Basic and Applied Social Psychology

ISSN: 0197-3533 (Print) 1532-4834 (Online) Journal homepage: https://www.tandfonline.com/loi/hbas20

Can Ego-Depletion Be Helpful? Testing the Process

Model Implication That Ego-Depletion Reduces
Irrational Persistence

Jeffrey M. Osgood

To cite this article: Jeffrey M. Osgood (2018) Can Ego-Depletion Be Helpful? Testing the Process
Model Implication That Ego-Depletion Reduces Irrational Persistence, Basic and Applied Social
Psychology, 40:3, 161-170, DOI: 10.1080/01973533.2018.1449748

To link to this article: https://doi.org/10.1080/01973533.2018.1449748

Published online: 05 Apr 2018.

Submit your article to this journal

Article views: 406

View related articles

View Crossmark data

Citing articles: 1 View citing articles

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=hbas20
BASIC AND APPLIED SOCIAL PSYCHOLOGY
2018, VOL. 40, NO. 3, 161–170
https://doi.org/10.1080/01973533.2018.1449748

none defined

Can Ego-Depletion Be Helpful? Testing the Process Model Implication

That Ego-Depletion Reduces Irrational Persistence
Jeffrey M. Osgood
Florida Gulf Coast University

ABSTRACT
The present research investigated the effect of state self-control on irrational persistence, which
refers to pursuing a course of action beyond when the effort can no longer be justified given the
cost of persisting and/or probability of success. In 3 studies, ego-depletion reduced the level of
irrational persistence displayed by participants. In Study 1, ego-depleted participants were less
tolerant of ineffective default computer settings that wasted their time. In Study 2, ego-depleted
participants solved more anagrams than nondepleted participants when given only a limited
amount of time to solve anagrams that ranged from easy to difficult and skipping was allowed.
Study 3 conceptually replicate the effect of Study 2 but produced a smaller effect size.

In his historic 1941 speech at the Harrow School during rewards and/or probability of success. This is tanta-
the height of Nazi Germany’s offensive against Britain, mount to irrational persistence. Irrational persistence
Winston Churchill urged his people to “never give in. may occur in daily activities such as when a
Never, never, never, in nothing great or small, large or student refuses to stop studying and sleep, despite the
petty, never give in” (Churchill, 2003 p. 306). Such urgings of his or her body, or a graduate student who
quotes are common on motivational posters and in half- relentlessly pursues a fruitless line of research.
time speeches in sports movies. However, research in The psychological literature lacks a unified vein of
social psychology suggests that “never giving in” can research devoted to irrational persistence, but examples
be quite difficult. For example, the popular, albeit con- of similar ideas exist across the discipline. Perhaps the
troversial, strength theory of self-control (Muraven & most well-known example of irrational persistence is
Baumeister, 2000) proposes that a person’s motivation the sunk-cost fallacy—individuals choose to continue
to exert self-control temporarily wanes after use; the investing time or resources in something that clearly
phenomenon is called “ego depletion.” Although the is not working after making an initial investment (Arkes
implications of ego depletion are often bemoaned, a & Blumer, 1985). The motivation in sunk-cost is often
natural limit to one’s ability to easily persist using to try to recoup the lost investment (Arkes & Blumer,
self-control may not always be a detriment. Indeed, less 1985). More recently, Koole and colleagues introduced
frequently included on inspirational posters is the next the idea of ego fixation (Koole et al., 2014). Ego fixation
phrase of Winston Churchill’s speech: “Never give in … is defined as the involuntary persistence of self-control
except to convictions of honor, and good sense.” It is processes and is hypothesized to alienate one from
doubtful that ego depletion could lead directly to “good their emotions and cause an involuntary inhibition of
sense,” but the present research will test if ego depletion emotional preferences. Another related concept is
may lead to more productive choices in certain hyperopia (see Kivetz & Simonson, 2002), which refers
circumstances. to the pursuit of long-term goals at the expense of
Effective decision making requires balancing both immediate needs or desires to an extent that creates
grit and time management. On one hand, there are significant regret later. For example, Keinan and Kivetz
salient examples in which people give in too quickly, (2008) interviewed college students and alumni about
such as the fickle dieter or an athlete who quits a train- either a recent school vacation or a school vacation from
ing routine because it is uncomfortable. Less examined a long time ago (either 1 year or 40 years earlier).
is the idea that giving in is sometimes a wise course of Students responded that they wish they had spent more
action, such as when the effort or resources required time on work (over pleasure) when asked about a
to persist are no longer justifiable given the potential recent school vacation. However, students and alumni

CONTACT Jeffrey M. Osgood [email protected] Florida Gulf Coast University, 10501 FGCU Blvd. South, Fort Myers, FL 33965-6565.
© 2018 Taylor & Francis
162 J. M. OSGOOD

regretted using too much self-control (i.e., choosing pool is spent, one loses the ability to exert self-control.
work over pleasure) when asked about school vacations Subsequent attempts were made by researchers to
either 1 year or 40 years earlier, respectively. This identify the nature of this resource. The most well-
suggests that people may sometimes prioritize self- known attempt was conducted by Gailliot et al. (2007)
control goals (i.e., long-term goals over immediate and proposed that the limited resource may be brain
pleasure) to an extent that they regret later. glucose. However, the limited resource account is
The existing research literature suggests many difficult to reconcile with other research (see Lange &
reasons why a person might persist irrationally. One Eggert, 2014; Osgood, 2017). Specifically, critics have
reason is that quitting may evoke negative emotions lamented that the glucose explanation has been difficult
such as shame (Nelissen, Breugelmans, & Zeelenberg, to replicate (e.g., Lange & Eggert, 2014) and seems
2013). For instance, people are socialized from a young implausible with what biologists have discovered about
age to resist giving up easily (Fox & Calkins, 2003). brain metabolism (Kurzban, 2010).
Indeed, many adages convey this message (e.g., “When A further criticism is that ego depletion is moderated
the going gets tough, the tough get going”). Further, by many manipulations that do not involve altering a
many formative environments of childhood (e.g., school mental resource including lay beliefs about self-control
& sports) are designed to reward hard work and persist- (Job, Dweck, & Walton, 2010), positive mood (Tice,
ence, but perhaps are not designed to teach young Baumeister, Shmueli, & Muraven, 2007), monetary
people how to determine in which tasks it is worth incentives (Muraven & Slessareva, 2003), and brief
persisting. Consequently, people may have learned exercise (Osgood, 2015) to name a few (for more, see
positive associations with persistence and negative Osgood & Salamone, 2016). Thus, the cause of this
associations with giving up during their development. effect does not seem to be a depleted physical resource
In fact, Staw (1981) observed that social norms exist within the brain. Further the reliability and robustness
such that an individual will often feel motivated to of the ego depletion effect has been challenge by recent
persist in a task until completion, even in the presence high-profile failures to replicate (see Hagger et al.,
of negative feelings and when progress is not being 2016). Even though the methods and analytical techni-
made on the task. Therefore, when experiencing the ques of these failed replications have themselves been
negative affect associated with a lack of progress, criticized (see Cunningham & Baumeister, 2016; Dang,
individuals may feel it would be unwise, socially 2018). Such multidirectional criticism of ego depletion
inappropriate, or even immoral to succumb to these suggests that the originally proposed model of ego
emotions. Ironically, some evolutionary psychologists depletion may need revising or even abandonment. It
argue that emotional reactions such as frustration and is likely that forcing participants to complete a task that
boredom were reproductively advantageous because is unpleasant and requires a large exertion of self-
they led to better decision making in an ancestral control will affect them either cognitively or affectively
context (Kurzban, Duckworth, Kable, & Myers, 2013). for a short time afterward. Less certain is whether this
Persisting in the face of such negative emotions requires subsequent effect is diminished self-control. Thus, the
exerting self-control (Tice & Bratslavsky, 2000). “ego depletion effect” may be just one manifestation
Ego depletion theory proposes that the motivation to of the cognitive or motivational processes altered by
exert effortful control diminishes with use (Muraven & an unpleasant exertion of self-control.
Baumeister, 2000), particularly in the face of negative Inzlicht and Schmeichel (2012) and Inzlicht,
emotions (Tice & Bratslavsky, 2000). Indeed, ego Schmeichel, and Macrae (2014) proposed an alternative
depletion may cause people to experience greater explanation of ego depletion called the “process model,”
sensitivity to boredom (Osgood, 2015). Although the which proposes that as people use self-control, their
emotional impulsivity engendered by ego depletion is motivation to continue using self-control wanes and
generally undesirable, it may confer an advantage in the individual shifts his or her attention to other desires
contexts where such emotions are conducive to effective and/or goals. This theory draws on evolutionary psy-
decision making. The purpose of this article is to test the chology to explain why ego depletion occurs. Specifi-
hypothesis that ego depletion may operate as a natural cally, the process model argues that natural selection
circuit breaker against persisting irrationally. would favor mechanisms that reduce the opportunity
The classic explanation of ego depletion theorizes cost associated with failing to disengage from a task
that all acts of self-control draw from a limited mental that is not worth the effort (Kurzban et al., 2013). Proxi-
resource (called ego strength or self-control strength) that mally, the process model argues that effortful control is
becomes exhausted through strong exertions of self- inherently unpleasant and becomes more unpleasant the
control (Muraven & Baumeister, 2000). As this resource longer one is engaged in using self-control, particularly
 BASIC AND APPLIED SOCIAL PSYCHOLOGY 163

if progress is not being made toward a goal. Thus, the Participants would have a fixed amount of time to read
process model suggests that ego depletion may be adapt- each article displayed to them, after which the computer
ive in forcing individuals to be more efficient in how they would move on to the next one. However, participants
use their time and effort when working on a task. If were also told that they could start selecting topics for
noticeable progress is not being made, ego depletion themselves by pressing a button visible on the computer
may encourage the individual to try doing something else. screen. This button was available on a waiting screen
Inzlicht and colleagues’ (Inzlicht & Schmeichel, 2012; between each article as well as prior to the first article.
Inzlicht et al., 2014) theory fits existing research; The computer was rigged to select only issues expected
however, the implication that ego depletion may lead to be unimportant to students (e.g., type of salt on the
to better decision making in contexts where unfettered road, janitorial policies, squirrel population on campus)
persistence is not the optimal strategy remains untested. while ignoring more relevant topics (e.g., tuition
Although the present research does not test the process increases and course offerings). The computer secretly
model directly, it does draw from the model in making recorded how many unimportant issues the participants
its prediction. The present research tests the hypothesis read before selecting to override the computer and
that ego depletion will lead to better decision making begin making their own choices. Participants were told
and/or performance in contexts where persistence in that the point of the task was for them to review topics
one’s initial course of action is not the best strategy. they cared about. Although not explicitly stated to part-
icipants, the best strategy was to oust the computer as
Study 1 quickly as possible so to respond to as many topics of
interest as possible in the limited time available. Thus,
Methods the true dependent variable in the study was how long
Participants participants would patiently tolerate the computers
Fifty participants (56% female) were recruited from poor default settings before electing to choose which
the undergraduate research pool at a university in the issues to read for themselves. Following this activity,
United States. Participants were randomly assigned to participants completed a last set of questionnaires, were
either the ego depletion or nondepletion manipulation. probed for deception, debriefed, and exited the lab.
A fire alarm drill occurred in the building during one
participant’s appointment. As such, this participant Materials and manipulations
was unable to complete the experiment. Thus, data were White bear task. Participants assigned to the ego
analyzed for the remaining 49 participants. depletion condition were instructed to write a free associ-
ation of their thoughts for 5 min while not thinking about
Procedure a white bear. The act of trying to inhibit a specific thought
All participants were administered informed consent or concept (e.g., white bear) ironically makes it more dif-
and given an overview of the study. Participants in the ficult to do and requires ample self-control (Wegner,
ego depletion condition then completed a white bear 1992). This task has been used in prior ego depletion
thought suppression task, whereas participants in the research and has been shown to require a large exertion
nondepletion condition completed arithmetic problems. of self-control (e.g., Osgood & Muraven, 2015, 2016).
Immediately following this, participants completed a
short manipulation check question and the Brief Mood Arithmetic problems. Participants in the nondepletion
Introspection Scale (BMIS; Mayer & Gaschke, 1988). condition solved moderately difficult math problems
Participants then proceeded to the dependent measure (multiplication) for 5 min. There were more problems
task. For this task, participants were falsely told they than the participant could solve in the given time, and
would provide feedback to university officials on issues participants were instructed to do their best but not to
facing the university. Participants were given a list of 21 worry if they were unable to solve most of the problems.
potential issues (participants retained this list for the Previous research has demonstrated that this is an
duration of the experiment) they might review and were appropriate neutral activity in ego depletion research,
told that they would have time to review only nine as this requires less self-control than the white bear task
issues and would be allowed to submit opinions only while being approximately equivalent in difficulty (e.g.,
on issues they reviewed during the study. The parti- Baumeister, 2003).
cipants were told that the computer would automati-
cally select an article related to an issue for them (one Brief mood introspection scale. The BMIS is a 16-item
at a time) to review from the list they were given based scale (a ¼ :84Þthat assesses one’s current mood (Mayer
on what the computer expects is important to students. & Gaschke, 1988).
164 J. M. OSGOOD

Manipulation check. Following the experimental from analyses. Therefore, data from 85 participants
manipulation (white bear/arithmetic), participants were analyzed.
responded to the following item on an 11-point Likert
scale from 1 (very little) to 11 (very much): “How much Procedure
were you fighting against an urge during the previous Participants were administered informed consent and
task?” This was intended to assess the use of self-control led to a computer to complete the remainder of the
during the manipulation. study. Participants were randomly assigned to complete
either an ego-depleting typing task or a nondepleting
Stimuli (university issues list). This was the list of 21 version of the same typing task. Next, the experimenter
issues that the participant (or computer) could select explained the anagram-solving task. Participants
to review and provide feedback on to university received one of two sets of instructions that were
officials. The list for students to review contained some identical except that in one version (high awareness of
important/interesting topics (e.g., tuition and course opportunity cost) participants were told that the com-
offerings) and some issues less relevant to students puter bank of anagrams contained a mix of both easy
(e.g., changes to janitorial policies and brand of salt to and difficult anagrams, whereas in the other version
be used on university roads during winter). (low awareness of opportunity cost) participants were
not informed of this. Thus, the study had a 2 (ego
depletion/no depletion) � 2 (high awareness/low aware-
Results ness) design. Participants were instructed to solve as
many anagrams as possible in the allotted time and were
Manipulation check
told they may skip as many as they like (no penalty for
Participants assigned to the white bear thought skipping). Participants were not told exactly how much
suppression condition reported more fighting against time they would have but were told they would be given
an urge during the experimental manipulation “around five minutes” (actually 4 min) and that the
(M ¼ 7.3, SD ¼ 3.6) than those assigned to solve experimenter would tell them when to begin and when
arithmetic problems (M ¼ 4.5, SD ¼ 3.4), d ¼ .80. to stop.

Materials and manipulations

Baseline mood (BMIS) Typing task. Participants were asked to retype a passage
The BMIS was scored for pleasant/unpleasant and about a neutral topic (Brazil nuts) without looking at the
uncovered no starting differences between the ego keyboard. All participants were instructed to retype the
depletion (M ¼ 6.2, SD ¼ 6.3) and nondepletion passage as quickly and accurately as possible. Parti-
(M ¼ 6.7, SD ¼ 6.0) groups prior to completing the cipants in the nondepletion condition were given no
university issues task (d ¼ −.08). further instructions. Participants in the ego depletion
condition were given the additional instructions to avoid
using the “n,” “a,” spacebar, and backspace keys.
University issues task Previous research has found that this is an effective
Ego-depleted participants (M ¼ 4.8 articles, SD ¼ 3.1) method for exerting self-control (to stop oneself from
overrode the computer sooner (read fewer unimportant pressing the illegal keys; Janssen, Fennis, & Pruyn, 2010).
articles on average) than nondepleted participants
(M ¼ 6.6 articles, SD ¼ 3.0) producing a medium effect Anagram task. Participants were presented with
size (d ¼ −.59). anagrams on the computer screen one at a time. Parti-
cipants were asked to unscramble the letters such that
all the letters were used to form one word (multiple
Study 2 answers would not be counted more than once; the
participant was informed of this). Participants were
Methods
informed that scoring would be based on the overall
Participants number of anagrams that they solved correctly and that
Eighty-six undergraduate students (56.4% female) from they may skip as many anagrams as they want but they
a university in the northeastern United States completed cannot go back once they skip. The computer contained
this experiment for partial course credit. One more than 100 anagrams so that participants would
participant was a non-native English speaker who did not run out. The first four anagrams were always very
not understand the study instructions and was dropped difficult (the first anagram was impossible). This was
 BASIC AND APPLIED SOCIAL PSYCHOLOGY 165

done to create a situation where simple persistence

would result in wasted time and a lower score. After
this, the remaining anagrams varied in difficulty from
very easy (e.g., “pcu” becomes “cup”) to very difficult
(e.g., “gesandio” becomes “diagnose”). Participants were
told they may skip as many anagrams as they wanted
without penalty and that scoring was based on the total
number of correctly solved anagrams.

Results
Number of anagrams solved Figure 2. Use of time on difficult anagrams (Study 2). Note.
Error bars are standard errors.
Participants in the key-restricted typing condition (ego
depletion) who were aware of the difficulty range solved
Use of time during task
the same mean number of anagrams (M ¼ 8.3,
SD ¼ 5.8), as participants in the key-restricted typing Mean time spent on the difficult anagrams was strongly
condition who were unaware of the difficulty range and negatively correlated with overall number of ana-
(M ¼ 8.3, SD ¼ 5.2). Participants in the normal typing grams solved (r ¼ −.74), the percentage of attempted
condition who were unaware of the difficulty range anagrams correctly solved (r ¼ −.51), and the number
solved slightly more anagrams (M ¼ 6.8, SD ¼ 5.0) than of anagram puzzles viewed (reached; r ¼ −.77).
participants in the normal typing condition who were Key-restricted typing (ego-depleted) participants who
aware of the difficulty range (M ¼ 6.2, SD ¼ 4.6). Over- were aware of the difficulty range spent the least amount
all, participants in the key-restricted typing condition of time on difficult anagrams (M ¼ 36.3 s, SD ¼ 20.7 s),
solved more anagrams (M ¼ 8.3, SD ¼ 5.5) than normal followed by key-restricted typing participants who were
typing participants (M ¼ 6.4, SD ¼ 4.7), d ¼ .37. unaware of the difficulty range (M ¼ 41.4 s, SD ¼ 31.5 s),
Participants who were told that the puzzles would vary followed by normal typing (nondepleted) participants
in difficulty actually solved slightly fewer anagrams on who were unaware of the difficulty range (M ¼ 44.6 s,
average (M ¼ 7.1, SD ¼ 5.2) than those not told that SD ¼ 28.2 s) and normal typing participants who were
the puzzles would range in difficulty (M ¼ 7.5, aware of the difficulty range (M ¼ 51.3 s, SD ¼ 46.0 s).
SD ¼ 5.1), d ¼ .07. Means and standard errors are Overall, participants who completed the key-restricted
displayed in Figure 1. The interaction of typing con- version of the typing task spent less time on difficult
dition and awareness condition produced no effect anagrams (M ¼ 38.7 s, SD ¼ 26.0 s) than participants
ðg2 ¼ :001Þ: Participants in the key-restricted typing who completed the normal typing version of the task
condition also solved a higher percentage of the puzzles (M ¼ 48.3 s, SD ¼ 38.7 s), d ¼ .29. Participants who were
they reached (M ¼ 48.2%, SD ¼ 18.6) than participants aware of the difficulty range spent slightly more time
in the normal typing condition (M ¼ 41.7%, SD ¼ 23.8), on difficult anagrams (M ¼ 44.3 s, SD ¼ 36.9 s) than
d ¼ .30. participants not informed about the difficulty range
(M ¼ 43.1 s, SD ¼ 29.4 s), d ¼ −.03. The interaction of
typing condition and awareness condition produced no
effect (η2 ¼ .008; see Figure 2). Overall, participants in
the key-restricted typing condition reached (viewed)
more anagrams (M ¼ 14.4, SD ¼ 7.8) than participants
normal typing condition (M ¼ 11.8, SD ¼ 7.2), d ¼ .34.

Study 3
Methods
Participants
Eighty-nine undergraduate students (53.9% female)
Figure 1. Mean number of anagrams solved per condition from a university in the northeastern United States
(Study 2). Note. Error bars are standard errors. completed this experiment for partial course credit.
166 J. M. OSGOOD

Data for two participants failed to collect during the interviewed and avoid looking at the words. This
experiment due to computer problems. Therefore, data required the individual to exert effortful control to
from 87 participants were analyzed. maintain focus on the woman while ignoring the dis-
tracting changing stimuli on another part of the screen.
Procedure
Participants completed the experiment individually at Anagram task. The anagram task was similar to that
computers separated by partitions in groups of one to used in Study 2, with the exception that the anagram
six. Participants’ cell phones were confiscated during set the participants worked from contained either a
the study to limit distractions. Upon arriving at the mixture of easy and difficult anagrams or exclusively
lab, participants were administered informed consent difficult anagrams. Participants were not informed
and then given an overview of the study. Participants about the composition of the anagram set beforehand.
were told that they would be solving anagrams and that
their job was to solve as many anagrams as possible
within a fixed amount of time (“less than five minutes”; Results
actually 4 min). They were told that all anagrams were Manipulation checks
equally important, that they could skip as many times
as they wanted without penalty, and that the computer In reference to the video attention control task, parti-
had many anagrams so they would not run out. Parti- cipants in the ego depletion condition reported exerting
cipants were told that they would first complete a visual more effort to control thoughts (M ¼ 6.6, SD ¼ 2.5) than
attention task and answer some questions before con- nondepleted participants (M ¼ 5.7, SD ¼ 3.0), d ¼ .33;
tinuing on to the anagram task. The video was followed slightly more effort to control themselves (M ¼ 6.4,
by several manipulation-check questions, then the SD ¼ 2.7, vs. M ¼ 5.8, SD ¼ 3.0), d ¼ .21; greater
anagram task. For half of the participants, the anagram fighting of an urge (M ¼ 6.1, SD ¼ 3.3, vs. M ¼ 5.2,
set contained only difficult anagrams, whereas for the SD ¼ 3.3), d ¼ .27; and that the task was more challenging
other half of the participants, the anagram set contained (M ¼ 5.5, SD ¼ 3.2 vs. M ¼ 3.3, SD ¼ 3.0), d ¼ .71.
a mixture of easy and difficult anagrams. All participants
were instructed to attempt to solve as many as possible.
Anagram task
Materials and manipulations Time spent on each anagram was negatively correlated
Manipulation check with total anagrams solved for both the difficult-only
Following the video attention task, participants were (r ¼ −.33) and mixed-difficulty (r ¼ −.73) anagram sets.
asked to respond on a 7-point Likert scale from 1 (none) Participants who worked from the mixed-difficulty
to 7 (very much) to the following questions to assess anagram set solved considerably more anagrams
their use of self-control: “How much effort did you exert (M ¼ 11.7, SD ¼ 4.1) than those who worked from the
to control distracting thoughts while working on this difficult-only set (M ¼ 1.8, SD ¼ 1.3), producing a very
task?”; “How much did you have to control yourself large effect (d ¼ 3.26). For those who solved exclusively
while working on the task?”; “How much were you difficult anagrams, participants in the video attention
fighting an urge during the task?”; and “How challeng- control (ego depletion) condition solved slightly more
ing was the task?” anagrams (M ¼ 1.86, SD ¼ 1.46) than participants in
the normal video watching (nondepletion) condition
Video attention control task. The video attention (M ¼ 1.68, SD ¼ 1.25), producing a very small effect
control task used in this experiment was developed by (d ¼ .13). For those who solved mixed-difficulty
Gailliot et al. (2007) to induce ego depletion. During anagrams, participants in the video attention control
the task, participants watched a silent 7-min clip of a condition solved more anagrams (M ¼ 12.29,
woman being interviewed. The view of the woman SD ¼ 4.39) than participants in the normal video watch-
being interviewed appeared on the left side of the screen ing condition (M ¼ 11.23, SD ¼ 3.84), producing a small
while a series of words (each presenting for about 10 s) effect (d ¼ .26). No interaction between video condition
appeared on the lower right corner of the screen. and anagram set condition was observed (g2 ¼ .005).
Participants in the nondepletion condition were told Means and standard errors are displayed on Figure 3.
to watch this video normally as if they were watching Regarding the use of time on the anagram task, for
TV at home. In the attention control (ego depletion) those who solved exclusively difficult anagrams, parti-
condition, participants were instructed to focus their cipants in the video attention control condition used
visual attention exclusively on the woman being slightly less time per anagram (M ¼ 28.7 s, SD ¼ 14.6 s)
 BASIC AND APPLIED SOCIAL PSYCHOLOGY 167

Although the hypotheses tested in this article were

derived from the process model, the proximal mechan-
isms of the process model (that exerting self-control is
unpleasant and drives a redirection of goal-related
attention) were not directly tested. Thus, it is important
to consider other theories that fit the results found in
the present research. For instance, Masicampo and
Baumeister (2012) found that strong goal commitment
caused individuals to miss nonperseverative opportu-
nities to reach their goal (an unexpected method of
reaching their goal). The present findings could be
explained by ego depletion reducing goal commitment.
Lower goal commitment could mean that people would
Figure 3. Mean number of anagrams solved by condition override the default computer settings sooner (Study 1),
(Study 3). Note. Error bars are standard errors. spend less time on difficult anagrams, and solve more
mixed anagrams by switching from hard ones (Studies
than participants in the normal video viewing condition 2 and 3). Thus, the underlying mechanism of the
(M ¼ 33.0 s, SD ¼ 16.3 s), producing a small effect present findings could be reduced goal commitment.
(d ¼ .27). Similarly, for those who solved mixed-difficult Future research should seek to measure or manipulate
anagrams, participants in the video attention control goal commitment alongside ego depletion on dependent
condition spent less time attempting to solve each measures such as those used in the present research.
anagram (M ¼ 14.0 s, SD ¼ 4.9 s) than participants The results presented here have special implications
who viewed the video normally (M ¼ 15.1 s, SD ¼ 8.8 s), for certain theories of self-control. In particular, the
producing a very small effect (d ¼ .15). process model effectively integrates ego depletion into
a larger model of decision making without resorting
to the tenuous position that ego depletion is reducible
Discussion
to an unavoidable natural limit. Despite this, the present
The present research represents one of only a handful of findings suggest that the process model needs further
investigations to report a situation where ego depletion refinement and empirical testing. On one hand,
improves decision making and/or outcomes on a ego-depleted participants persisted less when it would
performance-based measure. The present research was be imprudent to do so than nondepleted participants.
designed to test the hypothesis that ego depletion leads This is consistent with the process model’s suggestion
to better use of time in contexts where persistence is a that ego depletion may serve the purpose of protecting
maladaptive strategy. It was predicted that this effect the individual from opportunity cost. On the other
would be particularly strong when the opportunity cost hand, attempts to adjust perceived opportunity cost by
associated with persistence is high. This hypothesis either manipulating participants’ awareness of the
is derived from a revised theory of ego depletion opportunity cost (Study 2) or by varying opportunity
developed by Inzlicht and Schmeichel (2012) and cost (Study 3) did not meaningfully interact with ego
Inzlicht, Schmeichel, and Macrae (2014), which suggests depletion. Indeed, in Study 2, participants did not
that ego depletion may prevent overengaging in actions appear to be affected by the awareness manipulation
at the cost of missing opportunities elsewhere. at all. Thus, the effect of perceived opportunity cost
From a behavioral standpoint, the results of the on decision making while depleted will require further
studies reported here are generally supportive of this investigation. However, participants in Studies 2 and 3
theory. Across three studies, ego-depleted participants were not offered incentives to do well; thus, overall
made better use of their time and/or outperformed motivation may have been too low for the opportunity
nondepleted participants. In Study 1, ego-depleted part- cost manipulations to have an effect. Future research
icipants wasted less time before excusing an ineffective should vary the progress being made on tasks such as
computer decision maker. In Studies 2 and 3, ego- those used here and look for an interaction with ego
depleted participants used their time more efficiently depletion or provide tangible incentives (like cash
to solve more anagrams than nondepleted participants payments) for better performance.
by skipping difficult anagrams and solving more easy The results of the present research can be contrasted
ones. In addition, each of these effects were produced with those of previous studies that have found that ego
using a different ego depletion procedure. depletion worsens performance on difficult but solvable
168 J. M. OSGOOD

anagrams (e.g., Baumeister, Bratslavsky, Muraven, & For instance, perhaps certain socialization forces
Tice, 1998; Boucher & Kofos, 2012). Contrary to those experienced during childhood development may cause
prior studies, ego-depleted participants did not perform children to use self-control indiscriminately and impru-
worse when trying to solve exclusively difficult (but dently as adults. Finally, this research may provide
solvable) anagrams in Study 3. However, ego-depleted important insights for clinical practitioners and applied
participants did persist less time (skipped anagrams industrial and organizational psychologists. Clinicians
more quickly) on these anagrams on average, which is who study disorders involving acts of unhealthy persist-
consistent with prior research. Although this was not ent actions such as obsessive compulsive disorder may
the main focus of the present investigation, this incon- find this research informative for their theories. Indus-
sistency with previous findings adds to the ongoing trial/organizational psychologists seeking to improve
discussion on the reliability of ego depletion effects. worker performance using self-control training or
Few previous findings support the notion that ego selection instruments would be wise to carefully train
depletion may have limited benefits. The only similar employees on when it is (and is not) rational to persist
findings to the authors’ knowledge are that ego in a work strategy.
depletion may decrease unethical decision making in Despite the contributions of this research, the studies
certain circumstances (Yam, Chen, & Reynolds, 2014) presented here have important limitations. First, the
and may reduce innattentional blindness (Schofield, effects observed in this research were small. Smaller
Creswell, & Denson, 2015). Previous research has likely effects are inherently less reliable and more vulnerable
missed potential benefits of ego depletion because the to type I error than larger effects. Thus, additional
dependent measures in other ego depletion research research will be required to conceptually replicate the
are often designed such that persistence is always inter- effects reported here in other samples and using different
preted as a desirable response. For example, Muraven, dependent measures. Despite the small effect sizes, the
Tice, and Baumeister (1998) interpreted persistence on effects reported here represent changes in actual behavior
unsolvable anagrams as indicating effective self-control. (number of anagrams solved and number of articles read)
However, as argued earlier in this paper, persisting on rather than changes in abstract self-report items. Medium
an impossible task might not deserve such high praise. to large changes in underlying attitudes and intentions
To the contrary, the present research used dependent are generally associated with only modest effects on
measures that rewarded a strategic use of time and behavior (Webb & Sheeran, 2006). Thus, even these small
found that ego-depleted participants outperformed to medium changes in behavior should be considered
nondepleted participants. This represents a novel practically meaningful. Further, these effects are not
contribution to the empirical literature as the absence due to a “file-drawer problem,” as the three preceding
of findings demonstrating either a potential upside to investigations represent all efforts on the part of the
ego depletion or a downside to excessive self-control author to investigate this line of research (i.e., the file
has been interpreted by some as indicating that strong drawer is empty). Second, participants in Studies
willpower is an unqualified advantage (e.g., Baumeister 2 and 3 were not offered any incentives when solving
& Alquist, 2009). anagrams. This was evident in Study 2, as the opport-
The findings of the present study have several impor- unity cost manipulation did not produce an effect. The
tant implications for future research directions. First results may have turned out differently if participants
and foremost, these results suggest that theorists reeval- had been offered an incentive for every correctly solved
uate the role of ego depletion in models of decision anagram. Given the importance of motivation in self-
making. The popular view that ego depletion is simply control and ego depletion (see Muraven & Slessareva,
an inconvenience that only occurs due to the physical 2003; Osgood, 2017), measures or manipulations of
limitations of the human brain seems inadequate. motivation should be included in future research on this
Indeed, recent advances in social neuroscience suggest topic. This was likely less of a problem in Study 1, in
that ego depletion does not occur because of metabolic which participants thought they would provide feedback
limitations (Kurzban, 2010), and the present findings to university officials (which would have been
suggest that ego depletion may play a constructive role motivating). Third, the studies presented here suffer
in decision making. New theories and research should from several other key methodological shortcomings.
consider ego depletion as an important and instrumen- One such shortcoming is small sample sizes. In each
tal part of a larger decision-making and behavioral of the studies, the sample size per group was 25 or fewer.
modification process. Second, future research should However, in Study 2, the awareness manipulation
seek to understand what cognitive or affective processes was not central to the main hypothesis being tested.
trigger or inform a person to cease exerting self-control. The reader could reasonably collapse across this
 BASIC AND APPLIED SOCIAL PSYCHOLOGY 169

manipulation and consider the ego depletion versus Boucher, H. C., & Kofos, M. N. (2012). The idea of money
nondepletion groups, thus resulting in a more counteracts ego depletion effects. Journal of Experimental
satisfactory sample of 40þ ego-depleted participants Social Psychology, 48(4), 804–810. doi:10.1016/j.jesp.2012.
02.003
versus 40þ nondepletion participants. Nevertheless, Churchill, W. S. (2003). The Best of Winston Churchill’s
the combination of modest sample size and small to Speeches/Selected by his Grandson Winston S. New York:
moderate effect sizes suggests that the reader should Hyperion.
use care when interpreting the results and encourages Cunningham, M. R., & Baumeister, R. F. (2016). How to make
others to attempt to replicate these findings. Another nothing out of something. Frontiers in Psychology, 7, 1639.
doi:10.3389/fpsyg.2016.01639
methodological shortcoming was the absence of manipu-
Dang, J. (2018). Can the ego be depleted?: Attempts to
lation checks in Study 2. This may have been particularly replicate the ego depletion effect and integrate its explana-
problematic, as there was no verification that the tions (Masters thesis). Retrieved from https://www.lup.lub.
awareness instructions induced different levels of task lu.se
awareness and this variable did not produce an effect Fox, N. A., & Calkins, S. D. (2003). The development of self-
on number of anagrams solved. Thus, it is unclear if control of emotion: Intrinsic and extrinsic influences.
Motivation and Emotion, 27(1), 7–26.
the lack of effect is due to awareness actually being Gailliot, M. T., Baumeister, R. F., DeWall, C. N., Maner, J. K.,
unimportant or a failed manipulation. The concern over Plant, E. A., Tice, D. M., Brewer, L. E., & Schmeichel, B. J.
lack of manipulation checks was addressed with the (2007). Self-control relies on glucose as a limited energy
addition of manipulation checks in Study 3. Further, source: Willpower is more than a metaphor. Journal of
some methodologists have argued that the absence of Personality and Social Psychology, 92(2), 325–336.
doi:10.1037/0022-3514.92.2.325
self-report manipulation checks in research are not neces-
Hagger, M. S., Chatzisarantis, N. L., Alberts, H., Anggono,
sarily problematic when the results of a study support the C. O., Batailler, C., Birt, A. R., … , & Calvillo, D. P.
research hypothesis (see Trafimow & Rice, 2009). (2016). A multilab preregistered replication of the
In closing, the findings presented here should not ego-depletion effect. Perspectives on Psychological Science,
leave the reader with impression that perseverance or 11(4), 546–573. doi:10.1177/1745691616652873
self-control is undesirable. Indeed, success either Inzlicht, M., & Schmeichel, B. J. (2012). What is ego
depletion? Toward a mechanistic revision of the resource
personally or professionally is likely often the result of model of self-control. Perspectives on Psychological Science,
long periods of dutiful persistence, effort, trial, and 7(5), 450–463. doi:10.1177/1745691612454134
error. Further, strong self-control and the ability to Inzlicht, M., Schmeichel, B. J., & Macrae, C. N. (2014). Why
exert effortful control is nearly universally beneficial self-control seems (but may not be) limited. Trends in
(see Baumeister & Alquist, 2009). However, the impli- Cognitive Sciences, 18(3), 127–133. doi:10.1016/j.tics.2013.
12.009
cations of these findings are that individuals must
Janssen, L., Fennis, B. M., & Pruyn, A. T. H. (2010). Fore-
be careful to not commit to persisting too easily. Do warned is forearmed: Conserving self-control strength to
not devote yourself unwaveringly to any romantic resist social influence. Journal of Experimental Social
relationship or to any employment opportunity. Hard Psychology, 46(6), 911–921. doi:10.1016/j.jesp.2010.06.008
work is needed for success, but so are effective Job, V., Dweck, C. S., & Walton, G. M. (2010). Ego
decision-making and self-monitoring processes to depletion—Is it all in your head? Implicit theories
about willpower affect self-regulation. Psychological Science,
guide that hard work. As Winston Churchill advised, 21(11), 1686–1693.
we should be sensitive to our “good sense” when Keinan, A., & Kivetz, R. (2008). Remedying hyperopia: The
deciding whether to persist. effects of self-control regret on consumer behavior. Journal
of Marketing Research, 45(6), 676–689. doi:10.1509/jmkr.
45.6.676
References Kivetz, R., & Simonson, I. (2002). Self-control for the
righteous: Toward a theory of precommitment to indul-
Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. gence. Journal of Consumer Research, 29(2), 199–217.
Organizational Behavior and Human Decision Processes, doi:10.1086/341571
35(1), 124–140. doi:10.1016/0749-5978(85)90049-4 Koole, S. L., Tops, M., Strübin, S., Bouw, J., Schneider, I. K., &
Baumeister, R. F. (2003). Ego depletion and self regulation Jostmann, N. B. (2014). The ego fixation hypothesis:
failure: A resource model of self control. Alcoholism: Involuntary persistence of self-control. In J. P. Forgas and
Clinical and Experimental Research, 27(2), 281–284. E. Harmon-Jones (Eds.), The Control Within: Motivation
Baumeister, R. F., & Alquist, J. L. (2009). Is there a downside and Its Regulation (pp. 95–112). New York: Psychology
to good self-control? Self and Identity, 8(2–3), 115–130. Press.
doi:10.1080/15298860802501474 Kurzban, R. (2010). Does the brain consume additional
Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. glucose during self-control tasks? Evolutionary Psychology,
(1998). Ego depletion: Is the active self a limited resource? 8(2), 147470491000800208. doi:10.1177/147470491000
Journal of Personality and Social Psychology, 74(5), 1252–1265. 800208
170 J. M. OSGOOD

Kurzban, R., Duckworth, A., Kable, J. W., & Myers, J. (2013). diminish prosocial behaviors. Basic and Applied Social
An opportunity cost model of subjective effort and Psychology, 37(1), 68–80. doi:10.1080/01973533.2014.996225
task performance. Behavioral and Brain Sciences, 36(6), Osgood, J. M., & Muraven, M. (2016). Does counting to ten
661–679. doi:10.1017/s0140525x12003196 increase or decrease aggression? The role of state self‐
Lange, F., & Eggert, F. (2014). Sweet delusion. Glucose drinks control (ego‐depletion) and consequences. Journal of
fail to counteract ego depletion. Appetite, 75, 54–63. Applied Social Psychology, 46(2), 105–113. doi:10.1111/
doi:10.1016/j.appet.2013.12.020 jasp.12334
Masicampo, E. J., & Baumeister, R. F. (2012). Committed Osgood, J. M., & Salamone, A. (2016). Can trait self-control
but closed-minded: When making a specific plan for a improve with training? A review of the evidence and
goal hinders success. Social Cognition, 30(1), 37–55. possible underlying mechanisms. Advances in Psychology
doi:10.1521/soco.2012.30.1.37 Research, 116, 1–30.
Mayer, J. D., & Gaschke, Y. N. (1988). The experience Schofield, T. P., Creswell, J. D., & Denson, T. F. (2015). Brief
and meta-experience of mood. Journal of Personality and mindfulness induction reduces inattentional blindness.
Social Psychology, 55(1), 102–111. doi:10.1037//0022- Consciousness and Cognition, 37, 63–70. doi:10.1016/j.
3514.55.1.102 concog.2015.08.007
Muraven, M., & Baumeister, R. F. (2000). Self-regulation and Staw, B. M. (1981). The escalation of commitment to a course
depletion of limited resources: Does self-control resemble a of action. Academy of Management Review, 6(4), 577–587.
muscle? Psychological Bulletin, 126(2), 247–259. doi:10.5465/amr.1981.4285694
Muraven, M., & Slessareva, E. (2003). Mechanisms of Tice, D. M., & Bratslavsky, E. (2000). Giving in to feel good:
self-control failure: Motivation and limited resources. The place of emotion regulation in the context of general
Personality and Social Psychology Bulletin, 29(7), 894–906. self-control. Psychological Inquiry, 11(3), 149–159.
doi:10.1177/0146167203029007008 doi:10.1207/s15327965pli1103_03
Muraven, M., Tice, D. M., & Baumeister, R. F. (1998). Self- Tice, D. M., Baumeister, R. F., Shmueli, D., & Muraven, M.
control as a limited resource: Regulatory depletion (2007). Restoring the self: Positive affect helps improve
patterns. Journal of Personality and Social Psychology, self-regulation following ego depletion. Journal of Experi-
74(3), 774–789. doi:10.1037//0022-3514.74.3.774 mental Social Psychology, 43(3), 379–384. doi:10.1016/j.
Nelissen, R., Breugelmans, S. M., & Zeelenberg, M. (2013). jesp.2006.05.007
Reappraising the moral nature of emotions in decision Trafimow, D., & Rice, S. (2009). What if social scientists had
making: The case of shame and guilt. Social and Personality reviewed great scientific works of the past? Perspectives on
Psychology Compass, 7(6), 355–365. doi:10.1111/spc3.12030 Psychological Science, 4(1), 65–78. doi:10.1111/j.1745-6924.
Osgood, J. M. (2015). Acute cardiovascular exercise 2009.01107.x
counteracts the effect of ego-depletion on attention: How Webb, T. L., & Sheeran, P. (2006). Does changing behavioral
ego-depletion increases boredom and compromises intentions engender behavior change? A meta-analysis of
directed attention. International Journal of Psychological the experimental evidence. Psychological Bulletin, 132(2),
Studies, 7(3), 85–96. doi:10.5539/ijps.v7n3p85 249–268. doi:10.1037/0033-2909.132.2.249
Osgood, J. M. (2017). Effect of ego-depletion typing task on Wegner, D. M. (1992). You can’t always think what you want:
Stroop does not extend to diverse online sample. Journal Problems in the suppression of unwanted thoughts.
of Articles in Support of the Null Hypothesis, 13(2), 83–89. In Advances in experimental social psychology (Vol. 25,
Osgood, J. M. (2017). Is revenge about retributive justice, pp. 193–225). Academic Press.
deterring harm, or both? Social and Personality Psychology Yam, K. C., Chen, X. P., & Reynolds, S. J. (2014). Ego
Compass, 11, 1. depletion and its paradoxical effects on ethical decision
Osgood, J. M., & Muraven, M. (2015). Self-control depletion making. Organizational Behavior and Human Decision
does not diminish attitudes about being prosocial but does Processes, 124(2), 204–214. doi:10.1016/j.obhdp.2014.03.008