weighted least squares analyses to examine the relationship
between self-report effects and the magnitude of the depletion effect. I regressed the
depletion effect onto each type of self-report rating while including sample size as a
weighting variable. In this context, r reflects the effect size between the self-report effect

22

and the depletion effect. These analyses revealed a medium effect of effort on the
depletion effect, r = 0.263, 90% CI [-0.03, 0.56], a small effect of difficulty on the
depletion effect, r = 0.213, 90% CI [0.00, 0.43], and no effect of fatigue on the depletion
effect, r = 0.048, 90% CI [-0.29, 0.39].
Discussion
The meta-analysis revealed a relationship between effort and the depletion effect
and between difficulty and the depletion effect. As predicted, the strongest effect
emerged when I included effort ratings in the WLS analysis. The confidence intervals for
these effects contained zero, but the self-report ratings generally came from a singleitem question, with unknown reliability and validity. Even so, I found a potentially
important relationship between effort and the depletion effect, a relationship unexplored
previously. With better measures and a larger sample of studies, this effect might prove
even stronger.
Contrary to my predictions, no relationship emerged between fatigue and the
depletion effect. Post hoc, this finding can be interpreted as support for the motivation
account versus the resource account. If an inability to perform drives the depletion effect,
individuals should be truly fatigued. Greater fatigue differences on the first task should
then correlate with an increased depletion effect. This effect did not come out in the
analysis. Moreover, it seems unlikely that regulating for 5-10 minutes, the length of most
initial tasks, could produce true fatigue.
Instead, it seems more likely that when individuals feel like they exerted
significant effort or when they feel that a task was quite difficult, they do not work as
hard when they encounter another effortful or difficult task. This pattern of results is

23

consistent with the concept of cognitive miserliness. Overall, this meta-analysis
provided evidence that connects the depletion effect to motivated allocation of effort.

24

CHAPTER 3: STUDY 2- NON-REGULATORY INITIAL TASKS
Study 2 Rationale and Predictions
In previous studies of the depletion effect, regulatory demands may have been
confounded with effort demands. As the meta-analysis results suggested, the depletion
effect relates to differences in effort exertion on the initial task, not necessarily to
differences in regulatory exertion. A few studies have examined this possibility. In
Schmeichel’s third study (2007), there were three levels of the first task: a low-demand
condition, a high-demand/no self-regulation condition, and a high-demand/regulation
condition. Participants rated the difficulty of these conditions in a graded fashion. On the
second task, only the high demand/regulation condition showed impaired performance.
The low demand and high demand/no self-regulation conditions did not differ from each
other. As another example, Muraven et al. (1998) found the depletion effect after a task
that involved thought suppression, but not after a 3-digit multiplication task, even though
participants rated the math task and the suppression task similarly in difficulty. Muraven
and colleagues interpreted these results as support for the resource account and
argued that the type of task does matter. That is, the depletion effect only occurs after
an effortful self-regulatory task, not after an effortful non-regulatory task.
However, the majority of studies on the depletion effect have reported differences
in effort and/or difficulty on the first task and researchers have rarely explored
performance outcomes after non-regulatory tasks. With this in mind, I designed the
second study to explore whether a depletion-like effect would occur using nonregulatory initial tasks that required different levels of effort. According to the resource
account, the depletion effect should not occur using non-regulatory initial tasks because

25

they do not expend the regulatory resource. By contrast, according to a motivation
account, individuals should be sensitive to effort demands, regardless of task type. I
predicted that participants who performed a more effortful task first would perform
significantly worse on the second task than those who performed a less effortful task
first.
The two non-regulatory initial tasks involved mathematics: a single-digit addition
task and a 3-digit multiplication task. Single-digit addition is relatively automatic and less
effortful than three-digit multiplication, a task that requires greater attention and which
people rarely do by hand. Muraven et al. (1998) argued people solve these types of
math problems using standard procedures that do not involve overriding responses;
thus, they are non-regulatory tasks. Muraven et al. even used 3-digit multiplication as a
control task in one study, whereas I am using it as the more effortful task in this study.
For the regulatory second task, I used an anagram task, in which approximately half of
the anagrams were unsolvable. Researchers have often used anagram tasks in
depletion studies, arguing that they involve effortful regulation because individuals must
persist in the face of failure and override the desire to quit (Baumeister et al., 1998;
Vohs & Heatherton, 2000).
Method
Participants
Thirty-two undergraduates from the subject pool at Michigan State University
participated in this study for partial course credit. Three participants were removed from
the low-effort addition condition, one because there was a timing error and two because
they completed the provided sheets of math problems before the time limit was up. The

26

final sample (low-effort condition, n = 13; high-effort condition, n = 16) had a mean age
of 19.45 (SD = 1.50) and was 72% female. A post hoc power analysis indicated I
achieved acceptable power (> .80, alpha = .05, two-tail) to reject the null hypothesis that
there would be no difference in performance across conditions (Faul, Erdfelder, Lang, &
Buhner, 2007).
Procedure and Materials
Participants completed the study individually in five steps: (a) a timed math task
and manipulation check questions, (b) the Brief Mood Introspection Scale (Mayer &
Gaschke, 1988; BMIS), (c) an anagram task and manipulation check questions, (d) a
demographics questionnaire, and (f) additional manipulation check questions.
Participants were randomly assigned to a low-effort condition or a high-effort
condition. The procedure for the two conditions differed only on the initial math task.
Participants in the low-effort condition solved single-digit addition problems and
participants in the high-effort condition solved 3-digit multiplication problems.
Participants in both conditions worked on this task for 8 minutes using paper and pencil,
without the aid of a calculator. After the first task, participants answered several
manipulation check questions regarding their effort and the difficulty of the math task
(see Appendix B for the manipulation check questions used for the tasks in this and
subsequent studies). On all manipulation check questions, unless otherwise noted,
participants read each question and rated themselves on a scale from 1 to 100.
Participants also completed the BMIS to rule out differences in mood or arousal as an
alternative explanation for the results (e.g., Muraven et al., 1998). On this questionnaire,

27

participants used a 4-point scale to rate how well mood adjectives described
themselves at that moment.
Then, participants in both conditions completed an anagram task in which
approximately half of the items were unsolvable, although participants were not aware
of this (see Appendix C). The task consisted of a list of 48 anagrams. An experimenter
told the participants to solve the anagrams by unscrambling the letters to form a
common English word. The experimenter gave an example of an anagram and
identified the correct solution. The experimenter instructed the participant to begin
working and stated, “When you have decided that you have worked on the task long
enough (or when you want to stop), please tell me that you are done.” The experimenter
started a stopwatch as soon as the participant began the task and stopped it when the
participant quit. Persistence (time on task) was the dependent variable.
After the second task, participants answered several manipulation check
questions regarding their effort and the difficulty of the anagram task. Then they
completed a demographics questionnaire on which they reported their gender, age, and
ACT scores (see Appendix D). I used ACT scores to ensure that performance
differences on the anagram task across conditions were not attributable to differences in
general cognitive ability. Finally, participants completed a study questionnaire that
asked them to provide ratings for the first and second task on several aspects such as
task difficulty, motivation to perform, effort required, and effort invested.

4

4

In Studies 3-5, participants did not rate effort and difficulty of each task on the final
study questionnaire. Thus, I only reported effort and difficulty ratings from Study 2 that
were given directly following each task.
28

Results
Manipulation Checks
Table 1 displays the means, standard deviations, and inferential statistics for the
manipulation check questions. After the first task, participants who completed the
single-digit addition task indicated that they invested less effort than participants who
completed the 3-digit multiplication, t(27) = -4.32, p < .001, d = -1.67. Participants also
rated the single-digit addition task as less difficult, t(27) = -5.33, p < .001, d = -2.06, and
easier to work on, t(27) = -5.39, p < .001, d = -2.08, than the 3-digit multiplication task.
Regarding the second task, there was a near significant difference in effort
investment, t(27) = 1.98, p = .058, d = 0.77. When looking at the t-test where equal
variances were not assumed, the result for this variable was significant, p = .044.
Participants in the low-effort addition condition indicated they invested more effort on
the anagram task than participants in the high-effort multiplication condition. Otherwise,
participants did not differ in difficulty ratings or effort given ratings (ps > .20).
Items on the BMIS can be used to create two mood scales: Pleasant-Unpleasant
and Arousal-Calm. There were no significant differences between conditions on the
BMIS scales or ACT scores (ps > .20).
Anagram Persistence and Number Correct
Table 2 displays the means, standard deviations, and inferential statistics for
anagram performance. As expected, participants in the high-effort multiplication
condition persisted significantly less on the anagrams than participants in the low-effort
addition condition, t(27) = 2.97, p < .01, d = 1.14. The difference in number of correct
anagram solutions across conditions was non-significant (p > .96).

29

Table 1: Descriptive and Inferential Statistics for Study 2 Questionnaires
High-Effort Multiplication
Condition
95 % CI
Variable
Math Task Ratings
Effort Invested
Task Difficulty
How hard did
you work?
BMIS
PleasantUnpleasant
Scale
Arousal-Calm
Scale
Anagram Task
Ratings
Effort Invested
Task Difficulty
How hard did
you work?
ACT

Low-Effort Addition
Condition
95 % CI

M SD

LL

UL

M SD

66.44 30.12
39.94 24.19

51.68
28.09

81.20
51.79

19.77 27.45
3.85 2.91

48.94 26.35

36.03

61.85

8.31

t(27)

a

LL

UL

4.85
2.27

34.69
5.43

6.97

4.52

12.10

-5.39

-4.32
-5.33

b

p

Cohen’s
d

<.001
<.001

-1.67
-2.06

<.001

-2.08

.54

0.26

.37

0.38

7.23

6.08

3.92

10.54

8.67

5.23

5.71

11.63

0.63

14.92

4.01

12.74

17.10

16.33

3.68

14.25

18.41

0.91

79.00 29.27
94.00 7.16

64.66
90.49

93.34
97.51

95.69
95.85

8.59
6.36

91.02
92.39

100.00
99.31

1.98
0.73

.058
.48

0.77
0.27

90.75 13.38

84.19

97.31

95.92

6.38

92.45

99.39

1.28
c
-1.27

.21

0.49

b

25.47 2.59
24.16 26.78
23.92 3.73
21.81
26.03
.21
-0.51
Note. CI = Confidence Interval; LL = lower limit; UL = upper limit.
a
The statistics reported here are the ts associated with the test when equal variances were assumed. Although there
were cases in which Levene’s test was significant, in no case was the t associated with equal variances assumed
b
significant when the t associated with equal variances not assumed was not significant. These t-tests had 23 df because
c

some participants did not receive all of the manipulation check questions. This t-test had 25 df because some
participants did not report an ACT score.

30

Table 2: Descriptive and Inferential Statistics for Study 2 Anagram Performance
High-Effort Multiplication
Condition
95 % CI
Variable
Persistence
(in minutes)

M SD
10.26 4.71

LL

UL

7.95

12.57

Low-Effort Addition
Condition
95 % CI
M SD
18.11 9.23

LL

UL

13.09

23.13

t(27)

a

2.97

p

Cohen’s
d

.006

1.14

Correct Solutions
1.56 1.26
0.94
2.18
1.54 1.33
0.82
2.26
-0.05
.96
-0.08
Note. CI = Confidence Interval; LL = lower limit; UL = upper limit.
a
The statistics reported here are the ts associated with the test when equal variances were assumed. Although there
were cases in which Levene’s test was significant, in no case was the t associated with equal variances assumed
significant when the t associated with equal variances not assumed was not significant.

31

Discussion
Baumeister and colleagues have argued that the depletion effect only occurs
after tasks that involve self-regulation and not after non-regulatory tasks (e.g., Muraven
et al., 1998). However, in this study, I induced a depletion-like effect using two nonregulatory tasks that differed in terms of effort exertion. Whereas Muraven et al. did not
find the depletion effect after a 3-digit multiplication task, I did find a depletion-like effect
after a multiplication task when compared to a less effortful single-digit addition task.
Specifically, participants persisted less after a multiplication task than after an addition
task. Performance did not differ in number correct; however, I expected no difference
here because about half of the anagrams were unsolvable and the remaining anagrams
were very difficult. These results suggest that researchers have overlooked the impact
of effort in previous research. Furthermore, these results are inconsistent with a
resource account of the depletion effect, but are entirely consistent with a motivation
account of this effect.

32

CHAPTER 4: STUDY 3- INFLUENCE OF GOAL SETTING
Study 3 Rationale and Predictions
When people act as motivated tacticians, they choose whether to use low- or
high-effort strategies. I designed Study 3 to investigate this idea by examining whether
performance on a second task in a depletion paradigm changes depending on the
motivation to exert effort, even when all participants have completed the same initial
task. According to the resource account, if the first task is the same across conditions,
and thus expends the regulatory resource to the same degree, then performance on the
second task should not differ. By contrast, according to a motivation account, if
individuals are sensitive to motivation manipulations, then performance on the second
task should depend on how motivated individuals are to expend effort, regardless of
prior regulatory expenditure.
Specifically, Study 3 investigated whether performance would increase if
participants were motivated to exert more effort. Participants completed the same two
tasks. After the first task, I manipulated performance goals. Half of the participants were
given a specific, difficult goal for the second task, whereas the other half of the
participants were given instructions to do their best.
I based this manipulation on the hypothesis that goals motivate action (Latham &
Locke, 1991). Goals can range from vague (e.g., do-your-best) to specific (e.g.,
complete 32 problems within 10 minutes) and from easy (e.g., perform better than 10%)
to difficult (e.g., perform better than 90%). A consistent finding coming out of goal
setting research is that individuals given specific, difficult goals perform better than
individuals given do-your-best instructions (e.g., Lock, Shaw, Saari, & Latham, 1981).

33

According to Latham and Locke (1991), do-your-best instructions lead to different
performance outcomes, many of which fall short of a person’s best. Research has
shown that people given do-your-best instructions evaluate their performance more
positively (Kernan & Lord, 1989) and anticipate that they will be satisfied at any level of
performance (Mento, Locke, & Klein, 1990). Latham and Lock suggested that people
with do-your-best instructions give themselves the “benefit of the doubt” when they
evaluate their performance. Specific, difficult goals, on the other hand, set the bar for
what constitutes good performance and motivates greater effort exertion than do-yourbest instructions.
Accordingly, I predicted that even though resource expenditure after the first task
should not differ across participants, performance on the second task would differ
depending on what instructions the participants received for that task. That is, I
predicted that participants given a specific, difficult goal would be more motivated to
continue exerting effort on the second task and would thus outperform participants told
to do their best on the second task.
Method
Participants
Seventy-three undergraduates were recruited from the subject pool at Michigan
State University to participate in this study for partial course credit. I removed three
participants from the analyses (one appeared intoxicated, one did not follow the typing
instructions, and one did not follow the typing instructions or the anagram timing
instructions). This left a sample of 70 (do-your-best condition, n = 36; specific-difficultgoal condition, n = 34). The final sample had a mean age of 20.17 (SD = 2.13) and was

34

54% female. A power analysis indicated a sample of this size would have acceptable
power (> .80, alpha = .05, two-tail) to reject the null hypothesis that there would be no
difference in performance across conditions (Faul et al., 2007).
Procedure and Materials
Participants were randomly assigned to one of two conditions: a do-your-best
condition or a specific-difficult-goal condition. Participants completed the study
individually in eight steps: (a) a timed typing task and manipulation check questions, (b)
the BMIS, (c) a goal manipulation segment, (d) an anagram task and manipulation
check questions, (e) filler tasks (this part was conditional, which I explain later), (f) a
demographics questionnaire, (g) a vocabulary test, and (h) a study questionnaire.
Participants expected the session would last one hour.
An experimenter seated the participant in front of a computer to perform the
typing task and read the instructions for the task aloud (see Appendix E for detailed
instructions and the text of the typing task). Researchers have used a related paper and
pencil task (e.g., Baumeister et al., 1998) and a computerized version of the typing task
(Muraven et al., 2006) in prior studies. I adapted the present version to collect key
presses and run with an eye-tracking device, which I needed for another study.
The instructions indicated a passage would appear on the screen and the
participant needed to type out the passage, including punctuation and capitalization.
However, the participant was instructed to type the passage without typing spaces or
the letter E. The instructions also indicated that the participant would be unable to see
what was typed, but that he or she should try to do his or her best, correct mistakes,
and continue typing until told to stop. After 6 minutes had elapsed, the typing task

35

ended and the experimenter asked the participant to fill out manipulation check
questions regarding their effort and the difficulty of the typing task. Participants also
completed the BMIS.
Then, participants in both conditions completed an anagram task. For this study,
80 moderately easy and solvable anagrams were taken from Gilhooly and Johnson
(1978; see Appendix F). Baumeister et al. (1998) used solvable anagrams as their
second task and argued that completing solvable anagrams still required self-regulation
because participants had to persist at making and breaking combinations of letters in
order to come up with the correct solution, despite initial failures. I used these easier,
solvable anagrams so the performance goal would be attainable.
The instructions for the anagram task were the same as those used in Study 2,
but were presented on the computer. The experimenter read them aloud to the
participant. An added section manipulated performance goals. Participants in the doyour-best condition were told, “Do your best to solve as many anagrams as you can.”
Participants in the specific-difficult-goal condition were told, “Your goal is to identify at
least 50 different words that can be created from the set of anagrams. Please try and
reach or surpass this goal.” This goal was the 85th percentile for the number of
anagrams participants solved correctly in a separate pilot study (see the results section
of Study 3 for more details).
To reduce experimenter error in timing persistence, the computer time-stamped
when the participant started and stopped working on the anagrams. Specifically, the
participant was instructed to press enter on the computer when starting the task and to
press escape on the computer after deciding to stop. The participant completed the task

36

itself with paper and pencil. If the participant had not quit voluntarily when 10 minutes
remained in the hour-long session, the experimenter stopped the participant. Several
participants who were told to stop were asked whether they understood that they could
have quit at any time; all of them indicated that they did. After the anagram task, all
participants completed manipulation check questions regarding their effort exertion on
the task, the difficulty of the task, and their desire to quit or keep working on the task.
At this point, the experimenter determined how long the participant had been in
the study. The informed consent had indicated the session would last one hour and I
wanted consistency across participants. That is, I did not want participants’ decisions
about how long to work on the anagrams to be influenced by whether they could finish
the study early. Thus, if more than 10 minutes remained in the hour-long session after
participants had quit voluntarily on the anagrams, the experimenter gave the
participants filler questionnaires to complete until the 10-minute mark and then went on
to the final steps of the study. If there were 10-minutes or less remaining after
participants stopped working on the anagrams, they went directly to the final steps of
the study.
For the final steps of the study, the experimenter gave the participant a
demographics questionnaire. Participants also completed a vocabulary test (Zachary,
1986). For each of 15 target words, the participant selected the closest synonym from a
group of four additional words. I used ACT scores and the vocabulary test to ensure that
any performance differences on the anagram task across conditions were not
attributable to differences in general cognitive ability or verbal ability. The participant

37

also completed a study questionnaire that assessed how the goal manipulation
influenced motivation and performance on the anagram task.
Results
Pilot Study
I ran a pilot study for two reasons: (1) to establish a performance goal for Study 3,
and (2) to assess whether the goal effect would occur when participants performed the
anagram task alone. Thirty-one participants each were assigned to a do-your-best
condition and a specific-difficult-goal condition (age M = 20.21, SD = 4.37; 74% female).
In these conditions, participants did not perform the typing task; they just performed the
anagram task. Table 3 displays the means and standard deviations for anagram
performance in the pilot study.
Using the number of anagrams that participants solved correctly in the pilot doyour-best condition, I computed the 85

th

percentile of performance, which was 50

anagrams. I subsequently used this number as the goal for the specific-difficult goal
condition in the pilot study and in Study 3. For the pilot conditions, independent sample
t-tests indicated a significant difference on number correct, t(60) = 2.16, p = .03, d =
0.56, but not on persistence t(60) = 1.06, p = .30, d = 0.27.
Study 3 Manipulation Checks
I performed independent sample t-tests on the manipulation check questions to
assess how participants rated the different tasks and how they reacted to the goal
manipulation. Table 4 displays means, standard deviations, and inferential statistics for
the manipulation check questions. I did not expect differences on manipulation check
questions for the typing task because the experimental procedure was the same for

38

both conditions. However, participants in the specific-difficult-goal condition reported
that they invested more of their total effort on the typing task than participants in the doyour-best condition, t(64) = 2.20, p < .05, d = 0.55. Participants rated task difficulty, how
hard they worked, and how much effort they had to give similarly (ps > .37).
Regarding the anagram task, I expected participants in the specific-difficult-goal
condition to rate their effort and work ethic higher than participants in the do-your-best
condition. Accordingly, participants in the specific-difficult-goal condition self-reported
that they invested more of their total effort than participants in the do-your-best condition,
t(65) = 2.25, p = .03, d = 0.56. Otherwise, participants did not differ on ratings of task
difficulty, how hard they worked, how much effort they had to give, or their desire to quit
or keep working (ps > .19).
I also predicted that participants in the specific-difficult-goal condition would rate
themselves as more motivated by the instructions. As expected, these participants selfreported that the goal motivated them to increase their effort, t(66) = 2.55, p = .01, d =
0.63, and persist longer on the anagram task, t(67) = 2.60, p = .01, d = 0.63, than
participants in the do-your-best condition. Self-reported differences in whether the
anagram instructions motivated participants to perform better did not reach a
conventional significance level, t(67) = 1.76, p = .08, d = 0.43.
There were no statistically significant differences across conditions on the
vocabulary measure or the BMIS pleasant-unpleasant scale (ps > .26). However, there
was a significant difference on the BMIS arousal-calm scale, t(68) = 2.31, p = .02, d = 0.56, and a marginal difference on ACT scores, t(60) = 1.71, p = .09, d = -0.44.

39

Participants in the do-your-best condition rated themselves as more aroused and
reported higher ACT scores than participants in the specific-difficult-goal condition.
Study 3 Anagram Persistence and Number Correct
Table 3 also displays the means and standard deviations for anagram
performance in Study 3. I performed independent sample t-tests on persistence time
and number correct in order to determine if the goal instructions affected performance
on the anagram task. As predicted, participants in the specific-difficult-goal condition
solved more anagrams correctly than participants in the do-your-best condition, t(68) =
2.27, p = .03, d = 0.55. Although the direction of the effect remained the same, the
difference between conditions did not reach conventional significance for persistence
time, t(68) = 1.81, p = .08, d = 0.44. When I included the BMIS arousal-calm scale and
ACT scores as covariates, the results for number correct did not change (p = .03); the
results for persistence time changed slightly with the p-value reaching .058. Some
participants did not report their ACT scores so these analyses were performed on a
subset of the whole sample.
I also computed effect sizes to compare performance in the Study 3 conditions to
performance in the pilot conditions. Participants in the Study 3 do-your-best condition
persisted less (d = -0.91) and solved fewer anagrams correctly (d = -0.51) than
participants in the pilot do-your-best condition. Participants in the Study 3 specificdifficult-goal condition persisted less (d = -0.80) and solved fewer anagrams correctly (d
= -.52) than participants in the pilot specific-difficult-goal condition. Notably, these
decreases were comparable between the two different instruction types.

40

Table 3: Descriptive Statistics for the Pilot Study and Study 3 Anagram Performance

Variable
Persistence (in minutes)
Correct Solutions
a

Pilot Study
Do-Your-Best
Specific-Difficult
Condition
Goal Condition
M (SD)
M (SD)
34.03 (12.11)
37.31 (12.36)
a
39.77 (13.77)
46.94 (12.25)

Indicates a significant difference for a variable within a study (p < .05).

41

Study 3
Do-Your-Best
Specific-Difficult
Condition
Goal Condition
M (SD)
M (SD)
23.87 (10.69)
28.41 (10.30)
a
31.75 (17.48)
40.29 (13.63)

Table 4: Descriptive and Inferential Statistics for Study 3 Questionnaires
Do-Your-Best
Condition
95 % CI
Variable
Typing Task Ratings
Effort Invested
Task Difficulty
How hard did you
work?
How much effort did
you give?
BMIS
Pleasant-Unpleasant
Scale
Arousal-Calm Scale
Anagram Task Ratings
Effort Invested

Specific-Difficult Goal
Condition
95 % CI
p

Cohen’s
d

2.19
-0.74
-0.88

.03
.46
.38

0.53
-0.18
-0.21

88.50

-0.11

.92

-0.03

3.11

6.65

0.66

.51

-0.16

3.60

13.26

15.68

2.31

.02

-0.56

11.51

87.82

95.80

2.25

.03

0.56
-0.32

M

SD

LL

UL

M

SD

LL

UL

83.81
72.06
76.25

21.26
19.11
18.06

76.87
65.82
70.35

90.75
78.30
82.15

92.35
68.00
71.53

8.28
26.31
26.08

89.57
59.16
62.76

95.13
76.84
80.30

83.61

18.54

77.55

89.67

83.18

15.84

77.86

5.86

6.98

3.58

8.14

4.88

5.28

16.61

4.10

15.27

17.95

14.47

82.09

21.86

74.85

89.33

91.81

t(68)

a

b
b

Task Difficulty

82.97

15.74

77.76

88.18

77.31

19.80

70.45

How hard did you
work?
How much effort did
you give?
How much did you want
to stop working?
Did you keep working
after you thought about
quitting?

83.86

15.43

78.75

88.97

86.09

12.10

81.90

84.17 -1.30 .20
90.28 0.66b .51

83.37

20.74

76.50

90.24

87.28

12.75

82.86

91.70

.36

0.23

68.71

28.58

59.24

78.18

61.28

36.71

48.56

74.00 -0.93b .36

-0.23

75.80

29.16

66.14

85.46

78.25

26.67

69.01

87.49

42

0.92

0.36

b

b

.72

0.16

0.09

Table 4 (cont’d)
How long did you think
about quitting before
you did?
How much did you
force yourself to keep
working?
Influence of Anagram
Instructions
Did goal motivate you
to perform better?
Did goal increase your
effort?
Did goal influence you
to persist longer?
ACT

55.11

28.94

45.52

64.70

53.84

33.23

42.33

65.35 -0.17b .87

67.37

26.82

58.48

76.26

74.06

25.25

65.31

82.81

1.05

58.34

28.81

48.80

67.88

70.82

30.17

60.68

80.96

1.76

59.03

30.65

48.73

69.33

76.09

24.21

67.95

84.23

2.55

63.66

31.90

53.09

74.23

80.76

21.67

73.48

88.04

2.60

25.32

2.98

24.27

26.37

23.87

3.68

22.57

25.17

1.71
1.11

-0.04

b

.30

0.26

c

.08

0.43

d

.01

0.63

c

.01

0.63

e

.09

-0.44

Vocabulary
9.78
2.82
8.86 10.70
9.12
2.09
8.42
9.82
.27
-0.27
Note. CI = Confidence Interval; LL = lower limit; UL = upper limit.
a
The statistics reported here are the ts associated with the test when equal variances were assumed. Although there
were cases in which Levene’s test was significant, in no case was the t associated with equal variances assumed
b
significant when the t associated with equal variances not assumed was not significant. These t-tests had 65 df because
c

some participants did not receive all of the manipulation check questions. These t-tests had 67 df because some
participants did not answer all of the questions.
questions.

e

d

This t-test had 66 df because some participants did not answer all of the

This t-test had 60 df because some participants did not report their ACT scores.

43

Discussion
The manipulation check items did not completely follow predicted patterns. Some
patterns suggested that participants in the specific-difficult-goal condition were more
depleted following the typing task and that participants in the do-your-best condition
were more aroused going into the anagram task and potentially had greater cognitive
abilities. If anything, this should have given participants in the do-your-best condition an
advantage on the anagram task. Nevertheless, the results indicated these differences
were not important; instead, the motivation manipulation had the greatest influence on
anagram performance. I expected differences on self-report questions on the anagram
task, but few differences were significant. However, participants in the specific-difficultgoal condition did indicate that they put in more total effort than participants in the doyour-best condition.
Overall, the results of Study 3 support a motivation account of the depletion
effect. Specifically, a motivation manipulation shifted performance among participants
objectively depleted to the same degree. This indicates that performing an effortful
regulatory task did not entirely hinder the ability to perform on the second task. Rather,
when participants were more motivated, they solved more anagrams correctly. The
effect of the motivation manipulation did not extend to persistence time. Given the
nature of the specific, difficult goal, which focused on solving a certain number of
anagrams, it makes sense that the largest effect was on number correct.
The comparison between Study 3 conditions and the pilot conditions also support
a motivation account of task performance. With no prior task, participants with a specific,
difficult goal solved more anagrams correctly than participants simply told to do their

44

best (d = .56). In other words, the goal motivated participants to exert more effort on the
task. In Study 3, this same pattern emerged even when participants had just completed
an effortful regulatory task (d = .55). Furthermore, the size of the goal effect within the
pilot conditions and within the Study 3 conditions was similar in magnitude. If true
depletion was a factor in Study 3 conditions, the goal effect should have been smaller
because motivation was competing against decrements in ability.
When compared to the pilot conditions, anagram performance decreased after
an effortful regulatory task, but the size of the decrease was comparable for participants
in do-your-best or specific-difficult-goal conditions. This suggests that the motivation to
exert effort generally declined after a prior task, but the specific, difficult goal sufficiently
motivated participants to exert more effort whether there was a prior task or not. In other
words, people are miserly when it comes to effort exertion, but that miserliness reflects
motivation and not capability.

45

CHAPTER 5: STUDY 4- INFLUENCE OF PERCEIVED EFFORT
Study 4 Rationale and Predictions
In Study 3, I investigated whether a specific, difficult goal would increase
participants’ effort exertion, despite equivalent levels of depletion with other participants.
In Study 4, I investigated whether greater perceived effort exertion on the first task
would decrease participants’ effort exertion, despite equivalent levels of depletion with
other participants. I manipulated perceived effort using a design similar to that of Kivetz
and Zheng (2006).
Kivetz and Zheng (2006) showed that manipulating perceived effort on an initial
task influenced subsequent choices. Their first task involved unscrambling sentences.
Participants were told that they would be randomly assigned to a particular level of effort;
participants in the low-relative-effort condition were told that the number of sentences
they would construct could be between 14 and 28 and participants in the high-relativeeffort condition were told the number could be between 4 and 18. All participants were
told they had been assigned to unscramble 16 sentences. Thus, in the low-relativeeffort condition, participants believed they had to exert less effort relative to other
participants and in the high-relative-effort condition, participants believed they had to
exert more effort relative to other participants.
Kivetz and Zheng (2006) predicted that those in the high-relative-effort condition
would be less motivated to continue exerting effort subsequently because they would
think they had already exerted enough effort while the opposite would be true for those
in the low-relative-effort condition. Indeed, even though all participants unscrambled the
same number of sentences, participants in the high-relative-effort condition

46

subsequently chose what Kivetz and Zheng deemed “self-indulgent” rewards more often
than participants in the low-relative-effort condition. They argued that even when the
first task was the same across conditions, perceptions of effort on that task influenced
participants to exert more or less effort subsequently.
In Study 4, I had all participants perform the same initial effortful regulatory task.
Afterward, I gave participants false feedback that suggested they either exerted
relatively low or high effort on the first task. Then, I had all participants perform a
second effortful regulatory task. According to a resource account, the false feedback
should not affect performance because of equivalent depletion across participants.
According to a motivation account, participants in the high-perceived-effort condition
should be less motivated to continue exerting effort and should perform significantly
worse on the second task than those in the low-perceived-effort condition.
Method
Participants
Seventy undergraduates from the subject pool at Michigan State University
participated in this study for partial course credit. After the study, the experimenter
debriefed the participants about the nature of the perceived-effort deception. The
experimenter also asked the participants whether they believed the effort feedback at
the time that it was given. Participants’ responses were coded as either believing the
feedback, partially believing, or not believing. In the analyses for this study, I included

47

5

only those participants who fully believed the report. This left a sample of 47 (lowperceived-effort condition, n = 21; high-perceived-effort condition, n = 26). The final
sample had a mean age of 19.91 (SD = 2.39) and was 81% female. An a priori power
analysis indicated a sample of 70 would have 80% power (alpha = .05, two-tail) to reject
the null hypothesis that there would be no difference in performance across conditions.
A post hoc power analysis indicated the final sample of 47 had 73% power (alpha = .05,
two-tail) to reject the null hypothesis that there would be no difference in performance
across conditions.
Procedure and Materials
Participants were randomly assigned to one of two conditions: a low-perceivedeffort condition or a high-perceived-effort condition. Participants completed the study
individually in seven steps: (a) a typing task and manipulation check questions, (b) the
BMIS, (c) an effort feedback segment, (d) an anagram task and manipulation check
questions, (e) filler tasks (this part was conditional, just like in Study 3), (f) a
demographics questionnaire, (g) a vocabulary test, and (h) a study questionnaire.
Participants were informed the session would last one hour.
During a session, the experimenter seated the participant in front of a computer.
The experimenter calibrated an eye tracker (an Eyelink II at 500 Hz) to the participant’s
eye movements and informed the participant that the eye tracker would be monitoring
5

I decided to only include those who believed the effort feedback in the analyses to
ensure that the results would be the cleanest test of my hypotheses. It is possible that
disbelieving the feedback could have interacted with the motivation to continue to exert
effort. Results on the main dependent variables did not differ if I included all participants
in the analyses. However, the difference on one manipulation check question, “Did your
effort report on the typing task influence you to decrease your effort on the anagram
task?” was not significantly different with the full sample, but was with the reduced
sample.
48

pupil diameter throughout the task. Then the experimenter read aloud the instructions
for the typing task. The typing task and instructions were the same as those used in
Study 3, with an added section instructing the participant not to look down while typing,
as this interrupted the eye tracking.
After 6 minutes had elapsed, the typing task ended and the experimenter stated,
“Here are two questionnaires for you to fill out. While you are filling them out, I will be
putting together a report regarding your pupil diameter during the typing task. The
computers are networked so I don’t have to get on your computer to do this.” The
experimenter handed the participant a sheet with several manipulation check questions
about the typing task and the BMIS. Then, the experimenter went just outside of the
study room to another computer that was networked with a printer. The experimenter
logged on to the computer, opened a document of the pre-set effort report, and printed
6

two copies of it. The report was set up to print the current date and time. These
procedures were implemented to lend credibility to the effort report. After the copies
were printed, the experimenter returned to the study room, gave the participant one
copy, and kept the other, telling the participant to follow along as the experimenter read
the report.
In both conditions, the report began by summarizing scientific research that used
pupil diameter as a measure of effort on behavioral tasks. The report then indicated that
6

During Study 4, I had technical difficulties with the printer. For 17 participants, it did
not connect to the network and failed to print the report. However, the printer made
printing-like sounds for 14 of these participants. I had extra copies of the report printed
and used these when the printer failed. To ensure that this did not affect participants’
belief of the pupil report, I performed an independent t-test with report belief as the
dependent variable and whether the printer worked or not as the grouping variable. The
result was not significant (p > .40).
49

the participant’s pupil data had been analyzed as an indicator of his or her effort
exertion on the typing task. The report showed a graph with two lines plotted, one
representing the participant’s data and one representing other participants’ data for
comparison. The report had a written summary of the participant’s data in comparison to
others’. The graph and summary were false and intended to deceive participants into
thinking they exerted more or less effort relative to other participants. Specifically, in the
high-perceived-effort condition, the graph showed that the participant’s pupil diameter
was higher than the other participants’ pupil diameter and the summary stated, “The
pupil data show that you exerted more effort than the average participant on the typing
task. Specifically, you exerted more effort than 81.5% of the other participants” (see
Appendix G for a complete copy of the high-effort pupil report). In the low-perceivedeffort condition, the graph showed that the participant’s pupil diameter was lower than
the other participants’ pupil diameter and the summary stated, ”The pupil data show that
you exerted less effort than the average participant on the typing task. Specifically,
81.5% of the other participants exerted more effort than you” (see Appendix H for a
complete copy of the low-effort pupil report).
Participants in both conditions then completed the anagram task. In Study 2, 27
of the 48 items on the anagram task were unsolvable. In Study 4, I adapted this list by
replacing the unsolvable anagrams with solvable ones. To replace two of the items, I
removed one extra letter that had made the anagram unsolvable (e.g., OERIRFPM
became OERRFPM: PERFORM). To replace the remaining unsolvable items, I used
the Medical Research Council Psycholinguistics database to generate a list of sevenletter nouns that had concreteness and familiarity ratings of 400 or higher (these ratings

50

range from 100 to 700; Wilson, 1998). I randomized this list and selected 25 anagrams
to replace the unsolvable anagrams from the original list (see Appendix I). The basic
instructions and computer methodology were the same as those used in Study 3.
As in Study 3, the experimenter stopped the participant on the anagrams if he or
she was still working on them 10-minutes before the end of the hour-long session. The
experimenter asked several of these participants whether they understood that they
could quit at any time; all of them indicated that they did. Otherwise, the experimenter
followed the same procedure involving filler tasks to ensure that participants were in the
study session for approximately one hour.
After the anagram task, the participant answered manipulation check questions
regarding the task. The final steps of the study were completed after the filler tasks (if
they were administered). The participant completed a demographics questionnaire and
the same vocabulary test used in Study 3. The participant also completed a study
questionnaire that assessed how the pupil data influenced their effort exertion on the
anagrams and whether they thought the pupil report was an accurate reflection of their
exerted effort.
Results
Manipulation Checks
I performed independent sample t-tests on the manipulation check questions to
assess how the participants rated themselves on different aspects for each task and
how they reacted to the effort feedback. See Table 5 for means, standard deviations,
and inferential statistics for the manipulation check questions. As predicted, there were
no differences in task difficulty, hard work, or effort ratings for the typing task (ps > .77).

51

Unexpectedly, there were also no significant differences on any self-report questions for
the anagram task (ps > .12). Participants rated task difficulty and their effort, work, and
desire to do the task similarly.
I also examined participants’ self-reported reaction to the pupil diameter
feedback. I predicted that participants in both conditions would indicate that the effort
reading from the pupil data was accurate and similar to their subjective feeling of effort.
However, compared to the high-perceived-effort participants, participants in the lowperceived-effort condition felt that the pupil diameter feedback was not similar to their
subjective feeling of effort, t(44) = -3.79, p < .001, d = -1.15, and was not an accurate
reflection of their effort, t(45) = -5.06, p < .001, d = -1.52.
However, if the pupil diameter feedback successfully manipulated participants’
motivation to exert effort, participants in the high-perceived-effort condition should have
been less motivated to perform better on the anagram task. To assess this, participants
were asked whether the pupil diameter feedback motivated them to increase their effort
on the anagram task and whether the pupil diameter feedback influenced them to
decrease their effort. While the difference between conditions was not significant on the
first question (p > .37), the difference was significant on the second question, t(45) = 2.04, p = .048, d = -0.61. Participants in the high-perceived-effort condition indicated
that the pupil diameter feedback influenced them to decrease their effort on the
anagram task.
There were no statistically significant differences across conditions on the mood
scales, ACT scores, or performance on the vocabulary measure (ps > .41). Thus, mood
or verbal ability did not differentially affect performance on the anagram task.

52

Anagram Persistence and Number Correct
I performed independent sample t-tests on persistence time and number correct
to determine if the perceived-effort manipulation affected performance on the anagram
task. Table 6 displays means, standard deviations, and inferential statistics for anagram
performance. As predicted, participants in the high-perceived-effort condition solved
fewer anagrams correctly than participants in the low-perceived-effort condition, t(45) =
2.57, p = .01, d = 0.77. Additionally, high-perceived-effort participants persisted less
than low-perceived-effort participants, t(45) = 2.95, p = .005, d = 0.89.
In a separate pilot study with 53 participants, individuals with no prior task or
feedback persisted on average 28.80 minutes (SD = 13.81) and solved on average 9.72
anagrams (SD = 5.05). Compared to this pilot study, performing a prior task decreased
the average amount of time that participants decided to persist on the anagrams, but
particularly so for participants in the high-perceived-effort condition (d = -1.31 versus d =
-0.58 for participants in the low-perceived-effort condition). In terms of number correct,
those in the high-perceived-effort condition solved fewer anagrams (d = -0.86) than
those in the pilot study, whereas participants in the low-perceived-effort condition solved
a comparable number to those in pilot testing (d = -0.19).

53

Table 5: Descriptive and Inferential Statistics for Study 4 Questionnaires
High-Perceived-Effort
Condition
95 % CI
Variable
Typing Task Ratings
Effort Invested
Task Difficulty
How hard did you
work?
How much effort did
you give?
BMIS
Pleasant-Unpleasant
Scale
Arousal-Calm Scale
Anagram Task Ratings
Effort Invested
Task Difficulty
How hard did you
work?
How much effort did
you give?
How much did you
want to stop
working?
Did you keep working
after you thought
about quitting?

Low-Perceived-Effort
Condition
95 % CI
t(45)

a

M

SD

LL

UL

M

SD

LL

UL

90.00
84.46
86.42

18.62
18.42
17.90

82.84
77.38
79.54

97.16
91.54
93.30

90.19
85.81
86.38

15.08
12.24
13.29

83.74
80.57
80.70

96.64
91.05
92.06

0.04
0.29
0.01

.97
.78
.99

0.01
0.09
0.00

90.54

17.92

83.65

97.43

91.14

11.55

86.20

96.08

0.13

.89

0.04

7.27

7.89

4.24

10.30

5.76

6.24

3.09

8.43

-0.71

.48

-0.21

16.12

4.10

14.54

17.70

15.52

3.52

14.01

17.03

-0.52

.60

-0.16

77.62
91.00
92.08

21.69
11.05
11.03

69.28
86.75
87.84

85.96
95.25
96.32

86.43
92.00
86.57

16.33
9.58
17.35

79.45
87.90
79.15

93.41
96.10
93.99

1.54
0.33
-1.32

.13
.75
.19

0.46
0.10
-0.40

87.31

20.98

79.25

95.37

88.57

13.44

82.82

94.32

0.24

.81

0.07

70.96

26.42

60.80

81.12

74.57

26.83

63.09

86.05

0.46

.65

0.14

82.46

24.11

73.19

91.73

82.43

24.31

72.03

92.83

0.01

.99

0.00

54

p

Cohen’s
d

Table 5 (cont’d)
How long did you
think about quitting
before you did?
How much did you
force yourself to keep
working?
Influence of Pupil
Diameter Feedback on
Anagram Task
How similar was
feedback report and
subjective feeling of
effort?
How accurate was
the feedback?
Did feedback
increase your effort?
Did feedback
decrease your effort?
ACT

69.69

25.61

59.85

79.53

73.81

32.12

60.07

87.55

0.49

.63

0.15

73.62

24.65

64.14

83.10

70.52

28.94

58.14

82.90

-0.40

.69

-0.12

80.92

22.96

71.92

89.92

45.86

39.01

29.18

62.54 -3.79b

<.001

-1.15

89.79

15.31

83.91

95.67

46.57

40.18

29.38

63.76

-5.06

<.001

-1.52

56.54

31.27

44.52

68.56

64.52

29.32

51.98

77.06

0.90

.38

0.27

28.15

31.21

16.15

40.15

12.52

17.92

4.86

20.18

-2.04

.048

-0.61

24.83

3.27

23.52

26.14

25.06

3.44

23.42

26.70

0.21
0.82

.83

0.07

c

Vocabulary
9.23
2.69
8.20 10.26
9.86
2.50
8.79 10.93
.42
0.25
Note. CI = Confidence Interval; LL = lower limit; UL = upper limit.
a
The statistics reported here are the ts associated with the test when equal variances were assumed. Although there
were cases in which Levene’s test was significant, in no case was the t associated with equal variances assumed
b
significant when the t associated with equal variances not assumed was not significant. This t-test had 44 df because
c

some participants did not answer all of the questions. This t-test had 39 df because some participants did not report their
ACT scores.

55

Table 6: Descriptive and Inferential Statistics for Study 4 Anagram Performance
High-Perceived-Effort
Condition
95 % CI
Variable
Persistence
(in minutes)

Low-Perceived-Effort
Condition
95 % CI

M

SD

LL

UL

M

SD

LL

UL

13.97

7.16

11.22

16.72

21.57

10.43

17.11

26.03

t(45)

a

2.95

p

Cohen’s
d

.005

0.89

Correct Solutions
6.04
3.13
4.84
7.24
8.76
4.15
6.99
10.53
2.57
.01
0.77
Note. CI = Confidence Interval; LL = lower limit; UL = upper limit.
a
The statistics reported here are the ts associated with the test when equal variances were assumed. Although there
were cases in which Levene’s test was significant, in no case was the t associated with equal variances assumed
significant when the t associated with equal variances not assumed was not significant.

56

Discussion
Manipulation check questions showed no differences between conditions on the
typing task. This indicated that participants in both conditions viewed the initial task
similarly and put forth the same amount of effort. Participants were depleted to the
same degree before they received the pupil diameter feedback and completed the
second task.
Regarding the pupil feedback, some participants did not believe the report at the
time it was given to them. While initially surprising, this does make sense when looking
more closely at the procedure. Participants rated task difficulty, their effort, and how
hard they worked on the typing task and then received the pupil diameter feedback. As
the typing task was quite challenging, most participants reported that they put forth
significant effort and worked hard on the task. Therefore, when the experimenter told
low-perceived-effort participants that they did not put forth significant effort, this
feedback contradicted the self-report ratings they had just given.
There were no differences on manipulation check questions for the anagram task
although I had expected the ratings to coincide with actual performance. However, it is
possible that the pupil report changed participants’ perception of effort on the anagram
task. For example, participants told their effort was above average on the first task may
have inflated the anagram task ratings subsequently.
The anagram performance results suggest that participants act as motivated
tacticians. Participants told they exerted more effort initially withheld effort on the
anagram task and performed worse than participants told they exerted less effort initially.
This coincides with the self-report data where high-perceived-effort participants

57

indicated the pupil diameter feedback influenced them to decrease their effort. These
results also confirmed the meta-analysis. Study 4 showed that differences in perceived
effort led to a depletion-like effect. The meta-analysis showed that differences in selfreported effort related to the size of the depletion effect.
Furthermore, the low-perceived-effort participants performed similarly to the pilot
participants who had not had a prior task. Even though participants in the lowperceived-effort condition may have persisted less, they still solved about the same
number of anagrams. If performing the anagram task in a depleted state truly
incapacitated regulatory abilities, these participants should not have been able to
perform as accurately on the anagram task as participants in the pilot study.
These results, together with the results of the previous three studies, suggest
that the depletion effect and regulatory performance are often dictated by motivation,
rather than the availability of a regulatory resource. If individuals have already exerted
significant effort, they are less motivated to continue exerting effort. If initial effort
exertion is the same among participants, performance shifts up or down depending on
the willingness to continue exerting effort. An individuals’ willingness to exert effort
reflects many factors, including goals and perceptions of previous efforts.

58

CHAPTER 6: STUDY 5- CONSERVATION OF RESOURCE VS. EFFORT
Study 5 Rationale and Predictions
Earlier, I discussed research by Muraven et al. (2006) that suggested the
depletion effect partly results from resource conservation. Resource conservation is an
alternative explanation for Studies 3 and 4. That is, while the first task depleted
participants’ resource to the same extent, the motivation manipulations may have
encouraged some participants (e.g., in the low-perceived-effort condition) to continue
using what resources remained, while encouraging other participants (e.g., in the highperceived-effort condition) to conserve their remaining resources.
Muraven et al. (2006) investigated the concept of conservation across several
studies. In some of them, they used the typical two-ask procedure, but manipulated
instructions on the second task to make some participants believe there would be a
challenging self-regulatory third task. Other participants were led to believe there would
be a challenging non-regulatory third task or remained ignorant of a third task.
Participants who believed there would be a regulatory third task performed worse on the
second task than participants who thought it would be non-regulatory or did not know.
Muraven et al. argued that these results supported the resource account’s assertion that
the depletion effect, and now conservation, depend on whether or not self-regulation is
involved in prior or future efforts.
In their final study, Muraven et al. (2006) used a three-task procedure in order to
assess how conserving or not conserving affected performance when participants
actually completed the third task. After a depleting initial task, participants who
anticipated a challenging self-regulatory third task exhibited greater depletion effects on

59

a second task than those who anticipated a challenging non-regulatory third task. In
other words, when participants expected the need for future regulation, they were more
likely to conserve. Muraven et al. also found a trade-off between performance on the
second and third task. If depleted participants conserved on the second task (and
performed worse), they performed better on the third task, and vice versa. According to
the resource account, depleted participants who conserved on the second task still had
enough resources to perform the third task, whereas depleted participants who did not
conserve on the second task did not have sufficient resources to perform the third task.
Muraven et al.’s (2006) studies lend some support for the regulatory-specific
nature of the depletion and conservation effect. However, they did not address whether
the conservation effect is truly the conservation of a resource or whether it is simply
motivated conservation of effort. Even though the conservation concept introduces a
motivation element to the resource account, the underlying assumption remains the
same: after prior regulatory exertion, individuals operate with reduced resources. Thus,
the ability to regulate becomes increasingly limited for those who continue to expend
regulatory resources (e.g., like depleted participants who did not conserve in Muraven
et al.). If, however, the effects are motivational in nature, there should be no limits on
performance for depleted individuals who do not conserve.
I designed Study 5 to investigate whether it is more likely that individuals
conserve a limited, expendable resource or whether individuals simply conserve effort.
Although it was not a direct replication, I used the same three-task procedure as
Muraven et al.’s Study 4 (2006), where participants performed a typing task, a Stroop
task, and then an anagram task. On the first task, all participants received the same

60

instructions. When explaining the second task, half of the participants were told to
expect a challenging self-regulatory third task and half of the participants did not expect
a third task at all. Then, before the third task, there was a motivation manipulation
where half of the participants received a specific, difficult goal and half of the
participants received do-your-best instructions.
According to the resource account, participants who did not know about the third
task (i.e., low-conservation condition) should not hold back effort on the second task.
Thus, with doubly depleted resources, they should unavoidably perform worse on the
third task. No amount of motivation should enable these participants to perform at the
same level as those who did conserve on the second task (i.e., high-conservation
condition). If, however, no real reduction in resource availability occurs, participants who
did not conserve should still be capable of exerting a high level of effort and of
performing just as well or better on the third task than those who did conserve, as long
as they are sufficiently motivated to do so (e.g., when given a specific, difficult goal).
Another key to this design is that I used the same list of anagrams from Study 3.
This allowed me to examine performance on this task not only in the context of
conservation in this study, but also in the context of when the task was performed after
just one task and when it was performed with nothing prior. That is, this provided a way
to compare performance on the anagram task when participants had do-your-best
instructions or a specific, difficult goal and were presumably at full capacity, depleted, or
doubly depleted. I know from Studies 3 and 4 that performance declined to a certain
degree when comparing participants who performed the anagram task by itself to
participants who performed the anagram task after a regulatory task. The results from

61

this study allowed me to assess whether performance continued to decline for doubly
depleted participants.
Method
Participants
Participants were 118 undergraduate students from the subject pool at Michigan
State University who received partial course credit for participation. Six participants
were removed from all analyses. Four of these participants did not understand or follow
instructions on the Stroop task. One participant did not start typing on the first task as
directed and only typed four out of the six minutes. One participant’s computer
malfunctioned, disrupting the study procedure and data recording. This left a sample of
112 (high-conservation/do-your-best goal condition, n = 29; high-conservation/specificdifficult goal condition, n = 29; low-conservation/do-your-best condition, n = 26; lowconservation/specific-difficult goal condition, n = 28). The final sample had a mean age
of 20.36 (SD = 3.18) and was 51% female.
Some participants were not included in all analyses. Four participants were
removed from analyses relating to the anagram task because they did not follow the
instructions or the computer malfunctioned. Two participants were removed from the
anagram persistence analysis because the computer timing was inaccurate. Additionally,
there were some cases where participants did not answer all of the manipulation check
questions.
Procedure and Materials
Participants were randomly assigned to one of four conditions: a highconservation/do-your-best condition, a high-conservation/specific-difficult-goal condition,

62

a low-conservation/do-your-best condition, or a low-conservation/specific-difficult-goal
condition (see Table 7 for a layout of the design). Participants completed the study
individually in seven steps: (a) a timed typing task and manipulation check questions, (b)
a conservation manipulation segment, (c) a Stroop task and manipulation check
questions, (d) an anagram task and manipulation check questions, (e) a demographics
questionnaire, (f) a vocabulary test, and (g) a study questionnaire. Participants were
informed the session would last one hour.
Table 7: Design of Study 5
Condition
1st Task
High-Conservation/ Typing Task
Do-Your-Best
High-Conservation/ Typing Task
Specific-Difficult
Low-Conservation/ Typing Task
Do-Your-Best
Low-Conservation/ Typing Task
Specific-Difficult

2nd Task
Stroop Task
w/ Expectation
of Task 3
Stroop Task
w/ Expectation
of Task 3
Stroop Task
w/ No Expectation
of Task 3
Stroop Task
w/ No Expectation
of Task 3

3rd Task
Anagram Task
w/ Do-Your-Best
Instructions
Anagram Task
w/ Specific, Difficult
Goal
Anagram Task
w/ Do-Your-Best
Instructions
Anagram Task
w/ Specific, Difficult
Goal

Participants were seated in front of a computer. Unlike the previous studies, I
designed this study so that participants guided themselves through the majority of the
tasks by reading instructions on the computer. I made this change in order to improve
the design of the study and eliminate the possible effects of experimenter bias. Before
getting started, the experimenter explained to the participant that the computer would
guide them through the study and told the participant to read and follow all of the
instructions carefully. The experimenter encouraged the participant to contact the
experimenter with any questions not answered by the computer instructions.
63

Additionally, the experimenter told the participant that the computer instructions would
direct him or her to contact the experimenter at certain points to receive further
instructions.
After running several participants with this format, I found that more participant
errors occurred. For example, some participants used the wrong keys to respond during
the Stroop task or pressed certain keys during the anagram task that interrupted
persistence timing. To reduce participant error, I added more instructions at the
beginning of the study, such as pointing out certain keys the participant would use. After
these instructions, the experimenter left the room, but listened to make sure the
participant got started on the first task.
The first task was the same typing task used in Studies 3 and 4. The basic
instructions were the same, with some instructions added to facilitate the new
participant-guided procedure. Specifically, the instructions gave the participant an
example of what the typing should look like (e.g., If they saw, “The big red tree.” they
should type, “Thbigrdtr.”) and described different types of errors (e.g., typing spaces or
E’s, skipping over other characters, or typing other characters not in the paragraph).
After 6 minutes elapsed, the typing task ended and the participant answered
manipulation check questions regarding their effort and the difficulty of the task. The
participant entered their ratings for the questions using numbers on the keyboard and
pressing enter.
The computer then presented instructions for the Stroop task:
In the color-word identification task, you will see a color-word appear in the middle
of the screen and will need to correctly identify the COLOR of ink that the word is

64

written in. Your response will either be blue, green, or red. These responses
correspond, respectively, to the 1, 2, and 3 keys on the number pad of the
keyboard. They have been labeled B for blue, G for green, and R for red for your
convenience. For example, if you saw the following: blue (the word was printed in
red), the correct answer would be, “Red,” so you would press the key labeled “R”
on the number pad. At the beginning of each trial, you will see a fixation cross,
followed by a word. Once you make your response, another fixation cross will
appear, followed by another word, and so on. Research has shown that color-word
identification is challenging because it requires you to suppress or override the
automatic tendency to read the color-word in order to identify the color of ink the
word is printed in.
For participants in the low-conservation conditions, these instructions were
prefaced with, “You will complete a color-word identification task next.” For participants
in the high-conservation conditions, these instructions were prefaced with, “There are
two more tasks to complete. You will complete a color-word identification task next,
followed by an anagram task,” and were followed with these additional instructions,
“After performing the color-word identification task, you will complete an anagram task.
This task will also require your ability to suppress or override thoughts and intentions. In
order to correctly solve the anagrams, you should expect to work hard at overriding
impulses while working on them.” Thus, participants in the low-conservation conditions
did not know about a third task, whereas participants in the high-conservation conditions
knew about a third task and knew that it would be challenging in the same way that the
Stroop task was challenging. I adapted these instructions from Muraven et al.’s fourth

65

study (2006). They were designed to influence high-conservation participants, relative to
low-conservation participants, to withhold effort on the Stroop task because they knew
another challenging regulatory task would immediately follow.
Following these instructions, the participant completed the Stroop task, which
included 80 trials. Afterward, the participant answered manipulation check items
regarding the difficulty of the task, effort on the task, conservation on the task, and
anticipation of a future task. The participants also compared their effort on the Stroop
task to their effort on the typing task; responses ranged from 1 (a lot more effort on the
typing task than the Stroop task) to 5 (a lot more effort on the Stroop task than the
typing task). To assess participants’ comprehension of the conservation manipulation,
the participant chose out of five options what task they thought they would complete
next (1 = Don’t know, 2 = Color-word identification task, 3 = Anagram task, 4 = Typing
task, or 5 = Maze task). I anticipated that high-conservation participants would select
the anagram task most often, while the low-conservation participants would indicate
they did not know most often.
After the Stroop manipulation check questions, the computer provided
instructions for the anagram task. These instructions, the same as those used in
Studies 3 and 4, explained how to solve the anagrams and told the participant to press
enter when ready to begin and to press escape when done. Additionally, half of the
participants from the low- and high-conservation conditions were given do-your-best
instructions for the task and half were given a specific, difficult goal for the task (i.e., 50
anagrams). Then the computer directed the participant to contact the experimenter. The
experimenter handed the sheets of anagrams to the participant and reminded them to

66

press enter before starting. The experimenter then left the room while the participant
worked on the anagrams with paper and pencil. When the participant pressed enter,
there was a reminder on the screen about the participant’s condition-specific
instructions (i.e., complete a goal or do-your-best) and the quitting instructions (i.e.,
press escape when you feel you have worked on the task long enough). When the
participant decided to stop working on the anagrams and pressed escape, the computer
guided them through manipulation check questions regarding the difficulty of the task,
effort on the task, and how the option of quitting affected effort and the desire to quit.
Following that, the computer directed the participant to contact the experimenter.
The experimenter gave the participant the demographics questionnaire, the vocabulary
test, and a study questionnaire that assessed how the anagram instructions affected
effort and motivation on the task. The participants also compared their effort on the
anagram task to their effort on the Stroop task; responses ranged from 1 (a lot more
effort on the Stroop task than the anagram task) to 5 (a lot more effort on the anagram
task than the Stroop task).
After the participant completed these questionnaires, the experimenter assessed
how long the participant had been in the study. If there were more than 5 minutes
remaining in the hour-long session, the experimenter had the participant complete filler
questionnaires until the 5-minute mark and then the experimenter debriefed the
participant. If there were 5-minutes or less remaining, the experimenter went directly to
debriefing the participant. This methodology changed somewhat from Studies 3 and 4.
In these studies, participants completed fillers and then completed the last set of
questionnaires. In this study, participants completed the last set of questionnaires and

67

then did fillers. I changed this to keep the ratings on the final study questionnaire as
close to the final task as possible to reduce the possibility that participant ratings would
change by completing the filler tasks themselves or by the passage of time.
Results
Manipulation Checks
I performed 2 (Conservation: high, low) X 2 (Goal Type: do-your-best, specificdifficult) ANOVAs on several variables as manipulation checks. Table 8 displays means
and standard deviations for manipulation check questions administered following the
typing task. Since there were no manipulations on the first task, I predicted no
differences across conditions. Consistent with this prediction, none of the main effects
or interactions were significant for total effort exerted, task difficulty, hard work, effort
given, degree of self-control, or fighting an urge (ps > .11).

68

Table 8: Descriptive Statistics for Study 5 Manipulation Check Questions on the Typing Task

Variable
a

Effort Invested
Task Difficulty
How hard did you work?
How much effort did you give?
Self-Control
Fight Against an Urge
Note.

a

Do-Your-Best Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)
89.73
68.96
73.88
84.92
84.92
76.27

9.58
22.94
21.85
16.60
14.81
23.95

89.59
69.66
78.55
85.00
86.38
86.52

10.60
21.80
16.32
12.54
13.56
18.01

Specific-Difficult Goal Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)
86.04
65.11
71.57
74.89
83.71
78.21

17.27
22.83
23.01
25.22
12.71
26.93

Levene’s test of equality of error variances was significant (p < .05) for this question.

69

90.79
70.21
78.79
83.03
79.79
81.28

13.89
23.01
18.92
21.42
19.72
21.08

Table 9 displays means and standard deviations for manipulation check
questions administered following the Stroop task. I predicted that participants in the
high-conservation conditions would rate effort-type questions lower and rate selfcontrol- and conservation-type questions higher than participants in the lowconservation conditions. I did not predict a main effect of goal type or any interactions
because the goal manipulation had not taken place at the time participants answered
these questions.
However, there were main effects of goal type and conservation on the self2

control question, F(1, 106) = 5.63, p =.02, partial ! = .05, d = 0.46, and F(1, 106) =
2

3.94, p =.05, partial ! = .04, d = 0.39, respectively. Participants in the do-your-best
conditions indicated the task required more self-control than participants in the specificdifficult-goal conditions, and participants in the high-conservation conditions rated selfcontrol requirements higher than participants in the low-conservation conditions. The
Conservation x Goal Type interaction was significant for the item, “How difficult was the
2

Stroop task?” F(1, 106) = 5.00, p =.03, partial ! = .05. The pattern of means indicated
that for specific-difficult-goal conditions, participants in the high-conservation condition
rated the Stroop task as more difficult than those in the low-conservation condition (d =
0.32), whereas the opposite held for the do-your-best conditions (d = 0.33).
There was a marginally significant effect of conservation on how much
participants had to fight an urge on the Stroop task, such that participants in the highconservation conditions indicated they had to fight an urge more than participants in the
2

low-conservation conditions, F(1, 106) = 3.73, p =.056, partial ! = .03, d = 0.38. There

70

was also a marginally significant main effect of goal type for total effort exerted;
participants in the do-your-best conditions rated their total effort exertion higher than
participants in the specific-difficult-goal conditions, F(1, 107) = 3.67, p =.058, partial !

2

= .03, d = 0.37. Participants compared their effort on the Stroop task to their effort on
the typing task (ratings could range from 1 to 5, with 1 meaning more effort on the
typing task and 5 meaning more effort on the Stroop task). Participants generally
indicated they put more effort into the typing task, overall M = 1.79, SD = 1.07; there
were no significant differences across conditions (ps > .48). All other analyses on the
Stroop manipulation check questions (i.e., hard work, effort given, and conservation)
were also non-significant (ps > .10).
Following the Stroop task, participants indicated what task they thought they
would perform next. Participants were given five options, one being, “Don’t know,” and
another being, “Anagram task.” Of the participants in the low-conservation conditions,
who should not have known what task came next, 56% indicated they did not know
what the third task would be and 19% indicated they thought it would be an anagram
task. Of the participants in the high-conservation conditions, who had been told what
task came next, 85% indicated they thought it would be an anagram task. Thus,
participants appeared to comprehend the conservation manipulation instructions.

71

Table 9: Descriptive Statistics for Study 5 Manipulation Check Questions on the Stroop Task

Variable
Effort Invested
Task Difficulty

a

How hard did you work?
How much effort did you give?
Self-Control

Do-Your-Best Conditions
Low-Conservation High-Conservation
Condition
Condition
M
(SD)
M
(SD)
87.35
b
56.04
63.23
69.35
cd
71.96
54.73
1.69

Specific-Difficult Goal Conditions
Low-Conservation High-Conservation
Condition
Condition
M
(SD)
M
(SD)

18.85
28.33

89.24 15.80
42.34 24.57

78.25 25.02
42.61 28.58

81.11 31.70
52.78 30.28

28.67
28.45
27.81

54.21 26.93
69.52 25.69
74.69 23.06

47.96 31.18
56.79 32.94
52.36 30.39

57.78 30.99
67.74 31.85
69.96 25.78

Fight Against an Urge
33.18
62.62 27.29
46.11 32.10
60.81 30.07
Typing to Stroop Effort
0.97
1.66 1.01
1.89 1.17
1.74 1.13
Comparison
Conserving Energy
16.85
24.37
13.52 20.95
18.07 22.97
11.30 19.88
a
Thinking About Next Task
14.35
25.33
7.79 13.68
16.18 21.91
12.07 20.69
Note. Lower averages on the, “Typing to Stroop effort Comparison” mean participants rated their effort on the typing task
higher compared to their effort on the Stroop task.
a
b
Levene’s test of equality of error variances was significant (p < .05) for this question. Indicates a significant Goal Type
c

X Conservation interaction (p < .05). Indicates a significant main effect of goal type (p < .05).
main effect of conservation (p < .05).

72

d

Indicates a significant

Table 10 displays means and standard deviations for manipulation check
questions administered following the anagram task. I predicted that participants in the
specific-difficult-goal conditions would rate effort-type questions higher and rate desireto-quit-type questions lower than participants in the do-your-best conditions. I also
predicted that participants in the high-conservation conditions would rate effort-type
questions higher and rate desire-to-quit-type questions lower than participants in the
low-conservation conditions. The results indicated that goal type and conservation
interacted with several anagram self-report questions.
There was a significant interaction for the question, “What percentage of your
2

total effort did you invest in the anagram task?” F(1, 102) = 4.78, p =.03, partial ! = .05.
There was a near significant interaction for the question, “How hard did you have to
2

work at the anagram task?” F(1, 102) = 3.28, p =.073, partial ! = .03. The pattern of
means indicated that for the specific-difficult-goal conditions, participants in the highconservation condition rated their effort/work higher than participants in the lowconservation condition (d = 0.60/0.34), whereas the opposite held for the do-your-best
conditions (d = 0.36/0.38).
There were significant interactions for the following desire-to-quit-type items:
How long did you think about quitting the anagram task before you actually did? F(1,
2

102) = 4.26, p =.042, partial ! = .04; and How much did you force yourself to work on
2

the anagram task as long as you did? F(1, 102) = 6.81, p =.01, partial ! = .06. There
was a near significant interaction on the question, “Did you force yourself to keep
working on the anagram task after you first thought about quitting?” F(1, 102) = 3.60, p

73

2

=.06, partial ! = .03. The pattern of means indicated that for the specific-difficult-goal
conditions, high-conservation participants rated these items lower than low-conservation
participants (d = 0.55, 0.70, and 0.40 for the, “Think to quit,” “Force to work as long as
you did,” and, “Force to work after thinking about quitting questions,” respectively). The
opposite pattern held for the do-your-best conditions (d = 0.26, 0.31, and 0.35,
respectively). All other analyses on the anagram manipulation check questions (i.e.,
task difficulty, effort given, self-control, fight an urge, and desire to stop) were nonsignificant (ps > .09).
Table 10 also displays means and standard deviations for manipulation check
questions administered in the final study questionnaire. These questions assessed how
the anagram task instructions affected participants’ motivation and effort. I predicted
participants in the specific-difficult-goal conditions would rate these questions higher
than participants in the do-your-best conditions. This was the case: participants in the
specific-difficult-goal conditions indicated the instructions/goal on the anagram task
2

motivated them to perform better, F(1, 102) = 26.33, p <.001, partial ! = .21, d = 1.02,
2

increase their effort, F(1, 102) = 26.88, p < .001, partial ! = .21, d = 1.03, and work
2

harder, F(1, 102) = 19.56, p < .001, partial ! = .16, d = 0.88.
Participants compared their effort on the anagram task to their effort on the
Stroop task (ratings could range from 1 to 5, with 1 meaning more effort on the Stroop
task and 5 meaning more effort on the anagram task). Participants generally indicated
they put slightly more effort into the anagram task, overall M = 3.93, SD = 1.31; there
were no significant differences across conditions (ps > .76).

74

Table 10: Descriptive Statistics for Study 5 Manipulation Check Questions on the Anagram Task

Variable
a

Effort Invested
Task Difficulty
How hard did you work?
How much effort did you give?
Self-Control
Fight Against an Urge
How much did you want to
stop working?
Did you keep working after you
a
thought about quitting?
How long did you think about
a
quitting before you did?
How much did you force
a
yourself to keep working?
Did goal motivate you to
perform better?
Did goal increase your effort?

Do-Your-Best Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)
b
86.77 19.49
78.33 27.93
88.62
11.10
81.67 19.52
90.73
9.80
85.11 18.86
89.62
11.53
86.00 17.03
55.35
37.30
58.56 36.37
31.81
34.65
42.22 35.44
67.54
37.64
63.96 35.50
63.38

50.08
60.38

b

37.56

84.85
80.88
81.73
84.46
48.77
30.31
68.88

17.21
17.35
16.49
11.29
30.88
32.56
31.13

93.22
82.07
87.11
89.33
45.89
32.37
48.85

10.29
21.15
15.81
15.83
34.59
30.14
38.38

35.17

b

38.84

Specific-Difficult-Goal Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)

c
c

74.30 27.90

78.42 26.58

65.00 40.28

35.03

57.78 25.47

61.85 27.81

44.15 36.97

31.81

68.70 22.98

72.62 24.62

50.22 38.93

32.58

34.89 30.96

63.73 26.17

68.46 27.27

30.41

34.26 28.12

65.27 26.75

64.93 30.36

Did goal influence you to work
c
37.08 32.27
harder?
39.78 29.47
60.23 30.17
68.82 29.57
Stroop to Anagram Effort
Comparison
4.00
1.20
3.93 1.33
3.85 1.43
3.93 1.51
Note. Higher averages on the, “Stroop to Anagram effort Comparison” mean participants rated their effort on the anagram
task higher compared to their effort on the Stroop task.
a
b
Levene’s test of equality of error variances was significant (p < .05) for this question. Indicates a significant Goal Type
c

X Conservation interaction (p < .05). Indicates a significant main effect of goal type (p < .05).
75

Table 11 displays means and standard deviations for the BMIS scales,
vocabulary performance, and reported ACT scores. There was a near significant effect
of goal type on ACT scores; participants in the specific-difficult-goal conditions reported
higher ACT scores than participants in the do-your-best conditions, F(1, 97) = 3.84, p
2

= .053, partial ! = .04, d = 0.41. Otherwise, no other effects or interactions were
significant (ps > .10).

76

Table 11: Descriptive Statistics for Study 5 BMIS, Vocabulary, and ACT Variables

Variable
BMIS
Pleasant-Unpleasant Scale
Arousal-Calm Scale
ACT
Vocabulary
a

Do-Your-Best Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)
5.77
14.65
a
25.00
9.77

6.29
3.74
3.24
2.44

4.14 6.79
16.28 3.59
24.81 3.64
9.71

The main effect of goal type for this item was p = .053.

77

2.64

Specific-Difficult-Goal Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)
4.36 6.09
16.00 3.77
26.15 3.11

4.59 6.36
15.52 3.99
26.46 4.07

10.25 2.03

10.79 2.88

Conservation
I performed a 2 (Conservation: high, low) X 2 (Goal Type: do-your-best, specificdifficult) ANOVA on Stroop task reaction time and accuracy. I used two indicators of
reaction time. The first was the average reaction time for correct trials only. The second
was the average reaction time for correct trials that were under two seconds. I used this
measure to correspond with Muraven et al. (2006); in their fourth study, Stroop trials
were removed from analyses if they lasted more than two seconds. Table 12 displays
means and standard deviations for Stroop performance indices. I expected to find a
main effect of conservation, such that participants in the high-conservation conditions
performed worse in terms of reaction time (i.e., slower reaction times) and/or accuracy
(i.e., fewer correct trials). Since the goal manipulation had not taken place at this point, I
did not expect to find differences across the goal conditions or to find any interactions.
Contrary to my predictions, there were no significant effects on the performance indices
(ps > .48).

78

Table 12: Descriptive Statistics for Study 5 Stroop Task Performance
Do-Your-Best Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)
.97
.03
.97
.03
774.68 221.49
748.53 146.46
733.10 172.42
729.76 135.13

Variable
Accuracy
RT (correct trials only)
RT (correct and under
2 seconds)
Note. Reaction times are in milliseconds.

79

Specific-Difficult-Goal Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)
.97
.03
.98
.02
796.84 200.12
752.47 159.92
759.07 149.36
728.94 142.74

Anagram Persistence and Number Correct
I conducted a 2 (Conservation: high, low) X 2 (Goal Type: do-your-best, specificdifficult) ANOVA on anagram persistence and number correct. Table 13 displays means
and standard deviations for anagram performance. I expected to find a main effect of
goal type, such that participants in the specific-difficult-goal conditions persisted longer
and solved more anagrams correctly than participants in the do-your-best conditions. I
also expected a main effect of conservation such that the high-conservation conditions
would outperform the low-conservation conditions.
The results indicated a significant goal type effect on persistence, F(1, 102) =
2

4.57, p = .04, partial ! = .04, d = 0.42. Participants in the specific-difficult-goal
conditions persisted longer than participants in the do-your-best conditions. While the
effect of goal type on number correct demonstrated the same pattern, it did not reach
2

conventional significance, F(1, 104) = 2.97, p = .088, partial ! = .03, d = 0.34. There
was no main effect of conservation or an interaction on persistence (ps > .19) or number
correct (ps > .21).
Because there was a near significant main effect of goal type on ACT scores, I
performed the ANOVAS again, this time controlling for ACT scores. Some participants
did not report their ACT scores so these analyses was performed on a subset of the
whole sample. When the dependent variable was persistence, the main effect of goal
2

type was marginally significant, F(1, 91) = 3.85, p = .053, partial ! = .04, d = 0.48. The
main effect of conservation remained non-significant (p = .65), but the interaction
2

became marginally significant, F(1, 91) = 3.09, p = .082, partial ! = .03. The pattern of

80

means indicated that for the specific-difficult-goal conditions, participants in the highconservation condition persisted longer than participants in the low-conservation
condition (d = 0.47), whereas the opposite held for the do-your-best conditions (d =
0.28).
When the dependent variable was number correct, the main effect of goal type
was not significant (p = .14). The interaction remained non-significant (p = .17), but the
main effect of conservation became marginally significant, F(1, 93) = 3.47, p = .066,
2

partial ! = .04, d = 0.37. Participants in the high-conservation conditions solved more
anagrams correctly than participants in the low-conservation conditions.

81

Table 13: Descriptive Statistics for Study 5 Anagram Performance

Variable

a

Persistence (in minutes)
Number Correct

Do-Your-Best Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)
a
19.14
9.04
18.12 7.42
31.12 19.00
30.89 17.65

Indicates a significant main effect of goal type (p < .05).

82

Specific-Difficult-Goal Conditions
Low-Conservation High-Conservation
Condition
Condition
M (SD)
M (SD)
20.46 8.26
32.67 15.04

23.65 8.21
40.75 16.99

Anagram Performance Comparison
According to the resource account, people become increasingly depleted as they
continuously regulate. This predicts that performance should progressively worsen as
the number of prior tasks increases, regardless of motivation. According to the
motivation account, people become less and less willing to exert effort so performance
should generally decline as the number of prior tasks increases. However, motivation
should mitigate this decline.
I performed a 3 (Prior Task: none, 1, 2) X 2 (Goal Type: do-your-best, specificdifficult) ANOVA on anagram task performance. I combined the pilot data from Study 3
(no prior task), the data from Study 3 (1 prior task), and the data from Study 4 (2 prior
tasks) to perform this analysis. This analysis allowed me to see how performance on the
same set of anagrams changed depending on how many tasks had been completed
beforehand and what type of instructions were given for the task.
When the dependent variable was number correct, there were main effects of
2

goal type, F(1, 234) = 11.59, p = .001, partial ! = .05, and prior task, F(2, 234) = 7.20,
2

p = .001, partial ! = .06 (See Figure 1). Regarding the main effect of goal type,
participants in specific-difficult-goal conditions solved more anagrams correctly than
participants in do-your-best conditions (d = .43). Regarding the main effect of prior task,
Tukey post hoc tests indicated participants who had no prior task solved more
anagrams correctly than participants who had one prior task (p =.02) or two prior tasks
(p = .001). Performance did not differ between participants who had one or two prior
tasks (p = .70). The interaction was not significant (p = .85).

83

Figure 1: Number correct on the anagram task by goal type and prior task.

When the dependent variable was persistence, there were main effects of goal
2

type, F(1, 232) = 7.86, p = .005, partial ! = .03, and prior task, F(2, 232) = 45.16, p
2

< .001, partial ! = .28 (See Figure 2). Regarding the main effect of goal type,
participants in specific-difficult-goal conditions persisted longer than participants in doyour-best conditions (d = .32). Regarding the main effect of prior task, Tukey post hoc
tests indicated participants who had no prior task persisted longer than participants who
had one prior task (p < .001) or two prior tasks (p < .001). Participants who had one
prior task persisted longer than participants who had two prior tasks (p = .001). The
interaction was not significant (p = .92). In both of these analyses, Levene’s test for
unequal variances was significant.

84

Figure 2: Persistence on the anagram task by goal type and prior task.

Discussion
Participants rated the typing task similarly in terms of how difficult the task was,
how much effort they invested in the task, and how hard they worked on the task. This
suggests that participants were objectively depleted to the same degree before the
second task. On the Stroop task, participants in the high-conservation conditions rated
self-control requirements and the need to fight against an urge higher than participants
in low-conservation conditions. This corroborates the greater emphasis on suppression
in the task instructions for high-conservation participants. However, high-conservation
participants did not indicate that they conserved more or thought about the next task
more than low-conservation participants. Nor did high-conservation participants indicate
that they withheld effort compared to low-conservation participants. Several
manipulation checks on the Stroop task had main effects of goal type or an interaction

85

between goal type and conservation. These effects are not interpretable since the goal
manipulation had not taken place.
Furthermore, analyses of performance measures on the Stroop task did not
reveal any effects of conservation. There was no evidence that anticipating a future task
led participants to conserve resources, and thus perform worse on the Stroop task than
participants who had no anticipation of a future task. These results contradicted my
predictions. I ran two other versions of this study where participants completed a typing
task, a timed anagram task, and then another typing task. I found no evidence for
conservation on the second task in either of these cases.
Hagger et al. (2010) reported an averaged corrected standardized effect size of
+

d = 1.04 for the conservation effect. Although this effect was quite large, the
conservation effect is a relatively new concept in depletion literature. Hagger et al.
(2010) only used seven studies from two research articles to meta-analyze the effect.
Even though my study was designed as a partial replication of Muraven et al.’s fourth
study (2006), I was unable to replicate the conservation effect. Since conservation is a
newer aspect of depletion research, known methods of producing the effect are
relatively limited and require further investigation.
It is possible that the conservation manipulation in this study was not strong
enough to produce the effect. When I conducted Studies 2-4, a steady flow of students
signed up for participation; during Study 5, the number of students available in the
subject pool was drastically lower and they signed up for participation sporadically. This
may have affected the results. In Study 5, participants were primarily guided through the
study by computer instructions, whereas in Studies 2-4, participants were primarily

86

guided through the study by the experimenter. This change in procedure may have
affected the results. However, I found no conservation in the two other versions of this
study I conducted, in which the experimenter guided participants through the study.
It is also possible that individual differences affect whether conservation occurs
or not. For example, individuals may differ in how much effort they are normally willing
to give and how consistently they exert effort. As another example, individuals may
differ in how much they enjoy doing anagrams or other cognitive tasks.
In the future, a very direct manipulation of conservation may be necessary. For
example, participants may need explicit instructions telling them that they would be
more efficient on a future regulatory task if they conserve effort on the current task.
Other participants would not be told about the future task. Then, goal instructions could
be manipulated to see if participants who did not conserve can perform as well or better
than participants who were explicitly told to conserve.
Concerning the anagram manipulation check items, I had predicted that highconservation participants would rate effort-type questions higher and desire-to-quit-type
items lower than low-conservation participants. If participants had conserved more on
the Stroop task, they should have been more willing to continue exerting effort on the
anagram task. This pattern of results appeared in the specific-difficult-goal conditions,
but was the opposite in the do-your-best conditions.
Participants in the specific-difficult-goal conditions tended to report higher ACT
scores than participants in the do-your-best conditions. Even though this effect did not
reach the conventional level of significance, I examined anagram performance in two
ways: controlling for ACT scores or not. When I did not control for ACT scores, there

87

was a significant main effect of goal type. Consistent with my predictions, participants
given a specific, difficult goal persisted longer on the anagrams than participants given
do-your-best instructions. The effect of goal type on correct anagram solutions did not
reach significance.
When I controlled for ACT scores, the results changed somewhat. For
persistence, the main effect of goal type was qualified by the interaction. The results
followed the same pattern as the manipulation check items for the anagram task. When
high-conservation participants received a specific, difficult goal, they persisted longer
than low-conservation participants. When high-conservation participants received doyour-best instructions, they persisted less than low-conservation conditions. For number
correct, the main effect of goal type went away and a small effect of conservation
emerged. High-conservation participants solved more anagrams correctly than lowconservation participants.
These results suggest that participants generally persisted longer and solved
more anagrams if they had a specific, difficult goal. However, this effect may depend on
conservation; high-conservation participants performed better when they had a specificdifficult goal. This may suggest that conservation promotes greater effort exertion on
subsequent tasks if motivation is high, but promotes continued conservation if
motivation is not as high. This pattern is similar to learned industriousness where people
continue to exert high effort when it is rewarded. These interpretations are speculative,
however, because I did not actually observe conservation on the Stroop task.
The absence of a conservation effect implies that participants were doubly
depleted going into the third task. In other words, because performance did not differ

88

across conditions on the first or second task, all participants regulated to the same
degree and went into the third task depleted to the same degree. If depletion truly
hinders the ability to regulate, performing well after two tasks should be more difficult
than after one task. Muraven et al. (2006) found participants who had not conserved on
a second task performed worse on a third task. I found that even when participants had
presumably not conserved on a second task, I could manipulate some participants to
perform better by giving them a goal. This suggests that when participants are doubly
depleted, they can continue to exert effort and improve performance.
In an additional set of analyses, I compared participants who presumably
performed the anagram task at full capacity (i.e., Study 3 pilot conditions), participants
who performed the anagram task depleted (i.e., Study 3 conditions), and participants
who performed the anagram task doubly depleted (i.e., Study 5 conditions). Participants
who had no prior task persisted the longest and solved more anagrams correctly.
Persistence then decreased as the number of prior tasks increased. However,
participants who were depleted and doubly depleted solved the same number of
anagrams, suggesting that the ability to perform was not exhausted by prior regulatory
experience. Additionally, the specific, difficult goal led to better performance across the
board. Overall, despite null results for conservation, this study provides evidence that
even after two demanding regulatory tasks, participants are able to regulate, particularly
when they are more motivated.

89

CHAPTER 7: GENERAL DISCUSSION
When individuals engage in self-regulation, regulatory performance subsequently
suffers. Known as ego-depletion, or the depletion effect, this phenomenon has been
demonstrated in many studies. Hagger et al.’s (2010) recent meta-analysis provided
further evidence that the depletion effect is medium to large in magnitude. Nevertheless,
the phenomenon is not fully understood.
The depletion effect has typically been explained in terms of a resource account.
The primary assumptions of the resource depletion model are that self-regulation relies
on a domain-general resource, that the resource is limited, and that the resource is
expendable (Muraven & Baumeister, 2000). Proponents of the resource account have
attributed the depletion effect to an inability to regulate because of diminished resources,
or as an attempt to conserve what resources remain. Proponents of the resource
account have also argued that the depletion effect will only occur after a task that
involves self-regulation.
However, I identified a potential confound in previous studies of the depletion
effect. The majority of depletion studies have one initial task that involves self-regulation
and one initial task that does not. Individuals rate the task that involves self-regulation
as more difficult and more effortful than the task that does not involve self-regulation.
Thus, the depletion effect may arise from differences in self-regulation or from
differences in task difficulty and effort exertion. In support of the impact of effort, my
meta-analysis suggested there is a small relationship between effort ratings on the first
task and the depletion effect on the second task. Additionally, I found a depletion-like
effect after two initial tasks that presumably did not involve self-regulation. This

90

evidence suggests that motivated allocation of effort poses an alternative account of the
depletion effect.
A motivation account derives from the law of least effort and from the notions that
people are cognitive misers and motivated tacticians. While people generally resort to
low-effort strategies, effort exertion is adaptive. People will increase or decrease their
effort depending on many factors. In my studies, I found that participants increased their
effort if they were given specific, difficult goals. I also found that participants decreased
their effort if they believed that they had already exerted more effort than others. These
performance shifts occurred even when participants had completed the same depleting
task. Thus, motivation manipulations had a greater influence on performance than
“objective” levels of depletion.
Recent findings suggest that the motivation to conserve influences the depletion
effect (Muraven et al., 2006; Tyler & Burns, 2009). That is, individuals perform worse or
behave in a depleted manner as a way of preserving their diminishing resource.
However, if conservation ties to motivated effort allocation rather than resource
preservation, then individuals should be capable of performing well even when they
have not conserved, so long as they are motivated. In Study 5, I investigated whether
people conserve a resource or conserve effort. While I was unable to detect a
conservation effect, I did find that motivation manipulations continued to affect
performance, even after two depleting tasks.
Theoretical Implications
The findings from my research are consistent with a motivation account of the
depletion effect and suggest that effort and the motivation to exert effort have a large

91

impact on self-regulatory performance. However, as Hagger et al. (2010) pointed out in
their meta-analysis, the motivation and resource account may not be mutually exclusive.
If this is the case, depletion may occur as the resource account explains, but motivation
may moderate this effect. This would suggest that motivation would not always
overcome depletion. In other words, people could become depleted to the point that no
amount of motivation would improve performance. My research showed that motivation
improved performance even after two prior tasks. Further research will have to examine
whether this improvement continues beyond two prior tasks.
Nevertheless, my findings showed that performance generally declined as the
number of prior tasks increased. Hence, instead of being exclusive, the two accounts
may be describing the same resource, albeit differently. According to the resource
account, the resource is self-regulatory; by contrast, the resource may be effort. In other
words, performance may decline across tasks because of people’s unwillingness to
continue using their effort resources, rather than people’s inability to continue using their
regulatory resources. The key distinctions are that resource account assumes the
resource is readily expendable and regulatory specific. The motivation account
assumes that the motivation to use effort resources decreases after any effortful
behavior. Additionally, the ability to exert effort does not decline; rather, the choice to
exert effort changes over time. The resource depletion model focuses on the availability
of a resource, whereas the motivation account focuses on decisions about using a
resource.
In this sense, depletion effects become the result of dynamic decision-making
processes over time. This perspective is not new, but draws on an evolving set of

92

regulatory theories. Many control theories of self-regulation hypothesize that people
make decisions to reduce the discrepancy between a current state and a desired goal
state; these decisions determine regulatory behavior, effort allocation, and strategies
(e.g., Carver & Scheier, 1998). These decisions are thought to rely on economics
principles such as expected utility, or combining the value of an outcome with the
likelihood of it occurring. One criticism of these theories is that parameters, such as the
valence of an outcome, are modeled as static.
In recent years, Vancouver, Weinhardt, and Schmidt (2010) have begun
modeling self-regulation by incorporating control theories, decision-making theories, and
the theory of multiple-goal pursuit. This theory postulates that people are continuously
allocating resources between competing goals over time. Goals compete due to both
physical and psychological limitations; when a person chooses to pursue one goal, it
inevitably means the individual cannot pursue other goals. This model allows
parameters to be dynamic, to parallel how goal priorities, effort toward goals, and so on
change over time.
When self-regulation is viewed in terms of dynamic decisions between multiple
goals over time, depletion effects become a piece of a larger puzzle. For example, as a
participant enters a depletion study, both proximal and distal goals may affect
motivation and effort allocation on the administered tasks. Proximally, he or she may
have certain goals about the study itself, such as wanting to perform well on the
presented tasks. Distally, he or she may have certain goals for the day, such as
preparing for an upcoming test. These two goals compete; the participant cannot study
for the exam while completing the study. Thus, effort allocation within the study

93

becomes a dynamic decision between multiple goals. In my studies in particular,
competing goals create a situation where the participant is constantly deciding whether
to quit or keep working on the anagrams.
The concepts of the cognitive miser and motivated tactician easily fit within a
dynamic decision-making framework. With such a complex system of goal pursuit,
efficiency depends on adaptively adjusting effort according to goal priorities and other
factors. Models such as Vancouver et al.’s (2010), that seek to integrate many areas of
research, may be the key to understanding self-regulatory behaviors generally, and the
depletion effect specifically. However, understanding the role of goal pursuit and effort
allocation in the depletion effect will also involve identifying task factors and individual
differences that affect motivation.
For example, future research is needed to fully understand the time course of
depletion effects. Several components may play a factor such as the length of each task,
the length of time between tasks, and the participants’ overall expectations of time in the
study. The length of initial tasks in most depletion studies is as little as five minutes and
typically no more than ten minutes (e.g., Muraven et al., 1998). This means the
depletion effect occurs after a relatively short period. If willingness to exert effort or the
availability of a resource decrease with time, then depletion effects should be greater
after longer initial tasks, but this type of manipulation has rarely been examined. Some
research has shown that the length of time between tasks is important and that longer
breaks lead to better performance on a second task (e.g., Tyer & Burns, 2008).
Expectations also seem to impact regulatory performance; participants performed better
when they thought the study was almost over compared to when they expected to

94

continue for some time (e.g., Tyler & Burns, 2009). In my studies, participants were told
how long the study would last (e.g., one hour). Some participants may have persisted
longer because they knew they would have to remain in the study anyway. Others may
have disregarded this information and persisted less with the hopes of leaving the study
early.
If time expectations affect regulatory performance, other exogenous factors may
influence depletion effects, such as task complexity or task importance. Additionally,
factors such as the reward structure within tasks may influence whether motivation
increases or decreases across tasks (e.g., Eisenberger, 1992). In my studies, I used
several sets of anagrams that differed in terms of solvability and difficulty. These task
properties may differentially affect participants’ persistence. For example, when
participants encounter unsolvable anagrams, they may decide more quickly that effort is
futile, and persist less.
Regarding individual differences, many factors may influence regulatory behavior
and depletion effects, ranging from personality differences, to cognitive ability
differences, to personal preferences, and to life experience. Recent research on
regulatory focus, or whether individuals are oriented towards promoting gains or
preventing losses, suggested that individuals’ motivation to perform is influenced by
whether their preferred orientation fits their approach on a current task or goal (e.g.,
Avnet & Higgins, 2003; Kruglanski et al., 2000). Wan and Sternthal (2008) found that
individuals who were more likely to monitor their behavior against a standard were less
likely to exhibit depletion effects. Job, Dweck, and Walton (2010) have shown that
individuals’ beliefs about self-regulation impact their regulatory behavior such that those

95

who believe self-regulation is limited are more likely to behave in a depleted manner
than those who believe self-regulation is not limited.
Other individual difference factors that may influence the depletion effect and the
motivation to exert effort include work ethic, stress levels, and sleep patterns. According
to the resource account, training reduces depletion effects (e.g., Muraven et al., 2010).
This training may reflect learned industriousness, or developing the skill to continue
exerting effort over time (e.g., Eisenberger, 1992). Individuals may differ in how much
self-regulatory training they engage in or how responsive they are to learned
industriousness. In my studies, some individuals may have persisted longer because
they have trained themselves to continue exerting effort for longer periods of time, even
when the desire to quit is strong.
Relating to physiological factors, Hagger et al. (2010) suggested that there are
significant differences in ratings of fatigue between control and depletion conditions. In
my meta-analysis, I did not find a relationship between self-reported fatigue and the
depletion effect, suggesting that any differences were subjective. However, after
continued exertion, there may be a point for individuals when true physical fatigue
affects regulatory performance.
Further Issues, Limitations, and Future Directions
The present research focused on a limited number of tasks and motivation
manipulations. This was beneficial in one sense because it allowed me to determine if
the effects of motivation would occur consistently across the narrow set of tasks.
However, in order to determine whether the impact of motivation generalizes to the
depletion effect at large, future studies should use additional tasks and manipulate

96

motivation and effort in new ways. Future motivation manipulations could include more
than two levels in order to investigate whether the motivation to exert effort has a linear
relationship with performance. It would also be interesting to see whether individuals are
able to estimate how much effort they are willing to give and whether this changes over
time. This becomes a question of meta-cognition: how aware are individuals about the
dynamic decisions they make to allocate effort among goals?
The current research is also limited because it did not include a direct test of the
resource account against the motivation account and did not replicate the typical
depletion effect. One step towards distinguishing whether these two accounts are
exclusive or work together would be to cross regulation and motivation in a subsequent
study. On the first task, participants would complete a regulatory or non-regulatory task.
Then, the second task would be the same for all participants, but a motivation
manipulation would encourage participants to exert more or less effort. Perceived effort
feedback or goal instructions could achieve this kind of manipulation.
This type of study would allow an assessment of how the motivation to exert
effort affects performance after a regulatory and non-regulatory initial task. If regulation
is critical, as the resource account suggests, performance should be worse after the
initial regulatory task compared to the initial non-regulatory task. If motivation plays a
role in performance, participants manipulated to exert more effort should outperform
participants manipulated to exert less effort. If both motivation and a resource affect
performance, there should be an interaction. Participants who performed a regulatory
task first should not perform as well as participants who performed a non-regulatory

97

task first, even when they are more motivated. This finding would suggest that
motivation was unable to overcome decrements in self-regulatory ability.
Manipulating regulation on a second task, in addition to the manipulations
described above, would also contribute to the distinction between a resource and
motivation account. According to the resource account, the depletion effect should not
occur if the second task does not involve self-regulation. According to the motivation
account, the depletion effect should occur regardless of whether the task is regulatory
or non-regulatory, so long as both the first and second task are effortful and demanding.
With these types of manipulations, researchers can begin to distinguish the role of selfregulatory behavior and effortful behavior in depletion effects.
In the development of a direct test of the resource and motivation account,
researchers must address several issues. First, what is the theoretical distinction
between a regulatory and non-regulatory task? Because the resource account does not
clearly define the nature of the resource, this issue is challenging. The most consistent
definition of self-regulation in depletion studies is the ability to override responses, or
inhibition (see Muraven et al., 1998). This may be the starting point to distinguish
regulatory from non-regulatory tasks. Second, the regulatory and non-regulatory tasks
must be equally difficult and effortful. This would ensure that differences in performance
are due to inhibitory requirements between tasks and not a confound between effort and
regulation. Addressing these issues would involve extensive pilot testing. However, if
self-regulation is operationally defined as dynamic decision making amongst concurrent
goals, then distinguishing between regulatory and non-regulatory tasks may not be
possible.

98

Regardless of whether the resource and motivation account are mutually
exclusive, future research is needed to investigate what mechanism underlies depletion
effects. Even with dynamic decision-making processes, task factors, and individual
differences affecting self-regulation and the depletion effect, is there a specific cognitive
mechanism that underlies this phenomenon? While this question has not been
thoroughly examined, some researchers have begun to investigate what specific
regulatory process might affect depletion.
Specifically, this research links changes in monitoring processes to depletion.
Lorist, Boksem, and Ridderinkhof (2005) examined continuous task performance on a
flanker task, in which participants identify a central stimulus when flanking stimuli match
or mismatch. Post-error slowing, or the phenomenon that people take longer to respond
on trials following an error, decreased as time on task increased. Event-related
negativity (ERN), which is associated with performance monitoring in the anterior
cingulate cortex, also decreased with time-on-task. Boksem et al. (2006) replicated this
finding of reduced post-error slowing on a different task. Several event-related potential
components also indicated that with increasing time-on-task, participants were less
capable of recognizing and reacting to response conflict.
Inzlicht and Gutsell (2007) also hypothesized that depletion may weaken
performance monitoring. To test this hypothesis, they examined ERN in the Stroop task
after control and depletion conditions. ERN data indicated that the control condition had
greater negativity on incorrect trials than correct trials; there was no difference in
negativity on incorrect and correct trials for the depletion condition. This suggested that
the conflict monitoring system might have been inefficient in the depletion condition

99

compared to the control condition, or may have been incapable of detecting a
discrepancy between intentions and actions.
These studies have examined the depletion effect from a mechanism perspective,
with the aim of determining what cognitive processes or functions link directly to the
depletion effect. If the resource account and motivation account are not exclusive, this
research may be the bridge between the two accounts. The ability to regulate may
decline over time, but the mechanism that breaks down may tie to effort allocation. That
is, a key component of decision-making, or recognizing discrepancies between current
states and goal states, may break down, thus making it difficult for individuals to know
where to allocate resources or effort.
While these studies suggest that monitoring processes deteriorate over time,
motivation may still have a significant impact on performance. For example, Wan and
Sternthal (2008) eliminated conflict monitoring deficits associated with the depletion
effect simply by encouraging participants to monitor their performance against a
standard. Thus, participants may be able to overcome presumed deficits in cognitive
processing, so long as there is motivation to do so.
Conclusion
While my research cannot completely rule out the resource account, it ultimately
demonstrates that researchers need to expand their thinking about the potential causes
and theoretical underpinnings of depletion effects. Furthermore, my research shows that
the ability to self-regulate may not be as limited as the resource account implies. Rather,
the motivation to exert effort has a great impact on regulatory behavior in depletion
paradigms.

100

Researchers should continue to explore mechanistic explanations of the
depletion effect and integrate this with research concerning the effects of motivation and
effort allocation on self-regulatory performance. Additionally, researchers should explore
how the depletion effect, cognitive miserliness, and the motivated tactician concept fit
into regulatory goal theories (e.g., see Vancouver et al., 2010). As individuals regulate,
they must pursue multiple goals simultaneously, and effort allocation between these
goals is likely a complex system involving past experience, current contexts, and future
expectations. Ultimately, integrating different avenues of research will increase
understanding of depletion effects and will broaden scientific understanding of selfregulation.

101

APPENDICES

102

APPENDIX A

META-ANALYSIS DETAILS
Table A1: Condition ns and Sample Weights for the Focused Meta-Analysis
Author

Alberts, Martijn, Greb,
Merkelbach, & Vries (2007)
Baumeister Bratslavsky,
Muraven, & Tice (1998)
Bray, Ginis, Hicks, &
Woodgate (2008)
Burkley (2008)
Ciarocco, Sommer, &
Baumeister (2001)
DeWall, Baumeister,
Stillman, & Gailliot (2007)
Fennis, Janssen, & Bohs
(2009)
Finkel & Campbell (2001)
Finkel et al. (2006)

Friese, Hofmann, & Wanke
(2008)
Gailliot & Baumeister
(2007)
Gailliot, Schmeichel, &
Baumeister (2006)
Govorun & Payne (2006)
Hofmann, Rauch, &
Gawronski (2007)

Study

Depletion
Effect

Self-Report
Rating Effect

Type of
Rating
Provided

WLS
Sample
Weight

2

nc
20

nd
20

nc
40

nd
40

F

40

1

23

23

23

23

D, E, F

46

3
4
1

15
41
23

15
41
26

15
41
23

15
41
26

D, E
E, F
E, F

40
82
49

3
4
1

39
30
18

39
30
19

39
30
18

39
30
19

E
E
D

78
60
37

2
2

12
13

12
14

12
26

12
27

D
D

24
27

4
3

48
19

49
18

48
19

49
18

D
E

97
37

2
1
2
3
4
2

23
13
14
23
16
33

23
13
15
23
16
33

23
13
14
23
16
33

23
13
15
23
16
33

E
F
F
F
F
E, F

46
26
29
46
32
66

3
1

22
16

23
16

22
16

23
16

E, F
D

45
32

2
2

13
9

14
10

13
9

14
10

D
D

27
19

36
24

36
26

36
24

36
26

D, F
D

72
50

103

Table A1 (cont’d)
Joireman, Balliet, Sprott,
Spangenberg, & Schultz
(2008)
Legault, Green-Demers, &
Eadie (2009)
Martijn et al. (2007)
Martijn, Tenbult,
Merckelbach, Dreezens, &
de Vries (2002)
Muraven & Slessareva
(2003)

Neshat-Doost, Dalgleish, &
Golden (2008)
Oikawa (2005)
Ostafin, Marlatt, &
Greenwald (2008)
Park, Glaser, & Knowles
(2008)
Pocheptsova, Amir, Dhar,
& Baumeister (2009)
Schmeichel (2007)

Schmeichel & Vohs (2009)
Schmeichel, Demaree,
Robinson, & Pu (2006)

50

49

50

D

99

2

34

34

67

67

F

68

18
16

19
17

18
16

19
17

F
D

37
33

1

11

11

21

22

D, E

22

21
12
29

20
12
29

41
23
29

41
22
29

D, E
E
D, E

41
24
58

1

23

23

23

23

E

46

2
4
1

17
19
20

17
19
20

17
19
20

17
19
20

E
E
E, F

34
38
40

2
3

Muraven, Tice, &
Baumeister (1998)

49

2
3
Muraven, Collins, &
Nienhaus (2002)
Muraven, Shmeuli, &
Burkley (2006)

3

17
23
14

17
26
18

17
23
14

17
26
18

D
D
D, F

34
49
32

1
2

19
20
40

19
20
45

19
20
40

19
20
45

F
F
E

38
40
85

28

29

28

29

D, E

57

2

249

252

1S1
1S2
2S1
3
4
1
2

38
31
29
30
33
29
12
25

41
31
32
15
32
30
12
25

94
95
69
69
29
30
33
29
36
25

94
95
72
72
32
15
32
30
36
25

D
F
D
D
D
D
D
E
D
D

188
190
79
62
61
45
65
59
24
50

104

Table A1 (cont’d)
Schmeichel, Vohs, &
Baumeister (2003)

Tyler & Burns (2008)
Tyler & Burns (2009)

Vohs & Faber (2007)
Vohs & Schmeichel (2003)
Vohs, Baumeister, &
Ciarocco (2005)

Wan & Sternthal (2008)

Webb & Sheeran (2003)
Wheeler, Brinol, &
Hermann (2007)
Wright et al. (2007)
Wright, Martin, & Bland
(2003)

12

12

12

12

D

24

2
3
Segerstrom & Nes (2007)
Shamosh & Gray (2007)
Shmueli & Prochaska
(2009)
Stewart, Wright, Hui, &
Simmons (2009)
Stucke & Baumeister
(2006)
Tyler (2008)

1

18
18
20
28
48

19
18
21
29
49

18
18
41
28
50

19
18
42
29
50

D
D
E, F
E
D, E

37
36
41
57
97

21

19

21

19

D, F

40

3

22

23

22

23

E

45

1
2
3
4
1
2
1
2S1
2S2
3
1
2
3
3
4
3

16
15
15
30
10
10
10
10
10
20
17
34
19
24
23
15

17
15
15
30
10
10
10
10
10
20
18
34
20
24
24
15

16
15
15
30
30
20
20
30
30
40
17
34
19
24
24
15

17
15
15
30
30
20
21
30
30
40
18
34
20
24
25
15

D
D
D, F
D, F
D
D
D
D
D
D
D
D
D
D
D
D

33
30
30
60
20
20
20
20
20
40
35
68
39
48
47
30

4
6
7
8
1
2
3
4
1
2

30
14
16
16
12
13
25
21
15
14
17

30
15
16
16
13
14
25
21
16
14
17

30
29
35
16
25
40
25
41
15
28
64

30
29
36
16
25
41
25
42
16
29
64

D
D
D
D
E
E
E
E
D, E, F
D, F
D, E, F

60
29
32
32
25
27
50
42
31
28
34

2S2
S2

23
18

24
19

23
18

24
19

D
D, F

47
37

105

Table A1 (cont’d)
Wright, Stewart, & Barnett
(2008)

S1

26

27

53

55

D

53

Note. nc = sample size of the control condition used to compute the effect size. nd =
sample size of the depletion condition used to compute the effect size. WLS = weighted
least squares analysis. D = difficulty rating. E = effort rating. F = fatigue rating.

106

APPENDIX B

MANIPULATION CHECK QUESTIONS
Study 2
Questions Administered Following the Math and Anagram Task:
How difficult was this task?
(1=Not Difficult at All; 100= Very, Very Difficult)
How hard did you have to work at this task?
(1= Not Hard at All; 100= Very, Very Hard)
How much effort did you have to invest to perform this task?
(1= Not Very Much Effort at All; 100= All of my Effort)
Studies 3, 4, & 5
Questions Administered Following the Typing, Stroop, and Anagram Task
What percentage of your total effort did you invest in the [typing/Stroop/anagram] task?
(1= 1% of my Effort; 100= 100% of my Effort)
How difficult was the [typing/Stroop/anagram] task?
(1=Not Difficult at All; 100= Very, Very Difficult)
How hard did you have to work at the [typing/Stroop/anagram] task?
(1= Not Hard at All; 100= Very, Very Hard)
How much effort did you have to give to perform the [typing/Stroop/anagram] task?
(1= Not Very Much Effort at All; 100= All of my Effort)
Questions Specific to the Anagram Task
How much did you want to stop working on the anagram task?
(1= Not Very Much at All; 100= Very, Very Much)
Did you force yourself to keep working on the anagram task after you first thought about
quitting?
(1= Not at All; 100= To a Great Degree)
How long did you think about quitting the anagram task before you actually did?
(1= Not Very Much at All; 100= Very, Very Much)
How much did you force yourself to work on the anagram task as long as you did?
(1= Not Very Much at All; 100= Very, Very Much)
Study 3
Questions Regarding the Goal Manipulation (Given on study questionnaire)
How much did the goal you were given motivate you to perform better on the anagram
task?
(1= Not Very Much at All; 100= Very, Very Much)

107

How much did the goal you were given increase your effort on the anagram task?
(1= Not Very Much at All; 100= Very, Very Much)
How much did the goal you were given influence you to persist longer on the anagram
task?
(1= Not Very Much at All; 100= Very, Very Much)
Study 4
Questions Regarding the Pupil Diameter Manipulation (Given on study questionnaire)
How similar was your feeling of how much effort you put into the typing task and the
report on your effort from the pupil diameter data?
(1= Not Very Similar; 100= Very Similar)
How accurately did the pupil diameter measure your effort on the typing task?
(1= Not Very Accurately; 100= Very Accurately)
Did your effort report on the typing task motivate you to increase your effort on the
anagram task?
(1= Did Not Motivate at All; 100= Motivated Me Very Much)
Did your effort report on the typing task influence you to decrease your effort on the
anagram task?
(1= Did Not Decrease Effort at All; 100= Decreased Effort Very Much)
Study 5
Questions Added to those Administered after the Typing, Stroop, and Anagram Task
How much self-control did you have to use to perform the [typing/Stroop/anagram] task?
(Where 1= Not Very Much Self-Control at All and 100= All of my Self-Control)
How much did you have to fight against an urge to perform the [typing/Stroop/anagram]
task?
(Where 1= Not Very Much at All and 100= Very, Very Much)
Questions Specific to the Stroop Task
How would you compare your effort on the typing task to the color-word identification
task?
Select one:
__________ A lot more effort on the typing task than the color-word task
__________ Somewhat more effort on the typing task than the color-word task
__________ The same amount of effort on the typing and color-word task
__________ Somewhat more effort on the color-word task than the typing task
__________ A lot more effort on the color-word task than the typing task
Were you conserving energy for the next task?
(Where 1= Not Very Much at All and 100= Very, Very Much)
How much did you think about the next task while working on the color-word
identification task?
(Where 1= Not Very Much at All and 100= Very, Very Much)

108

What task do you expect to perform next?
__________Don’t know
__________Color-word identification task
__________Anagram Task
__________Typing Task
__________Maze Task
Questions Regarding the Goal Manipulation (Given on study questionnaire)
How much did the instructions/goal you were given motivate you to perform better on
the anagram task?
(1= Not Very Much at All; 100= Very, Very Much)
How much did the instructions/goal you were given increase your effort on the anagram
task?
(1= Not Very Much at All; 100= Very, Very Much)
How much did the instructions/goal you were given influence you to work harder on the
anagram task?
(1= Not Very Much at All; 100= Very, Very Much)
How would you compare your effort on the color-word identification task to the anagram
task?
Select one:
__________ A lot more effort on the color-word task than the anagram task
__________ Somewhat more effort on the color-word task than the anagram task
__________ The same amount of effort on the color-word task and anagram task
__________ Somewhat more effort on the anagram task than the color-word task
__________ A lot more effort on the anagram task than the color-word task

109

APPENDIX C

48 ANAGRAMS FROM STUDY 2
LTEUBLA (bullate)
GROADNE (groaned)
LENPTAE
UOLDIBE
FSNAITE (fainest)
OECARDE
TRAETCR (retract)
MRBTHUE (thumber)
AEDRNOM (madrone)
ARVHTEL
SHMCUEL (mulches)
THATROE
RPSEONH
KECUBEL
RATSIID (diarist)
DNOWIWE
SCUREED (recused, reduces, rescued,
secured, seducer)
TNHRCIE (cithern)
LODLANE
NPGRISA (parings, parsing, rasping,
sparing)
AICOLST (stoical)
MBYLSOA
ONADESN

ENELGTD (gentled)
OETKPCH
CSEDOLA (solaced)
LEYPSET (steeply)
CABLHED
PLECINA (capelin, panicle, pelican)
OMCNMOT
POSTSGI (spigots)
EMKOONY
GLUAERR (regular)
IIDVEDE
OEFSWLR (flowers, fowlers, reflows)
ROFAVSL (flavors)
CMBHAOT
CTWSIHT
GAWHEIT
TETLELI
EODCMYN
URAHHCC
OERIRFPM
ONTCUESL
LSLIMNOI
SIVTION
LEKHPIC
NORCGEA (acrogen)

Solutions are in parentheses for the solvable anagrams.

110

APPENDIX D

DEMOGRAPHICS QUESTIONNAIRE
What is your age? _____ years
What is your gender?

M

F

What is your ethnic background (please circle all that apply)?
a. African-American or Black
b. American Indian or Alaska Native
c. Asian-American or Asian
d. Caucasian (non-Hispanic)
e. Hispanic
f. Native Hawaiian or Other Pacific Islander
Other (please specify): __________________________
What was your total score on the ACT or SAT? _____
* Are you a native English speaker?

Yes _____

No _____

*This question was included in the demographics questionnaire for Study 5.

111

APPENDIX E

TYPING TASK DETAILS
Text from Lentricchia (1990)
There’s a little story once told by Wallace Stevens that I have to replot as I retell it. The
story (Steven’s’ and mine) is actually an “anecdote”: from the Greek, anekdota, meaning
unpublished items. More familiarly, in English, a small gossipy narrative generally of an
amusing, biographical incident in the life of a famous person whose biography’s broad
outline has long been a matter of public record. And more: this biography is often—
when the famous person is also exemplary—a concentrated representation of the
idealized story that a culture would like to tell about itself. Like all anecdotes, then, the
one I have in mind can’t work as an anecdote unless it somehow tells a story beyond
the one it tells. So: an unpublished little story, funny and biographical, apparently stands
in for a bigger story, a socially pivotal and pervasive biography which it illuminates—in
an anecdotal flash the small story reveals the essence of the larger story and in that
very moment becomes exegesis of a public text; the hitherto unpublished items become
published. The teller of anecdotes has to presume the cultural currency of that large,
containing biographical narrative which he draws upon for the sharp point he would give
his anecdote, whose effect is ultimately political: to trigger a narrative sense of
community that the anecdote evokes by evoking the master biography. In evoking the
master biography, anecdote helps us to remember. And remembrance, so triggered, is
the power which sustains, by retrieving, our basic cultural fiction. One day, when he was
a little boy, George Washington chopped down a cherry tree in his father’s orchard.
Americans usually get the point in a hurry; we hardly need to finish the story and deliver
its famous moral punchline: the relationship of government and the people who elect it
is transparent, sincere because the origin of the USA was honest.
Number of Words in Text: 313
Number of E’s in Text: 195
Number of Spaces in Text: 309
Instructions:
Variation for Study 3
In the following task, a passage will fill the screen. You will need to type out this
passage, including punctuation and capitalization. However, you must type the passage
without typing spaces or the letter E. While you will be unable to see what you type,
type as quickly and as accurately as you can. Please do your best; correct any mistakes
that you make, and continue typing until you are told to stop.
Variation for Study 4
In the following task, a passage will fill the screen. You will need to type out this
passage, including punctuation and capitalization. However, you must type the passage

112

without typing spaces or the letter E. While you will be unable to see what you type,
type as quickly and as accurately as you can. Also, please do not look at your fingers or
the keyboard as you type, as this will interfere with the eye tracking. Please do your
best; correct any mistakes that you make, and continue typing until you are told to stop.
Variation for Study 5
In the following task, a passage will fill the screen. You will need to type out this
passage, including punctuation and capitalization. However, you must type the passage
without typing spaces or the letter E. For example, if the passage said the following: The
table had three chairs beside it. You would type the following: Thtablhadthrchairsbsidit.
While you will be unable to see what you type, type as quickly and as accurately as you
can, and avoid making errors. Errors are typing E's or spaces, typing other letters or
punctuation that are not in the paragraph, and skipping over other letters and
punctuation that are in the paragraph. Please do your best; correct any errors that you
make, and continue typing until the computer indicates you are done. As soon as you
see the passage, start typing without using spaces or the letter E.

113

APPENDIX F

80 ANAGRAMS FROM STUDIES 3 & 5
LTIGN (glint)
CNIHF (finch)
CHITP (pitch)
NRCIA (cairn)
FTEIH (thief)
HTIWD (width)
ITRUF (fruit)
TCUON (count)
ADEBL (abled, baled,
blade)
EUNDC (dunce)
HMIEC (chime, hemic)
TERDN (trend)
AUGRD (guard)
PHTED (depth)
AVTLU (vault)
OCPHR (porch)
CNHLU (lunch)
NRTOF (front)
JEGUD (judge)
OAJRM (major)
BOTIR (orbit)
SRIVU (virus)
HECPR (perch)
HAENY (hyena)
GIFTH (fight)
FTLRI (flirt)

KASNC (snack)
ORDCW (crowd)
RCNHA (ranch)
ONAPR (apron)
NBIAC (cabin)
HLVOE (hovel)
FIYRA (fairy)
KRTCU (truck)
OWNLC (clown)
IPTLU (tulip)
ONGYA (agony)
FNKIE (knife)
NEIBR (brine)
ZLTWA (waltz)
CRIAV (vicar)
IUNYT (unity)
YAORV (ovary)
HSAFL (flash)
HNDUO (hound)
LCKPU (pluck)
IMENC (mince, nemic)
GITHL (light)
GBRUY (rugby)
CIAHR (chair)
LKCOA (cloak)
HSLAC (clash)
HRCUS (crush)

Solutions are in parentheses.

114

GBTIO (bigot)
PLIMB (blimp)
KTENO (token)
TRHIM (mirth)
NIRDK (drink)
ACNFY (fancy)
NOEHY (honey)
OFREC (force)
EMCYR (mercy)
HACTM (match)
UTDBO (doubt)
ECRVO (cover)
MACPR (cramp)
YQERU (query)
BNHUC (bunch)
OWAMN (woman)
HTMON (month)
OUSTC (scout)
GEVOL (glove)
LAKPN (plank)
ORHCD (chord)
WSHLA (shawl)
IOGLC (logic)
RETIV (rivet)
TELSY (style)
IJNOT (joint)
CMRBU (crumb)

APPENDIX G

STUDY 4 PUPIL DATA REPORT FOR HIGH-PERCEIVED-EFFORT PARTICIPANTS
Background on Pupil Dilation:
Dilation of the pupil changes over time. It is commonly known that pupil dilation
changes as the amount of sunlight hitting the eye changes. It is also known that dilation
changes to reflect processing effort. In psychology, pupil dilation has been studied and
used as a reliable physiological measurement of effort in tasks such as mental
arithmetic and short-term memory (Hess & Polt, 1964; Kahneman, 1973). Pupil dilation
has also been used to measure processing effort during reading comprehension (Just &
Carpenter, 1993).
Explanation of Your Pupil Data:
During the typing task you just completed, the eye tracker was set up to measure
your pupil dilation. The data from your pupil dilation has been computed as a measure
of your processing effort during the task. On the y-axis of the figure below is pupil
diameter; pupil diameter is higher when more effort is being exerted. On the x-axis of
the figure is time. The figure shows the data relating to your processing effort over the 6
minutes of the task. As a comparison, the figure also shows the average processing
effort of participants who have already participated in this study.
Figure A1: False representation of pupil diameter data for participants in the highperceived-effort condition.

Summary of Your Pupil Data:
The pupil data show that you exerted more effort than the average participant on the
typing task. Specifically, you exerted more effort than 81.5% of the other participants.
115

References
Hess, E. H., & Polt, J. M. (1964). Pupil size in relation to mental activity during simple
problem solving. Science, 143, 1190–1192.
Just, M. A., & Carpenter, P. A. (1993). The intensity dimension of thought: Pupillometric
indices of sentence processing. Canadian Journal of Experimental Psychology, 47,
310–339.
Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall.

116

APPENDIX H

STUDY 4 PUPIL DATA REPORT FOR LOW-PERCEIVED-EFFORT PARTICIPANTS
Background on Pupil Dilation:
Dilation of the pupil changes over time. It is commonly known that pupil dilation
changes as the amount of sunlight hitting the eye changes. It is also known that dilation
changes to reflect processing effort. In psychology, pupil dilation has been studied and
used as a reliable physiological measurement of effort in tasks such as mental
arithmetic and short-term memory (Hess & Polt, 1964; Kahneman, 1973). Pupil dilation
has also been used to measure processing effort during reading comprehension (Just &
Carpenter, 1993).
Explanation of Your Pupil Data:
During the typing task you just completed, the eye tracker was set up to measure
your pupil dilation. The data from your pupil dilation has been computed as a measure
of your processing effort during the task. On the y-axis of the figure below is pupil
diameter; pupil diameter is higher when more effort is being exerted. On the x-axis of
the figure is time. The figure shows the data relating to your processing effort over the 6
minutes of the task. As a comparison, the figure also shows the average processing
effort of participants who have already participated in this study.
Figure A2: False representation of pupil diameter data for participants in the lowperceived-effort condition.

Summary of Your Pupil Data:
The pupil data show that you exerted less effort than the average participant on the
typing task. Specifically, 81.5% of the other participants exerted more effort than you.
117

References
Hess, E. H., & Polt, J. M. (1964). Pupil size in relation to mental activity during simple
problem solving. Science, 143, 1190–1192.
Just, M. A., & Carpenter, P. A. (1993). The intensity dimension of thought: Pupillometric
indices of sentence processing. Canadian Journal of Experimental Psychology, 47,
310–339.
Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall.

118

APPENDIX I
48 ANAGRAMS FROM STUDY 4
CSEDOLA (solaced)*
OEFSWLR (flowers, fowlers, reflows)*
RIEVNAG (vinegar)
OIOGBYL (biology)
RATSIID (diarist)*
AREEDMR (dreamer)
OERRFPM (perform)**
AICOLST (stoical)*
UDPIDNG (pudding)
SCTFIHA (catfish)
ULCHEKC (chuckle)
TNHRCIE (cithern)*
AFLNNLE (flannel)
NORCGEA (acrogen)*
FSNAITE (fainest)*
EAUQRRT (quarter)
TPAULAS (spatula)
TRAETCR (retract)
AOBCART (acrobat)
EEDSAEW (seaweed)
INAHPIR (hairpin)
CECRNON (concern)
SHMCUEL (mulches)*
POSTSGI (spigots)*
LEYPSET (steeply)*

PLECINA (capelin, panicle, pelican)*
RAECBEM (embrace)
TNETISD (dentist)
RASMDUT (mustard)
LTEUBLA (bullate)*
GROADNE (groaned)*
IZETICN (citizen)
SCUREED (recused, reduces, rescued,
secured, seducer)*
LLIMNOI (million)***
AEDRNOM (madrone)*
REHGTIF (fighter, freight, refight)
ELASSME (measles)
PEURTMT (trumpet)
GLUAERR (regular)*
NPGRISA (parings, parsing, rasping,
sparing)*
ENELGTD (gentled)*
RSEEPKA (speaker)
ROFAVSL (flavors)*
ETNRPAT (pattern)
EETSRGU (gesture)
IPLSREP (ripples, slipper)
ONTCUSE (contuse)**
CATILPA (capital)

Solutions are in parentheses.
*Signifies solvable anagrams from list in Study 1
**These anagrams were in the list for Study 1 with an extra letter that made them
unsolvable (e.g., in the Study 1 list OERRFPM (perform) was OERIRFPM).
***This anagram was solvable in the list for Study 1 (LSLIMNOI: millions); the “S” was
removed for this list to make it a 7-letter anagram.

119

REFERENCES

120

REFERENCES

References marked with an asterisk indicate studies included in the meta-analysis.
*Alberts, H. J. E. M., Martijn, C., Greb, J., Merckelbach, H., & de Vries, N. K. (2007).
Carrying on or giving in: The role of automatic processes in overcoming ego
depletion. British Journal of Social Psychology, 46, 383-399.
doi:10.1348/014466606X130111
Ashford, S. J., & Northcraft, G. (2003). Robbing Peter to pay Paul: Feedback
environments and enacted priorities in response to competing task demands.
Human Resource Management Review, 13, 537-559.
doi:10.1016/j.hrmr.2003.11.002
Avnet, T., & Higgins, E. T. (2003). Locomotion, assessment and regulatory fit: Value
transfer from "How" To "What". Journal of Experimental Social Psychology, 39,
525-530. doi:10.1016/S0022-1031(03)00027-1
*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. doi:10.1037/0022-3514.74.5.1252
Baumeister, R. F., & Heatherton, T. F. (1996). Self-regulation failure: An overview.
Psychological Inquiry, 7, 1-15. doi:10.1207/s15327965pli0701_1
Baumeister, R. F., Vohs, K. D., & Tice, D. M. (2007). The strength model of self-control.
Current Directions in Psychological Science, 16, 351– 355.
doi:10.1111/j.1467-8721.2007.00534.x
Boksem, M. A. S., Meijman, T. F. & Lorist, M. M. (2005). Effects of mental fatigue on
attention: an ERP study. Cognitive Brain Research, 25, 107-116.
doi:10.1016/j.cogbrainres.2005.04.011
Boksem, M. A. S., Meijman, T. F. & Lorist, M. M. (2006). Mental fatigue, motivation
and action monitoring. Biological Psychology, 72, 123-132.
doi:10.1016/j.biopsycho.2005.08.007
*Bray, S. R., Ginis, K. A. M., Hicks, A. L., & Woodgate, J. (2008). Effects of selfregulatory strength depletion on muscular performance and EMG activation.
Psychophysiology, 45, 337–343. doi:10.1111/j.1469-8986.2007.00625.x
Brehm, J. W., & Self, E. A. (1989). The intensity of motivation. Annual Reviews of
Psychology, 40, 109-131. doi:10.1146/annurev.ps.40.020189.000545

121

Brener, J., & Mitchell, S. (1989). Changes in energy expenditure and work during
response acquisition in rats. Journal of Experimental Psychology: Animal Behavior
Processes, 15,166-175. doi:10.1037/0097-7403.15.2.166
Broadbent, D. E. (1979). Is a fatigue test now possible. Ergonomics, 22, 1277-1290.
doi:10.1080/00140137908924702
*Burkley, E. (2008). The role of self-control in resistance to persuasion. Personality and
Social Psychology Bulletin, 34, 419–431. doi:10.1177/0146167207310458
Carver, C. S., & Scheier, M. F. (1981). Attention and self-regulation: A control-theory
approach to human behavior. New York: Springer-Verlag.
Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. New York,
NY: Cambridge University Press.
*Ciarocco, N. J., Sommer, K. L., & Baumeister, R. F. (2001). Ostracism and ego
depletion: The strains of silence. Personality and Social Psychology Bulletin, 27,
1156–1163. doi:10.1177/0146167201279008
Cleary, T. J. (2006). The development and validation of the Self-Regulation Strategy
Inventory—Self-Report. Journal of School Psychology, 44, 307-322.
doi:10.1016/j.jsp.2006.05.002
Cohen, J. (1987). Statistical power analysis for the behavioral sciences (2nd ed.).
Hillsdale, NJ: Erlbaum.
Converse, P. D., & DeShon, R. P. (2009). A tale of two tasks: Reversing the selfregulatory resource depletion effect. Journal of Applied Psychology, 94, 13181324. doi: 10.1037/a0014604
*DeWall, C. N., Baumeister, R. F., Stillman, T. F., & Gailliot, M. T. (2007). Violence
restrained: Effects of self-regulation and its depletion on aggression. Journal of
Experimental Social Psychology, 43, 62–76. doi:10.1016/j.jesp.2005.12.005
Ebenbach, D. H., & Keltner, D. (1998). Power, emotion, and judgmental accuracy in
social conflict: Motivating the cognitive miser. Basic and Applied Social
Psychology, 20, 7-21. doi:10.1207/s15324834basp2001_2
Eisenberger, R. (1992). Learned industriousness. Psychological Review, 99, 248-267.
doi:10.1037/0033-295X.99.2.248
Faul, F., Erdfelder, E., Lang, A.-G., & Buchner, A. (2007). G*Power 3: A flexible
statistical power analysis program for the social, behavioral, and biomedical
sciences. Behavior Research Methods, 39, 175-191. doi:10.3758/BF03193146

122

*Fennis, B. M., Janssen, L., & Vohs, K. D. (2009). Acts of benevolence: A limitedresource account of compliance with charitable requests. Journal of Consumer
Research, 35, 906–924. doi:10.1086/593291
*Finkel, E. J., & Campbell, W. K. (2001). Self-control and accommodation in close
relationships: An interdependence analysis. Journal of Personality and Social
Psychology, 81, 263–277. doi:10.1037//0022-3514.81.2.263
*Finkel, E. J., Dalton, A. N., Campbell, W. K., Brunell, A. B., Scarbeck, S. J., &
Chartrand, T. L. (2006). High-maintenance interaction: Inefficient social
coordination impairs self-regulation. Journal of Personality and Social Psychology,
91, 456–475. doi:10.1037/0022-3514.91.3.456
Fiske, S. T. (1992). Thinking is for doing: Portraits of social cognition from
Daguerreotype to laserphoto. Journal of Personality and Social Psychology, 63,
877-889. doi:10.1037/0022-3514.63.6.877
Fiske, S. T. (2004). Intent and ordinary bias: Unintended thought and social motivation
create causal prejudice. Social Justice Research, 17, 117-127.
doi:10.1023/B:SORE.0000027405.94966.23
Fiske, S. T., & Taylor, S. E. (1984). Social cognition. Reading, MA: Addison-Wesley.
Fiske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed.). New York: McGraw-Hill.
*Friese, M., Hofmann, W., & Wanke, M. (2008). When impulses take over: Moderated
predictive validity of explicit and implicit attitude measures in predicting food choice
and consumption behaviour. British Journal of Social Psychology, 47, 397–419.
doi:10.1348/014466607x241540
Funder, D. C., & Block, J. (1989). The role of ego-control, ego-resiliency, and IQ in
delay of gratification in adolescence. Journal of Personality and Social
Psychology, 57, 1041-1050. doi:10.1037/0022-3514.57.6.1041
Funder, D. C., Block, J. H., & Block, J. (1983). Delay of gratification: Some longitudinal
personality correlates. Journal of Personality and Social Psychology, 44, 11981213. doi:10.1037/0022-3514.44.6.1198
Gailliot, M. T. (2008). Unlocking the energy dynamics of executive functioning: Linking
executive functioning to brain glycogen. Perspectives on Psychological Science,
3, 245-263. doi:10.1111/j.1745-6924.2008.00077.x
*Gailliot, M. T., & Baumeister, R. F. (2007b). Self-regulation and sexual restraint:
Dispositionally and temporarily poor self-regulatory abilities contribute to failures at
restraining sexual behavior. Personality and Social Psychology Bulletin, 33, 173–
186. doi:10.1177/0146167206293472
123

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. doi:10.1037/0022-3514.92.2.325
Gailliot, M. T., Plant, E. A., Butz, D. A., & Baumeister, R. F. (2007). Increasing selfregulatory strength can reduce the depleting effect of suppressing stereotypes.
Personality and Social Psychology Bulletin, 33, 281-294.
doi:10.1177/0146167206296101
*Gailliot, M. T., Schmeichel, B. J., & Baumeister, R. F. (2006). Self-regulatory processes
defend against the threat of death: Effects of self-control depletion and trait selfcontrol on thoughts and fears of dying. Journal of Personality and Social
Psychology, 91, 49–62. doi:10.1037/0022-3514.91.1.49
Gigerenzer, G., & Goldstein, D. G. (1996). Reasoning the fast and frugal way: Models of
bounded rationality. Psychological Review, 103, 650-669. doi:10.1037/0033295X.103.4.650
Gilhooly, K. J., & Johnson, C. E. (1978). Effects of solution word attributes on anagram
difficulty: A regression analysis. The Quarterly Journal of Experimental
Psychology, 30, 57-70. doi:10.1080/14640747808400654
*Govorun, O., & Payne, B. K. (2006). Ego-depletion and prejudice: Separating
automatic and controlled components. Social Cognition, 24, 111–136.
doi:10.1521/soco.2006.24.2.111
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,
495-525. doi:10.1037/a0019486
Hess, E. H., & Polt, J. M. (1964). Pupil size in relation to mental activity during simple
problem solving. Science, 143, 1190–1192. doi:10.1126/science.143.3611.1190
Hirst, W., & Kalmar, D. (1987). Characterizing attentional resources. Journal of
Experimental Psychology: General, 116, 68-81. doi:10.1037/0096-3445.116.1.68
Hockey, G. R. J. (1997). Compensatory control in the regulation of human performance
under stress and high workload: A cognitive-energetical framework. Biological
Psychology, 45, 73-93. doi:10.1016/S0301-0511(96)05223-4
*Hofmann, W., Rauch, W., & Gawronski, B. (2007). And deplete us not into temptation:
Automatic attitudes, dietary restraint, and self-regulatory resources as
determinants of eating behavior. Journal of Experimental Social Psychology, 43,
497–504. doi:10.1016/j.jesp.2006.05.004

124

Inzlicht, M., & Gutsell, J. N. (2007). Running on empty: Neural signals for self-control
failure. Psychological Science, 18, 933-937.
doi:10.1111/j.1467-9280.2007.02004.x
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,
1686-1693. doi:10.1177/0956797610384745
*Joireman, J., Balliet, D., Sprott, D., Spangenberg, E., & Schultz, J. (2008).
Consideration of future consequences, ego-depletion, and self-control: Support for
distinguishing between CFC-Immediate and CFC-Future sub-scales. Personality
and Individual Differences, 45, 15–21. doi:10.1016/j.paid.2008.02.011
Just, M. A., & Carpenter, P. A. (1993). The intensity dimension of thought: Pupillometric
indices of sentence processing. Canadian Journal of Experimental Psychology,
47, 310–339. doi:10.1037/h0078820
Just, M. A., Carpenter, P. A., & Miyake, A. (2003). Neuroindices of cognitive workload:
Neuroimaging, pupillometric and event-related potential studies of brain work.
Theoretical Issues in Ergonomics Science, 4, 56-88.
doi:10.1080/14639220210159735
Kahneman, D. (1973). Attention and effort. Englewood Cliffs, NJ: Prentice-Hall.
Kahneman, D., & Tversky, A. (1973). On the psychology of prediction. Psychological
Review, 80, 237-251. doi:10.1037/h0034747
Kernan, M. G., & Lord, R. G. (1989). The effects of explicit goals and specific feedback
on escalation processes. Journal of Applied Social Psychology, 19, 1125-1143.
doi:10.1111/j.1559-1816.1989.tb01243.x
Kivetz, R., & Zheng, Y. (2006). Determinants of justification and self-control. Journal of
Experimental Psychology: General, 135, 572-587. doi:10.1037/00963445.135.4.572
Kruglanski, A. W., Thompson, E. P., Higgins, E. T., Atash, M. N., Pierro, A., Shah, J. Y.,
& Spiegel, S. (2000). To “do the right thing” or to “just do it”: Locomotion and
assessment as distinct self-regulatory imperatives. Journal of Personality and
Social Psychology, 79, 793-815. doi:10.1037/0022-3514.79.5.793
Latham, G. P., & Locke, E. A. (1991). Self-regulation through goal setting.
Organizational Behavior and Human Decision Processes, 50, 212-247.
doi:10.1016/0749-5978(91)90021-K

125

*Legault, L., Green-Demers, I., & Eadie, A. L. (2009). When internalization leads to
automatization: The role of self-determination in automatic stereotype
suppression and implicit prejudice regulation. Motivation and Emotion, 33, 10-24.
doi:10.1007/s11031-008-9110-4
Lentricchia, F. (1990). In place of an afterword—someone reading. In F. Lentricchia & T.
McLaughlin (Eds.), Critical terms for literary study (pp. 429). Chicago, IL, USA:
The University of Chicago Press.
Lewis, M. (1965). Psychological effect of effort. Psychological Bulletin, 64,183-190.
doi:10.1037/h0022224
Lock, E. A., Shaw, K. M., Saari, L. M., & Latham, G. P. (1981). Goal setting and task
performance: 1969-1980. Psychological Bulletin, 90, 125-152.
doi:10.1037/0033-2909.90.1.125
Lorist, M. M., Boksem, M. A. S., & Ridderinkhof, K. R. (2005). Impaired cognitive control
and reduced cingulate activity during mental fatigue. Cognitive Brain Research, 24,
199–205. doi:10.1016/j.cogbrainres.2005.01.018
Lorist, M. M., Klein, M., Nieuwenhuis, S., De Jong, R., Mulder, G. & Meijman, T. F.
(2000). Mental fatigue and task control: planning and preparation.
Psychophysiology, 37, 614-622. doi:10.1017/S004857720099005X
MacLeod, C. M. (1991). Half a century of research on the Stroop effect: An integrative
review. Psychological Bulletin, 109, 163-203. doi:10.1037/0033-2909.109.2.163
* Martijn, C., Alberts, H., Merckelbach, H., Havermans, R., Huijts, A., & De Vries, N. K.
(2007). Overcoming ego depletion: The influence of exemplar priming on selfcontrol performance. European Journal of Social Psychology, 37, 231–238.
doi:10.1002/ejsp.350 4089.52.1.17
*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, 441-460. doi:10.1521/soco.20.6.441.22978
Mayer, J. D., & Gaschke, Y. N. (1988). The experience and meta-experience of mood.
Journal of Personality and Social Psychology, 55, 102-111. doi:10.1037/00223514.55.1.102
Meijman, T. F. (1997). Mental fatigue and the efficiency of information processing in
relation to work times. International Journal of Industrial Ergonomics, 20, 31-38.
Mento, A. J., Locke, E. A., & Klein, H. J. (1992). Relationship of goal level to valence
and instrumentality. Journal of Applied Psychology, 77, 395-405.
doi:10.1037/0021-9010.77.4.395
126

Mischel, W., Shoda, Y., & Peake, P. K. (1988). The nature of adolescent competencies
predicted by preschool delay of gratification. Journal of Personality and Social
Psychology, 54, 687-696. doi:10.1037/0022-3514.54.4.687
Muraven, M. (2010). Building self-control strength: Practicing self-control leads to
improved self-control performance. Journal of Experimental Social Psychology,
46, 465-468. doi:10.1016/j.jesp.2009.12.011
Muraven, M. & Baumeister, R. F. (2000). Self-regulation and depletion of limited
resources: Does self-control resemble a muscle. Psychological Bulletin, 126,
247-259. doi:10.1037//0033-2909.126.2.247
Muraven, M., Baumeister, R., & Tice, D. M. (1999). Longitudinal improvement of selfregulation through practice: Building self-control strength through repeated
exercise. Journal of Social Psychology, 139, 446–457.
doi:10.1080/00224549909598404
*Muraven, M., Collins, R. L., & Nienhaus, K. (2002). Self-control and alcohol restraint:
An initial application of the self-control strength model. Psychology of Addictive
Behaviors, 16, 113–120. doi:10.1037/0893-164X.16.2.113
*Muraven, M., Shmueli, D., & Burkley, E. (2006). Conserving self-control strength.
Journal of Personality and Social Psychology, 91, 524-537. doi:10.1037/00223514.91.3.524
*Muraven, M., & Slessareva, E. (2003). Mechanism of self-control failure: Motivation
and limited resources. Personality and Social Psychology Bulletin, 29, 894-906.
doi:10.1177/0146167203253209
*Muraven, M., Tice, D. M. & Baumeister, R. F. (1998). Self-control as limited resource:
Regulatory depletion patterns. Journal of Personality and Social Psychology, 74,
774-789. doi:10.1037/0022-3514.74.3.774
Navon, D. (1984). Resources--a theoretical soup stone? Psychological Review, 91, 216234. doi:10.1037/0033-295X.91.2.216
*Neshat-Doost, H. T., Dalgleish, T., & Golden, A. J. (2008). Reduced specificity of
emotional autobiographical memories following self-regulation depletion. Emotion,
8, 731-736. doi:10.1037/a0013507
Oaten, M., & Cheng, K. (2006a). Improved self-control: The benefits of a regular
program of academic study. Basic and Applied Social Psychology, 28, 1–16.
doi:10.1207/s15324834basp2801_1

127

Oaten, M., & Cheng, K. (2006b). Longitudinal gains in self-regulation from regular
physical exercise. British Journal of Health Psychology, 11, 717–733.
doi:10.1348/135910706X96481
Oaten, M., & Cheng, K. (2007). Improvements in self-control from financial monitoring.
Journal of Economic Psychology, 28, 487–501. doi:10.1016/j.joep.2006.11.003
Oaten, M., Williams, K. D., Jones, A., & Zadro, L. (2008). The effects of ostracism on
self-regulation in the socially anxious. Journal of Social and Clinical Psychology,
27, 471–504. doi:10.1521/jscp.2008.27.5.471
*Oikawa, M. (2005). How do conscious and unconscious goals differ? Suppression of
stereotypes by instructions or priming. The Japanese Journal of Educational
Psychology, 53, 504-515.
Operario, D., & Fiske, S. T. (1999). Social cognition permeates social psychology:
Motivated mental processes guide the study of human social behavior. Asian
Journal of Social Psychology, 2, 63-78. doi:10.1111/1467-839X.00026
*Ostafin, B. D., Marlatt, G. A., & Greenwald, A. G. (2008). Drinking without thinking: An
implicit measure of alcohol motivation predicts failure to control alcohol use.
Behavior Research and Therapy, 46, 1210–1219. doi:10.1016/j.brat.2008.08.003
*Park, S. H., Glaser, J., & Knowles, E. D. (2008). Implicit motivation to control prejudice
moderates the effect of cognitive depletion on unintended discrimination. Social
Cognition, 26, 401– 419. doi:10.1521/soco.2008.26.4.401
*Pocheptsova, A., Amir, O., Dhar, R., & Baumeister, R. F. (2009). Deciding without
resources: Resource depletion and choice in context. Journal of Marketing
Research, 46, 344 –355. doi:10.1509/jmkr.46.3.344
Schellekens, J. M. H., Sijtsma, G. J., Vegter, E. & Meijman, T. F. (2000). Immediate and
delayed after-effects of long lasting mentally demanding work. Biological
Psychology, 53, 37-56. doi:10.1016/S0301-0511(00)00039-9
*Schmeichel, B. J. (2007). Attention control, memory updating, and emotion regulation
temporarily reduce the capacity for executive control. Journal of Experimental
Psychology: General, 136, 241-255. doi:10.1037/0096-3445.136.2.241
*Schmeichel, B. J., Demaree, H. A., Robinson, J. L., & Pu, J. (2006). Ego depletion by
response exaggeration. Journal of Experimental Social Psychology, 42, 95-102.
doi:10.1016/j.jesp.2005.02.005
* Schmeichel, B. J., & Vohs, K. D. (2009). Self-affirmation and self-control: Affirming
core values counteracts ego depletion. Journal of Personality and Social
Psychology, 96, 770–782. doi:10.1037/a0014635

128

*Schmeichel, B. J., Vohs, K. D. & Baumeister, R. F. (2003). Intellectual performance
and ego depletion: Role of the self in logical reasoning and other information
processing. Journal of Personality and Social Psychology, 85, 33-46.
doi:10.1037/0022-3514.85.1.33
*Segerstrom, S. C., & Nes, L. S. (2007). Heart rate variability reflects self-regulatory
strength, effort, and fatigue. Psychological Science, 18, 275–281.
doi:10.1111/j.1467-9280.2007.01888.x
*Shamosh, N. A., & Gray, J. R. (2007). The relation between fluid intelligence and selfregulatory depletion. Cognition & Emotion, 21, 1833–1843.
doi:10.1080/02699930701273658
*Shmueli, D., & Prochaska, J. J. (2009). Resisting tempting foods and smoking
behavior: Implications from a self-control theory perspective. Health Psychology,
28, 300-306. doi:10.1037/a0013826
Solomon, R. I. (1948). The influence of work on behavior. Psychological Bulletin, 45, 140. doi:10.1037/h0055527
*Stewart, C. C., Wright, R. A., Hui, S.-K. A., & Simmons, A. (2009). Outcome
expectancy as a moderator of mental fatigue influence on cardiovascular
response. Psychophysiology, 46, 1141–1149.
doi:10.1111/j.1469-8986.2009.00862.x
*Stucke, T. S., & Baumeister, R. F. (2006). Ego depletion and aggressive behavior: Is
the inhibition of aggression a limited resource? European Journal of Social
Psychology, 36, 1–13. doi:10.1002/ejsp.285
Tice, D. M., Baumeister, R. F., Shmueli, D., & Muraven, M. (2007). Restoring the self:
Positive affect helps improve self-regulation following ego depletion. Journal of
Experimental Social Psychology, 43, 379-384. doi:10.1016/j.jesp.2006.05.007
*Tyler, J. M. (2008). In the eyes of others: Monitoring for relational value cues. Human
Communication Research, 34, 521–534. doi:10.1111/j.1468-2958.2008.00331.x
*Tyler, J. M., & Burns, K. C. (2008). After depletion: The replenishment of the self’s
regulatory resources. Self and Identity, 7, 305–321.
doi:10.1080/15298860701799997
*Tyler, J. M., & Burns, K. C. (2009). Triggering conservation of the self’s regulatory
resources. Basic and Applied Social Psychology, 31, 255–266.
doi:10.1080/01973530903058490

129

Vancouver, J. B., Weinhardt, J. M., & Schmidt, A. M. (2010). A formal, computational
theory of multiple-goal pursuit: Integrating goal-choice and goal-striving processes.
Journal of Applied Psychology, 95, 985-1008. doi:10.1037/a0020628
van der Linden, D., & Eling, P. (2006). Mental fatigue disturbs local processing more
than global processing. Psychological Research, 70, 395-402.
doi:10.1007/s00426-005-0228-7
van der Linden, D., Frese, M. & Meijman, T. F. (2003a). Mental fatigue and the control
of cognitive processes: Effects on perseveration and planning. Acta Psychologica,
113, 45-65. doi:10.1016/S0001-6918(02)00150-6
van der Linden, D., Frese, M. & Sonnentag, S. (2003b). The impact of mental fatigue on
exploration in a complex computer task: rigidity and loss of systematic strategies.
Human Factors, 45, 483-494. doi:10.1518/hfes.45.3.483.27256
van der Linden, D., Massar, S. A. A., Schellekens, A. F. A., Ellenbroek, B. A., & Verkes,
R-J. (2006). Disrupted sensorimotor gating due to mental fatigue: Preliminary
evidence. International Journal of Psychophysiology, 62, 168-174.
doi:10.1016/j.ijpsycho.2006.04.001
Vohs, K. D., & Baumeister, R. F. (2004). Understanding self-regulation: An introduction.
In R. F. Baumeister & K. D. Vohs (Eds.), Handbook of self-regulation: Research
theory, and applications (pp. 1-9). New York, NY, US: Guilford Press.
*Vohs, K. D., Baumeister, R. F., & Ciarocco, N. J. (2005). Self-regulation and selfpresentation: Regulatory resource depletion impairs impression management and
effortful self-presentation depletes regulatory resources. Journal of Personality and
Social Psychology, 88, 632– 657. doi:10.1037/0022-3514.88.4.632
*Vohs, K. D., & Faber, R. J. (2007). Spent resources: Self-regulatory resource
availability affects impulse buying. Journal of Consumer Research, 33, 537–547.
doi:10.1086/510228
Vohs, K. D. & Heatherton, T. F. (2000). Self-regulatory failure: A resource-depletion
approach. Psychological Science, 11, 249-254. doi:10.1111/1467-9280.00250
*Vohs, K. D., & Schmeichel, B. J. (2003). Self-regulation and extended now: Controlling
the self alters the subjective experience of time. Journal of Personality and Social
Psychology, 85, 217-230. doi:10.1037/0022-3514.85.2.217
*Wan, E. W., & Sternthal, B. (2008). Regulating the effects of depletion through
monitoring. Personality and Social Psychology Bulletin, 34, 32–46.
doi:10.1177/0146167207306756

130

*Webb, T. L. & Sheeran, P. (2003). Can implementation intentions help to overcome
ego depletion. Journal of Experimental Social Psychology, 39, 279-286.
doi:10.1016/s0022–1031(02)00527-9
Wegner, D. M. (1994). Ironic processes of mental control. Psychological Review, 101,
34-52. doi:10.1037/0033-295X.101.1.34
*Wheeler, S. C., Brinol, P., & Hermann, A. D. (2007). Resistance to persuasion as selfregulation: Ego-depletion and its effects on attitude change processes. Journal of
Experimental Social Psychology, 43, 15-156. doi:10.1016/j.jesp.2006.01.001
Wickens, C. D. (1980). The structure of processing resources. In R. Nickerson (Ed.),
Attention and performance VIII. (pp. 239-257). New York: Erlbaum.
Wilson, M. (1988). MRC psycholinguistic database: Machine-usable dictionary,
version 2.00. Behavioural Research Methods, Instruments and Computers, 20, 610. doi:10.3758/BF03202594
*Wright, R. A., Junious, T. R., Neal, C., Avello, A., Graham, C., Herrmann, L., Junious,
S., . . . Walton, N. (2007). Mental fatigue influence on effort-related cardiovascular
response: Difficulty effects and extension across cognitive performance domains.
Motivation and Emotion, 31, 219–231. doi:10.1007/s11031-007-9066-9
*Wright, R. A., Martin, R. E., & Bland, J. L. (2003). Energy resource depletion, task
difficulty, and cardiovascular response to a mental arithmetic challenge.
Psychophysiology, 40, 98–105. doi:10.1111/1469-8986.00010
*Wright, R. A., Stewart, C. C., & Barnett, B. R. (2008). Mental fatigue influence on effortrelated cardiovascular response: Extension across the regulatory (inhibitory)/nonregulatory performance dimension. International Journal of Psychophysiology, 69,
127–133. doi:10.1016/j.ijpsycho.2008.04.002
Zachary, R. A. (1986). Shipley Institute of Living Scale: Revised manual. Los Angeles,
CA: Western Psychological Services.

131