Self depletion ego

History (H) does not refer to the occurrence of events before the experiment. Rather, history refers to specific events that are external to the experiment but occur at the same time as the experiment. These events may affect the dependent variable

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

Other-To link to this article: https://doi.org/10.1080/00913367.2021.1887012

View supplementary material

Published online: 02 Apr 2021.

Submit your article to this journal

Article views: 147

View related articles

View Crossmark data

Ego Depletion and Charitable Support: The Moderating Role of Self-Benefitand Other-Benefit Charitable Appeals

Hyun Seung Jina
, Hyoje Jay Kima
, Jaebeom Suhb

, Ben Sheehanaand Robert Meedsc

self-Charitable organizations are facing increased tion In the United States, donations remain stable atapproximately 2% of GDP; however, the growth rateof nonprofit organizations has been estimated at 3.4%per annum (Harrison and Irvin 2018) This competi-tion can lead to an excessive, inefficient level of fund-raising (Castaneda, Garen, and Thornton 2007;Thornton 2006) Further compounding the problem,donors have shown a preference for charities with lowadministrative expenses As a result, charities maydecrease their fundraising expenses, which reducesfuture donations and their capacity to deliver socialprograms (Burkart, Wakolbinger, and Toyasaki 2018).This feedback loop has been termed the nonprofitstarvation cycle (Lecy and Searing2015) Competitionis forcing charities to invest in advertising, but donorsperceive this advertising as diverting their contribu-tion away from potential beneficiaries This paper

competi-provides a means by which charities can increase theeffectiveness of their donation appeals.

Benefactors are under pressure themselves Modernconsumer culture provides endless choices; managingworkplace and relational stressors requires self-regula-tion and achieving personal goals taps one’s self-con-trol resources These pressures can result in egodepletion, a state in which one’s self-control resourceshave been temporarily exhausted after exertion(Baumeister et al 1998) Existing literature suggeststhat ego depletion reduces prosocial behavior, asdepleted individuals feel less guilt and are thereforeless inclined to help others (Xu, B
egue, and Bushman2012) Furthermore, donors are often exposed to mul-tiple donation requests (Erceg et al 2018) as it ismore cost-effective to reapproach a known donorthan to attract a new one In response, donors mayreduce subsequent donations after having previouslymade a contribution (Adena and Huck 2019) In line

CONTACTHyun Seung Jinhs.jin@qut.edu.auQueensland University of Technology, 2 George Street, Brisbane, QLD, Australia.
The first two authors contributed equally to this article.

Hyun Seung Jin (PhD, University of North Carolina) is a Senior Lecturer, QUT Business School, Queensland University of Technology.

Hyoje Jay Kim (PhD, Queensland University of Technology) is a Research Lab Coordinator, QUT Business School, Queensland University of Technology.Jaebeom Suh (PhD, University of Alabama) is an Associate Professor of Marketing, College of Business Administration, Kansas State University.Ben Sheehan (MPhil, Queensland University of Technology) is a PhD candidate, QUT Business School, Queensland University of Technology.Robert Meeds (PhD, University of Missouri) is a Professor of Communications, College of Communications, California State University at Fullerton.Supplemental data for this article is available online athttps://doi.org/10.1080/00913367.2021.1887012.

Copyrightß 2021, American Academy of Advertising

with the consequences of ego depletion, otherresearchers have shown that individuals are morelikely to engage in a selfish act after a prosocial act(Krishna2011; Schwabe, Dose, and Walsh2018).

This research examines a means by which charityorganizations can mitigate the negative effects of egodepletion We examine the moderating role of mes-sage appeal type Two types of appeals are frequentlyexamined in the prosocial literature: other-benefit andself-benefit messages (Nelson et al 2006; White andPeloza 2009) Three experiments compare the timedonation intent (studies 1 and 2) and actual monetarydonations (study 3) of depleted (vs non-depleted)individuals who have been exposed to either a self-benefit message, highlighting the gains to be accruedto the donor themselves, or an other-benefit messagewhich focuses on the welfare of beneficiaries.

These studies contribute to the literature in severalways First, we provide the first empirical demonstra-tion of the process model of ego depletion of Inzlichtand Schmeichel (2012) in a charity context We dem-onstrate that as donors become depleted, their atten-tion shifts from cues requiring self-control towardcues which signal self-benefit Previous research sug-gests that ego depletion increases selfishness, harmingcharity donations We show that appealing to this self-ishness can promote time donation intent and actualmonetary donations, mitigating or reversing theeffects of ego depletion Second, as a processing mech-anism, we empirically demonstrate that depleted indi-viduals pay more attention to self-benefit messages,which in turn increases charitable support Third, byexamining time-of-the-day effects (morning vs even-ing), we offer a practical implication The data suggestthat charities can maximize donations using other-benefit messages in the morning and self-benefit mes-sages in the evening Last, we treat depletion as both amanipulated and measured variable We manipulatedepletion in study 1 using a standard method takenfrom the depletion literature Then, to enhance exter-nal validity, we measure depletion without a manipu-lation (studies 2 and 3).

Ego Depletion and Prosocial Behaviors

Self-regulation refers to one’s conscious efforts toregulate their emotions, thoughts, impulses, desires,and automatic behavioral responses in order toachieve a goal (Vohs and Schmeichel 2003) A bodyof literature suggests that people have a limited cap-acity for self-regulation For example, the strength-resource model of self-control posits that when people

engage in a self-regulatory activity, the self-controlresource is temporarily exhausted Consequently, theyare likely to fail in subsequent attempts at self-regula-tion (Baumeister et al 1998; Muraven, Tice, andBaumeister 1998) In this strength model, the regula-tory resource is thought to work like a muscle, in thesense that strength decreases after muscle use andremains exhausted until a sufficient recovery periodhas elapsed This reduction in the self-controlresource is called ego depletion (Baumeisteret al 1998).

The ego depletion effect suggests that when control resources are used, subsequent self-controlsuffers; thus, people are less able to override theirimpulses Considerable evidence supports this view ofself-regulation as a limited resource (Hagger et al.2010) Furthermore, the ego depletion effect has beenobserved in a wide range of contexts For instance,depleted people are more likely to engage in behaviorsproviding immediate gratification (Metcalfe andMischel 1999), engage in impulsive buying (Vohs andFaber2007), have a greater temptation to cheat (Meadet al 2009), binge eat (Joiner, Vohs, and Heatherton2000), engage in unhealthy food consumption (Job,Dweck, and Walton 2010), and have violent impulses(Finkel et al.2009).

self-Prior research in self-regulation has also examinedthe relationship between ego depletion and prosocialbehaviors A line of research posits that helping othersrequires self-control Baumeister and Exline (1999)view self-control as a “moral muscle” because self-con-trol curbs selfishness in favor of other-focused behav-iors, enabling society to function In a similar vein,DeWall et al (2008) argue that “to help others, peoplemay overcome a natural impulse toward selfishness andself-interest—but overcoming it may require advancedpsychological processes, such as self-regulation”(1653–1654) As such, a selfish or less altruistic behav-ior becomes more likely when one’s self-controlresource is depleted Hence, depletion increases self-serving behaviors and/or decreases prosocial behaviors.

Empirical evidence has demonstrated that whendepleted, people are less likely to override their selfishinclinations and, thus, subsequent prosocial behaviorsand intent are reduced (Achtziger, Alos-Ferrer, andWagner 2015; Balliet and Joireman 2010; DeWallet al 2008; Osgood and Muraven 2015; Xu, B
egue,and Bushman 2012) For example, DeWall et al.(2008) found that depleted participants were less will-ing to engage in helping strangers in six hypotheticalscenarios Osgood and Muraven (2015) found thatego depletion negatively influences cooperation

behaviors by reducing motivation to overcome egoisticdesires when helping others comes at a cost to theself Depleted participants chose to allocate a greaterreward to themselves as compared to non-depletedparticipants in a dictator game where one player (dic-tator) makes a decision on how a reward is dividedbetween herself/himself and the other player(Achtziger, Alos-Ferrer, and Wagner 2015; Xu, B
egue,and Bushman 2012) Other studies using a decom-posed game (i.e., a choice from various distributionsof resources between the self and another person)reported similar results (Balliet and Joireman 2010).These findings support the idea that ego depletionleads to selfishness.

Depletion, however, does not always lead to selfishbehaviors Some studies have found boundary condi-tions under which depletion does not decrease pro-social behaviors For example, DeWall et al (2008)reported that although depletion reduced helpingtoward strangers, it did not decrease helping towardfamily members Similarly, Balliet and Joireman(2010) found that there was no depletion effect onprosocial behaviors when participants had a prosocialorientation, a trait concerned with maximizing jointgain and quality with others.

Interestingly, a body of research has found thatdepletion can even increase prosocial behavior undercertain conditions Studies found that ego depletionenhanced compliance with charitable requests whensocial influence techniques were used (Fennis, Janssen,and Vohs 2009; Janssen et al 2008) For example,Janssen et al (2008) tested the heuristic principle ofauthority Participants in the ego depletion conditionwere exposed to a charitable message from either awell-known charity organization or an unknown char-ity organization Those who were depleted donatedmore than non-depleted people when the authorityprinciple was activated (i.e., well-known organization).However, for the unknown charity organization,depletion did not affect compliance Similarly, Fennis,Janssen, and Vohs (2009) investigated the reciprocityprinciple Depleted (vs non-depleted) participantswere assigned to either the reciprocity condition ornon-reciprocity condition, and then a compliancebehavior was measured In the reciprocity condition,the experimenter told participants she would make anexception and excuse them from the next quite boringtest, because she collected enough data from the pre-vious test Participants in the non-reciprocity condi-tion were not told about this exemption from anonexistent test Those who were depleted showedhigher compliance in volunteering than non-depleted

individuals in the reciprocity condition Fennis,Janssen, and Vohs (2009) also tested the heuristicprinciple of likability Participants in the likabilitycondition were complimented for their ability to com-plete a task In the control condition, no commentswere made Results were consistent Depleted partici-pants showed higher compliance in volunteeringunder the likability condition These studies suggestthat depletion fosters compliance with a charitablerequest through reliance on salient heuristics.

The Moderating Role of Self- versus Benefit Appeals

Other-Researchers have argued that people give for two basicreasons: altruistic and egoistic (Cialdini et al 1997).Altruistic giving refers to giving in order to enhancethe well-being of others, while the primary goal ofegoistic giving is increasing one’s own image andpositive affect Similarly, an other-benefit appealfocuses on altruistic reasons for giving, such as bene-fits to be accrued by people in need, whereas a self-benefit appeal heightens egoistic reasons for giving,such as incentives or rewards Self-benefit appealsvary in reward types, which can include tangible (e.g.,tax offset) or intangible (e.g., feeling good about one-self) benefits.

Charity organizations use two types of messageappeals: other- versus self-benefit appeals Strategicuse of other- versus self-benefit messages has been apopular topic in advertising and marketing literature(Brunel and Nelson 2000; Feiler, Tost, and Grant2012; Green and Peloza 2014; Kareklas, Carlson, andMuehling 2014; Nelson et al 2006; White and Peloza2009) For example, research has found that other-benefit appeals generated more favorable donationsupport (White and Peloza 2009) and environmentallyfriendly products (Green and Peloza 2014) than self-benefit appeals when consumers were publiclyaccountable for their responses, while the oppositepattern was found when consumers’ responses wereprivate Prior research has found gender and culturaldifferences For example, Brunel and Nelson (2000)showed that females respond more favorably to other-benefit appeals and males to self-benefit appeals Thefindings were replicated in masculine cultures; how-ever, the opposite pattern was observed in femininecultures (Nelson et al 2006) While it is possible for acharity organization to highlight both self- and other-benefits in a single message (Feiler, Tost, and Grant2012; Kareklas, Carlson, and Muehling 2014), we

separate the two appeal types in order to examinewhen each appeal is most effective.

In the previous section, we described how heuristiccue salience effects are thought to occur becausedepletion hinders systematic message-relevant process-ing and enhances the weight of heuristic processing indecision-making (Fennis, Janssen, and Vohs 2009).Other researchers provide a similar argument, suggest-ing that ego depletion increases susceptibility to situ-ational cues (Banker et al 2017) As far as depletionenhances the effects of heuristic cues, loss of self-con-trol may generate more selfish or more prosocial deci-sions, depending on what the cues advocate (Bankeret al 2017).

The process model of ego depletion proposed byInzlicht and Schmeichel (2012) provides a theoreticalexplanation for what makes a cue salient to depletedpeople; that is a rewarding cue They propose a “shiftsin attention” hypothesis It posits that “ego depletionleads to a shift in attention away from signs of goalconflict and discrepancy and instead toward signs ofpossible reward and gratification” (Inzlicht andSchmeichel 2012, 457) Using functional neuroimag-ing, Wagner et al (2013) examined brain activity inresponse to viewing food items among chronic dieters.They found that depletion enhances neural responsesto rewards Depleted dieters showed greater foodcue–related activity in the orbitofrontal cortex, whichis associated with coding the reward value.

Schmeichel, Harmon-Jones, and Harmon-Jones(2010) suggest that self-benefit messages may be moreengaging for depleted individuals as a result of shiftsin attention They assigned participants to depletedand non-depleted conditions Following the manipula-tion, participants were exposed to visual symbols thatare associated with reward (e.g., $ sign) or to symbolsthat are not associated with reward (e.g., % sign).They were asked to make quick identity judgmentsabout these symbols Results found that depleted par-ticipants perceived and detected the dollar signs moreaccurately compared to non-depleted participants,suggesting that depletion facilitates, rather than inter-feres with, attention to a cue that is associated withreward These empirical studies demonstrate thatdepletion increases the likelihood of attending andresponding to reward stimuli.

The critical aspect of self-benefit appeals is thatthey include a self-serving or self-rewarding stimuli orcue For example, other-benefit messages may say“save people’s lives and help others live,” while a self-benefit message would say “save your life and protectyour future” (Brunel and Nelson2000) Another study

used stimuli suggesting that volunteering can “helpmake the community a better place for everyone”(other-benefit) versus “build your resume by develop-ing and practicing job skills” (self-benefit) (White andPeloza 2009) Thus, shifts in attention toward reward-ing cues should make self-benefit appeals salient todepleted people.

Building upon the shifts in attention hypothesisand its empirical evidence, we propose that a self ver-sus other-benefit message appeal plays a moderatingrole between ego depletion and charitable support.When people are depleted, attention should shift toreward-seeking cues Therefore, depleted people paymore attention to self-benefit messages, which in turngenerates more charitable support As such,we propose:

H1: When depleted (vs non-depleted) individuals areexposed to a self-benefit (vs other-benefit) message,they are more likely to provide charitable support.H2: Self-benefit (vs other-benefit) messages are moreeffective in generating charitable support amongdepleted individuals as they pay more attention toself-benefit messages.

Study 1

The goal of study 1 was to test an interaction betweenego depletion and message appeals on time donationintent The experiment involved a 2 (self-regulation:depleted vs non-depleted)  2 (message appeals: selfvs other benefit) between-group design A total of225 college students from a major midwestern univer-sity in the U.S participated in the experiment inexchange for extra course credit (57% female, Mage ¼20.4, SD¼ 3.04).

Stimuli Development

For the ad stimuli, we used a health charity as the get organization The stimuli was a modified versionof that used by Brunel and Nelson (2000) and Whiteand Peloza (2009) In the other-benefit appeal, thestimuli described how the charity helped those inneed, while the self-benefit appeal highlighted thebenefits to be accrued by donors For example, theheadline of the other-benefit (vs self-benefit) messagestated that donating could “save people’s lives” (vs.“save your life”) In the body of the text, the other-benefit appeal stated that small gestures of caring,such as a meal or soft blanket, mean a lot to patients.The self-benefit message highlighted that employers

tar-are impressed by volunteer work on an applicant’sresume and that volunteers may meet important con-tacts who can help them to secure a good job (seeOnlineAppendix Afor stimuli).

Experimental Procedure

The experiment was conducted in a computer lab, ingroups of 10 to 20 participants Each participant sat ata computer and finished the experiment using a web-based interface Participants were randomly assignedto one of four conditions In order to avoid potentialdemand effects of the ego depletion manipulation(Stroop task), participants were told that they wereparticipating in two independent studies in which thefirst study involved a computer-based cognitive task(Stroop task) and the second one focused on charit-able behaviors.

For the ego depletion manipulation, we used theStroop task, which is a common depletion manipula-tion method used in the ego depletion literature.Participants were presented with 52 color words (e.g.,red, blue, yellow, and green), one at a time on thecomputer screen These words appeared in either thesame font color or a different color with the semanticmeaning of the word For example, the word “blue”could be written in a blue color or in red, yellow, orgreen Respondents were informed that the task wasto indicate the correct font color as quickly and accur-ately as possible Participants responded by clickingone of four color buttons at the bottom of the screen.Before participants began the task, they practiced thetask with two examples.

For participants in the depleted condition, 75% ofthe words were mismatched with the font colors Thistask requires self-regulation because the semanticmeaning of the word is an automatic response; hence,avoiding this response requires regulatory control(Fennis, Janssen, and Vohs 2009) For participants inthe non-depleted condition, all of the words werematched with font colors Thus, no self-regulation wasrequired After they completed the Stroop task, weadministered manipulation check questions and meas-ured current feelings, using the Positive and NegativeAffect Schedule (Watson, Clark, and Tellegen 1988).The proctor told participants to wait until furthernotice for the next study.

When all participants had completed the firststudy, the proctor explained the second study.Participants were told that this was a pilot test for acharity campaign in the future The proctor advisedthe participants that providing an honest responsewas important because the test results would help

develop an effective message for the target charity.Immediately after participants were exposed to thestimulus material, we measured the dependent andother variables.

Dependent measure Three questions were asked tomeasure the participant’s time donation intent (a¼.85): (a) How likely would you be to donate your timeby volunteering for the charity organization? (b) Howinclined are you to volunteer for the charity organiza-tion? (c) How willing are you to volunteer for the char-ity organization? A 7-point scale was used, where 7indicates more generous time donation intent.

Attention to the message Three questions were askedto measure overall attention to the message (a ¼.87).The scale was taken from Laczniak and Muehling(1993): (a) How much attention did you pay to the mes-sage? (b) How much did you concentrate on the mes-sage? (c) How involved were you with the message? A7-point scale was used, where 7 indicates high attention.Emotions We asked participants to indicate towhat extent they presently felt (a) enthusiastic, (b)active, (c) distressed, (d) tense, (e) irritable, and (f)frustrated A 7-point scale was used, with 1 being “notat all” and 7 being “very much.”

Manipulation checks Three questions were asked tomeasure the extent to which the Stroop task waseffortful (a ¼ 81): (a) How much effort did you exerton the task? (b) How difficult was the task? (c) Howmuch attention did the task require? All items weremeasured by a 7-point scale A composite variablewas created by averaging the items.

Self- versus other-benefit appeal We asked threequestions (a ¼ 67): (a) the message is trying to help : “people in general”–“me and my family,” (b)the message talked about how can benefit bydonating my time to the charity organization (areversed item): “people in general”–“I” and (c) themessage seemed like it was directed to“everyone”–“me specifically.” A 7-point scale wasused The reversed item was recoded Thus, a highervalue indicated that the message was more self-benefitoriented A composite score was generated averagingthe three items Finally, we collected participants’demographics (e.g., age and gender).

Analysis and Results

Manipulation Checks and Other Tests

The results indicate that ego depletion and messageappeal manipulations were evident Those who were

in the depleted condition indicated that the task wasmore effortful than those who were in the non-depleted condition (Mdepleted ¼4.08, SD ¼ 1.37; Mnon-depleted ¼ 2.66, SD ¼ 1.10, F (1, 223) ¼ 73.44, p <.001, gp2 ¼ 25) We also found that the self-benefitmessage was perceived as being more self-benefit ori-ented than the other-benefit message (Mself ¼4.13,SD¼ 1.17; Mother ¼ 3.17, SD ¼ 1.28, F (1, 223) ¼36.63, p < 001, gp ¼ 12) In addition, we examinedwhether participants’ mood states differed because ofthe different levels of the Stroop task Two positivemoods were averaged (enthusiastic and active, a ¼.76) as were four negative moods (a ¼ 79).Participants’ moods were not different across the non-depleted and depleted conditions: positive mood(Mnon-depleted ¼ 4.33 vs Mdepleted ¼ 4.51, F < 1, p¼.32, gp2 ¼ 004) and negative mood (Mnon-depleted ¼3.11 vs Mdepleted¼ 3.10, F < 1, p ¼.98, gp < 001).

Effects of Ego Depletion and Message Appeals onTime Donation Intent

The results of a two-way analysis of covarianceincluding age, gender, positive mood, and negativemood as covariates revealed that none of the covari-ates were associated with the dependent variable.

Thus, we report the results of 2 (depleted vs depleted)  2 (self- vs other-benefit appeal) between-subject analyses of variance (ANOVAs).

non-The main effect of ego depletion on time donationintent was marginally significant (F (1, 221) ¼ 2.77, p¼ 097, gp2 ¼ 012) The main effect of messageappeals was not significant (F (1, 221) < 1, p ¼ 81,gp2 < 001) Our focal interest was the interactioneffect The results indicate that there was a significantinteraction effect (F (1, 221) ¼ 12.74, p < 001, gp ¼.055) Figure 1A presents the means and standarddeviations for the four conditions A simple effectanalysis shows that when participants were notdepleted, the other-benefit appeal resulted in moregenerous time donation intent than the self-benefitappeal (Mother ¼ 4.66, Mself¼ 4.33, F (1, 221) ¼ 5.65,p ¼ 018, gp ¼.025) Conversely, when participantswere depleted, the self-benefit appeal was more effect-ive than the other-benefit appeal in generating timedonation intent (Mself ¼ 4.52, Mother ¼ 4.15, F (1,221) ¼ 7.12, p ¼ 008, gp ¼ 03) The findings sup-ported our interaction hypothesis (H1).

Moderated Mediation Analysis

We predicted that attention to the message wouldmediate the interaction effect of ego depletion andmessage appeal upon time donation intent To testthis mediated moderation model, PROCESS Model 8was used (Hayes 2018) The moderated mediationmodel is presented in Figure 2.

The self-benefit message was coded “1,” whereasthe other-benefit message was coded“0.” The depletedcondition was coded “1,” and the control conditionwas coded “0.” A bias-corrected confidence interval(CI; 95%) and bootstrapping with 5,000 repetitionswere employed to estimate the indirect effect.

First, we examined the effects of two independentvariables on attention to the message The ego deple-tion (b ¼ .18, SE ¼ 22, p ¼ 40) and messageappeal (b ¼ .21, SE ¼ 22, p ¼ 34) variables did

Figure 1 Study 1: time donation intent and attention to themessage.

Note: The values are means (standard deviations) The sameletter indicates a significant mean difference (p < 05).

Figure 2 Study 1: a moderated mediation model (ProcessModel 8)
p < 05;

p < 01;


p < 001.

not predict attention to the message However, theinteraction effect was significant (b¼ 74, SE ¼ 31, p¼ 02) Second, we looked at whether attention pre-dicted time donation intent We found that attentionwas positively associated with the dependent variable(b ¼ 15, SE ¼ 04, p < 001) Next, we observed asignificant index of moderated mediation (b ¼ 12, SE¼ 06) with a 95% CI of [.0174, 2508]; thus, a moder-ated mediation was established The conditional indir-ect effects were examined The indirect effect model,depletion (vs non-depletion) ! attention to the mes-sage! time donation intent was significant when themessage appeal was self-benefit (b ¼ 08, SE ¼ 04,95% CI [.0199, 1837]) However, when the messageappeal was other-benefit, the indirect effect was notsignificant (b ¼ .03, SE ¼ 04, 95% CI[.1092, 0449]).

The differential conditional indirect effects werederived from the interaction effect of depletion mes-sage appeals on attention More specifically, the self-benefit message had a significantly higher attentionscore than the other-benefit message when partici-pants were depleted (Mdepleted ¼ 4.73 vs Mnon-depleted

¼ 4.17, F (1, 220) ¼ 6.62, p ¼ 011) However, theother-benefit message did not differ in attentionscores between depleted and non-depleted conditions(Mdepleted ¼ 4.19 vs Mnon-depleted ¼ 4.37, F (1, 220) ¼.07, p ¼ 40) The results supported our secondhypothesis The means and standard deviations ofattention to the message for the four conditions arepresented inFigure 1B.

Study 1 Discussion

We proposed an interaction effect between ego tion and charitable message appeal on subsequenttime donation intent As hypothesized, we found asignificant interaction effect such that when partici-pants were depleted (vs non-depleted), the self-benefit(other-benefit) appeal was more effective in generatingtime donation intent The mediation analysis suggeststhat this is due to the self-benefit message attractingmore attention from depleted individuals.

deple-Although this study provides empirical evidence ofa reward-seeking tendency when control resources aredepleted, the study has some limitations First, collegestudents may be more generous with their time com-pared to the general public A study with an adultsample is needed Second, although the Stroop task asa manipulation of depletion is a commonly usedmethod in the ego depletion literature, it lacks eco-logical validity Ego depletion research is

overwhelmingly conducted in controlled laboratoryexperiments However, given the study’s managerialcontext, an alternative means of measuring depletionis warranted.

Study 2

Study 1 appears to support the theoretical model sented The purpose of study 2 was a replication andextension of study 1, designed to address the limita-tions detailed above As such, study 2 utilized a differ-ent charity organization and an adult sample Inaddition, this study treated ego depletion as a measuredvariable, as opposed to a manipulated variable In thisway, study 2 employed a single-factor, between-groupexperiment: self- versus other-benefit appeal.

Stimuli Development

Where our previous study used a health charity, thisstudy used a stimulus based on an appeal from acharity for children This charity was chosen as it ispolitically neutral, focusing on helping terminally illchildren Stimuli in the self-benefit condition featuredthe headline, “Charity benefits the giver too.” Thebody of the text supported this self-benefit message,stating “You will find it personally rewarding and thatit makes you happy knowing you’re doing somethingimportant, knowing that you’re contributing to agreater cause.” Conversely, the stimuli to be used inthe other-benefit appeal condition featured the head-line, “Help grant wishes, Transform lives.” The bodytext said that volunteering will make a significant dif-ference in the lives of ill children and that the chil-dren will benefit a lot from just a little of the donor’stime (see OnlineAppendix B for stimuli).

Experimental Procedure

A sample of 104 Americans (45% female, Mage ¼ 39.5,SD¼ 12.7) from Amazon’s Mechanical Turk (MTurk)website elected to participate in the study Participantswere randomly assigned to one of the two experimen-tal conditions and began by providing demographicdata Next, participants were provided with instruc-tions, advising them to pay attention to the charitymessage stimuli Immediately after exposure, partici-pants were asked to indicate their agreement withitems measuring the dependent variable, control varia-bles, and manipulation check questions.

Although MTurk is a popular platform for datacollection, there are some concerns regarding data

quality Following recommendations from the ture (Buhrmester, Talaifar, and Gosling 2018;Chmielewski and Kucker 2020), we screened MTurkparticipants to improve data quality for both study 2and study 3 All participants were required to (a) belocated in the United States, (b) have an approval rateabove 95%, and (c) have completed more than 1,000approved tasks.

The dependent variable (time donation intent) andpositive and negative affect were measured asdescribed in study 1 The major difference regardingmeasures in this study was the treatment of ego deple-tion Given that the Stroop task is an artificial meansof inducing ego depletion, we opted to increase real-ism in study 2 This was achieved by using a naturalstate measure of depletion A 6-item (a ¼ 89) deple-tion instrument was used (Lisjak and Lee 2014) tomeasure the degree to which participants agreed withthe following statements: (a) At this moment, I feelmy energy is running low, (b) At this moment, I feelmy willpower is gone, (c) At this moment, I feel men-tally exhausted, (d) Today, I have worked on mentallychallenging tasks, (e) Today, I have made an import-ant decision, and (f) Today, I have thought deeplyabout something A 7-point Likert scale was used,anchored at strongly disagree (1) and stronglyagree (7).

Analysis and Results

Manipulation Checks and Other Tests

The message appeal manipulation was evident Theself-benefit appeal was perceived as more self-benefitoriented than the other-benefit appeal (Mself ¼3.91,SD¼ 1.28; Mother ¼ 3.08, SD ¼ 1.15, F (1,102) ¼11.98, p ¼ 001, gp2¼ 11) In addition, we examinedwhether participants’ depletion and mood states wereaffected by the different message appeals Participants’mean scores for depletion across the two conditionsdid not differ (Mself ¼ 3.31, SD ¼ 1.61; Mother ¼ 3.53,SD¼ 1.54, F (1, 102) < 1, p ¼ 49, gp ¼ 005).Participants’ moods were not different either: positivemood (Mself ¼ 4.34, SD ¼ 1.66; Mother ¼ 4.42,SD¼ 1.46, F (1, 102) < 1, p ¼ 81, gp ¼ 001) andnegative mood (Mself¼ 2.50, SD ¼ 1.50; Mother ¼ 2.82,SD¼ 1.48, F (1, 102) ¼ 1.25, p ¼ 29, gp2¼ 012).

Effects of Message Appeals and Depletion on TimeDonation Intent

We used Process Model 1 (Hayes 2018), which tests amoderation effect The independent variable wasdepletion, which was a continuous variable The mod-erating variable was the message appeal The other-benefit message was coded “0,” whereas the self-bene-fit message was coded “1.” The dependent variablewas time donation intent.

Results show that the more participants weredepleted, the less time they intended to donate (b ¼.59, SE ¼ 13, p < 001) This depletion effect ontime donation intent was very strong The messageappeals did not show a significant effect on timedonation intent (b ¼ 1.15, SE ¼ 69, p ¼ 099).However, as per hypothesis 1, the interaction effectwas significant (b ¼ 39, SE ¼ 18, p ¼ 03) Next, weran a Johnson-Neyman analysis Figure 3 presents theinteraction pattern and the zone of significance Asshown, depletion had a negative impact on time dona-tion regardless of the message appeal used Both mes-sage appeals showed a downward slope However,when the level of depletion was 4.89 (80th percentile)or above, the self-benefit message generated signifi-cantly higher time donation intent than the other-benefit message Below the cutoff of 4.89, there wasno significant difference between the two appeals intime donation intent We did not see any evidence tosuggest that the other-benefit appeal was more effect-ive than the self-benefit appeal when depletion waslow, as in study 1 For example, when the depletionvalue was 1.5 (16th percentile), there was no differ-ence between the self- and other-benefit messages intime donation intent (b¼ .56, SE ¼ 45, p ¼.22).

Figure 3 Study 2: effects of ego depletion and messageappeals.

Note: There was a significant interaction effect of depletionand message appeals on time donation intentions The resultsof Johnson-Neyman analysis shows the zone of significance.The self-benefit appeal generated significantly higher volun-teering intentions than the other-benefit message appealwhen the level of depletion was 4.89 (80th percentile) orabove At depletion levels below 4.89, there is no significantdifference between the two appeals.

Study 2 Discussion

Study 2 also found an interaction effect of depletionand message appeals However, the patterns of inter-action were different from study 1 In study 1, theinteraction shows a X shape in which a self-benefit(other-benefit) appeal is more effective when peopleare depleted (not depleted) In study 2, where deple-tion was measured (not manipulated), the relationshipbetween depletion and time donation intent shows anegative slope for both appeals, but the other-benefitappeal is steeper This pattern indicates that a self-benefit appeal is more effective than an other-benefitappeal when the level of depletion is high However,when the level of depletion is low or moderate, nodifference between the self-benefit and other-benefitappeal was found In sum, consistent across bothstudies is the finding that when people are depleted,the use of a self-benefit appeal is more effective ingenerating time donation intent.

Study 3

The two previous studies had several limitations First,they involved a hypothetical volunteering intent.Second, no monetary donation was examined Third,the advertising messages differed in ways other thanthe experimental manipulation, that is, self- versusother-benefit For example, the other-benefit messageshad a more relational appeal and emphasis on people(plural), while the self-benefit messages featured atransactional tone and a single entity Addressingthese limitations, study 3 strengthened both theinternal and external validity of the research Morespecifically, study 3 used a genuine monetary donationto a real charity organization as the dependent meas-ure and featured two new ads in which the differencesbetween the two messages, except for other- versusself-benefit, were minimized.

Furthermore, study 3 examined time-of-the-dayeffects: morning versus evening Everyday decision-making requires exertion of self-regulation (Kouchakiand Smith 2014) For example, people often controltheir desires and impulses when deciding what to eatfor lunch, whether to travel on the weekend with fam-ily, or how to deal with a difficult client Therefore,self-control resources should deplete graduallythroughout the day This prediction is consistent withearlier work on ego depletion and self-control failure.For example, Kouchaki and Smith (2014) show thatpeople are more likely to make an impulsive decisionlater in the day These impulsive decisions are ofteninterpreted as reflecting the depletion of self-

regulatory capacity (Dewitte, Bruyneel, and Geyskens2009; Vohs and Faber 2007) In line with this theoriz-ing, we hypothesize the following:

H3: People are more likely to donate to a self-benefitmessage (vs an other-benefit message) in the evening(vs in the morning).

The study utilized a 2 (time of the day: morning vs.evening)  2 (message appeals: self-benefit vs other-benefit) between-group design The dependent vari-able was an actual monetary donation to areal charity.

Stimuli Development

We developed ad stimuli around raising funds to helpyoung people with a physical disability The self-bene-fit ad featured the headline, “Your donation can helpyou” and the sub-headline, “Donate now and feelgood You deserve happiness.” Conversely, the other-benefit ad’s headline and sub-headline read: “Yourdonation can help Maria” and “Donate now and helpMaria feel good She deserves happiness.” The bodycopy and the visual elements in the two ads wereidentical (see Online Appendix C for ad stimuli).

Experimental Procedure

To avoid sample selection bias in the morning versusevening conditions, we followed the two-part proced-ure used by Kouchaki and Smith (2014) In part 1, weposted an unrelated study (i.e., product evaluationsurvey) onto the MTurk platform at midmorning on aweekday A total of 700 MTurk participants completedthis unrelated study At the end of the survey, weasked respondents whether they were interested inparticipating in another study (part 2) on the follow-ing day in exchange for 60¢ A total of 537 partici-pants indicated that they were interested Theseparticipants were the base sample from which we ran-domly assigned them to either the morning (8–11a.m.) or evening condition (6–9 p.m.) for the mainstudy (part 2) The times were based on the partici-pants’ local time On the following day, an email wassent to each participant approximately one hourbefore the designated time window The email invita-tion included the study link and the instruction thatthey must complete the study during the desig-nated time.

Participants were randomly presented either theself-benefit or other-benefit message Then, they wereasked how much of their participation fee (60¢) they