Journal of Consumer Research, Inc....marketing, and Juliano Laran ([email protected]) is associate professor of marketing, University of Miami, Coral Gables, FL 33124-6524. Chris Jan-iszewski

Journal of Consumer Research, Inc.

Hedonic Eating Goals and Emotion: When Sadness Decreases the Desire to IndulgeAuthor(s): Anthony Salerno, Juliano Laran, and Chris JaniszewskiSource: Journal of Consumer Research, Vol. 41, No. 1 (June 2014), pp. 135-151Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/10.1086/675299 .

Accessed: 19/05/2014 15:15

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press and Journal of Consumer Research, Inc. are collaborating with JSTOR todigitize, preserve and extend access to Journal of Consumer Research.

http://www.jstor.org

This content downloaded from 128.227.198.238 on Mon, 19 May 2014 15:15:36 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ucpress

http://www.jstor.org/stable/10.1086/675299?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


135

� 2014 by JOURNAL OF CONSUMER RESEARCH, Inc. ● Vol. 41 ● June 2014All rights reserved. 0093-5301/2014/4101-0009$10.00. DOI: 10.1086/675299

Hedonic Eating Goals and Emotion: WhenSadness Decreases the Desire to Indulge

ANTHONY SALERNOJULIANO LARANCHRIS JANISZEWSKI

This article uses the functionalist perspective of emotion to demonstrate that theinfluence of sadness on indulgent consumption depends on the presence of ahedonic eating goal. Sadness heightens a person’s sensitivity to the potentiallyharmful consequences of indulgent consumption, which decreases indulgencewhen a hedonic eating goal is salient. As sadness is often associated with a loss,this protective function is geared toward preventing future losses. The executionof this function is mitigated by feelings of safety, a counterforce to concerns aboutthe harmful consequences of goal pursuit. Alternatively, when a hedonic eatinggoal is not salient, or a salient goal does not have harmful consequences, sadnessresults in emotion regulation (i.e., indulging as a means of feeling better). Thisconceptualization and findings show that the effects of emotions on indulgent con-sumption can be goal-dependent, and that emotions can aid consumers in thebalancing of long-term goals and well-being.

Anumber of models explain how people pursue goals(goal systems theory, Kruglanski et al. 2002; regula-

tory focus theory, Higgins, Shah, and Friedman 1997; test-operate-test-exit, Miller, Galanter, and Pribram 1960; goalsetting theory, Locke and Latham 1990). These models havebeen used to gain insight into factors that moderate goalpursuit, including goal accessibility (Kruglanski 1996; Laranand Janiszewski 2009; van Osselaer and Janiszewski 2012),inhibition (Fishbach, Friedman, and Kruglanski 2003), com-mitment/progress (Fishbach and Dhar 2005), regulatory fit(Crowe and Higgins 1997), and implementation intentions(Gollwitzer and Sheeran 2006). In general, these modelsassume that moderators of goal pursuit exert a consistentinfluence across all types of goals (e.g., increasing the ac-

Anthony Salerno ([email protected]) is a doctoral candidate inmarketing, and Juliano Laran ([email protected]) is associate professor ofmarketing, University of Miami, Coral Gables, FL 33124-6524. Chris Jan-iszewski ([email protected]) is the Russell Berrie Em-inent Scholar Chair and Professor of Marketing, Warrington College ofBusiness Administration, University of Florida, Gainesville, FL 32611-7155. The authors thank Keith Wilcox for his many helpful comments onan earlier draft of this work. The authors also thank the editor, associateeditor, and three anonymous reviewers for their guidance. This article isbased on the first author’s doctoral dissertation.

Ann McGill served as editor and Hans Baumgartner served as associateeditor for this article.

Electronically published January 31, 2014

cessibility of a goal should always increase the motivationto pursue the goal). That is, the influence of a moderatorshould not depend on the type of goal.

Research on the influence of emotions challenges the as-sumption that moderators of goal pursuit should exert aconsistent influence across goals. The functionalist per-spective of emotion proposes that the influence of an emo-tion on goal pursuit is a function of the type of goal beingpursued. Functionalists argue that certain emotions are in-formative to the pursuit of certain goals (Campos et al. 2004;Witherington and Crichton 2007). For example, happinessfacilitates the pursuit of a self-improvement goal but inhibitsthe pursuit of a mood management goal, whereas sadnessfacilitates the pursuit of a mood management goal but in-hibits the pursuit of a self-improvement goal (Fishbach andLabroo 2007). Goal-emotion (dis)synergies exist becauseeach emotion consists of a different set of motivational sub-routines (DeSteno, Valdesolo, and Bartlett 2006; Griskevi-cius, Shiota, and Neufeld 2010; Witherington and Crichton2007). For example, happiness consists of subroutines thatencourage (a) heuristic processing (Tiedens and Linton2001), (b) an abstract construal (Labroo and Patrick 2009),and (c) an approach orientation (Labroo and Rucker 2009),whereas sadness consists of subroutines that promote (a)systematic processing (Tiedens and Linton 2001), (b) a con-crete construal (Labroo and Patrick 2009), and (c) an avoid-ance orientation (Labroo and Rucker 2009). The implicationis that happiness versus sadness should influence goal pur-


mailto:[email protected]




136 JOURNAL OF CONSUMER RESEARCH

suit based on whether the goal is simple versus complex,flexible versus rigid, and based on seeking opportunitiesversus avoiding risks.

We document an additional influence of sadness on in-dulgent consumption, which depends on the activation of ahedonic eating goal. A hedonic eating goal typically resultsin indulgent consumption because a short-term focus onhedonic experiences takes precedence over long-term con-cerns like weight maintenance, disease prevention, and lon-gevity (Tice and Bratslavsky 2000). Sadness can mitigatethis influence of a hedonic eating goal because one functionof sadness is to make a person more vigilant about pre-venting future losses (Lazarus 1991; Lench, Flores, andBench 2011). Consequently, when a hedonic eating goal isactive, an experience of sadness should increase a person’ssensitivity to the potentially harmful consequences of in-dulgent consumption, which, in turn, should decrease thedesire to indulge. In contrast, when sadness is experiencedin the absence of an active goal, these harmful consequencesof indulgence are less apparent. As a result, sadness shouldfunction as a signal to regulate one’s negative emotions andresult in increased consumption of indulgent foods as ameans of feeling better (Andrade and Cohen 2007; Tice,Bratslavsky, and Baumeister 2001).

In the next section, we draw from the functionalist per-spective to better understand how sadness can decrease orincrease indulgent consumption. We begin our discussionwith an overview of the functionalist perspective, so as toprovide a foundation for understanding how emotions canserve as informational sources that can modify or initiategoal-directed behavior (Keltner and Gross 1999; Stein, Li-wag, and Wade 1996). We then transition into a discussionof why sadness is uniquely suited to modify hedonic eatinggoals. We close with a discussion of how people shouldrespond to sadness in the absence of active goals. Fourstudies test our hypotheses.

FUNCTIONALIST PERSPECTIVEThe functionalist perspective is a framework predicated

on the assumption that emotions facilitate adaptive re-sponses to the environment (Campos et al. 1994; Keltnerand Gross 1999; Lazarus 1991). The perspective assumesthat emotions are an evolutionary adaptation, and it is usedto explain how emotions influence different classes of be-havior. For example, evolutionary functionalists focus onhow emotions are necessary for survival and reproduction(Cosmides and Tooby 2000; Griskevicius, Shiota, and Neu-feld 2010; Tooby and Cosmides 1990). Social functionalistsfocus on how emotions help people engage in efficaciousinterpersonal, social, and cultural behaviors (Keltner, Haidt,and Shiota 2006). Communicative functionalists focus onhow emotional expression alters the behavior of others(Saarni et al. 2006). Constructionists focus on how parentalinteraction and childhood socialization influence how emo-tions are experienced and adapted in order to facilitate togoal pursuit (Thompson and Lagattuta 2006). The processesthat facilitate these adaptive responses vary by class of be-

havior. Some research streams speak primarily to how emo-tions initiate goal pursuit (e.g., social functionalism), someto how emotions facilitate or inhibit goal pursuit (e.g., com-municative functionalism), and some to how emotions mod-ify goal pursuit (e.g., evolutionary functionalism, construc-tionism).

The functionalist framework allows for a specific emotionto exert a different type of influence on each type of goal(i.e., an emotion may facilitate the pursuit of goal A butinhibit the pursuit of goal B). Functionalists account for thisadaptability through their conceptualization of an emotion.Functionalists assume that each emotion consists of a set ofneural programs (Tooby and Cosmides 1990). These neuralprograms can activate motivational subroutines that facili-tate adaptive responses to the environment (DeSteno et al.2006; Griskevicius, Shiota, and Nowlis 2010). The moti-vational subroutines are goal- and situation-specific (Toobyand Cosmides 1990). In effect, an emotion increases theaccessibility of an array of neural programs, of which thesubset that is applicable depends on the specific goal and/or context.

Emotion-Driven Goal Modification

Functionalism conceptualizes goal pursuit as an interac-tion between a person and an environment. The fact that aperson-environment interaction is ongoing (and environ-ments are complex) implies that emotions can modify goalpursuit in multiple ways. First, emotions can be used asfeedback during goal pursuit (Louro, Pieters, and Zeelenberg2007). For example, successful (unsuccessful) goal pursuitcan generate satisfaction (frustration) which, in turn, en-courages (discourages) continued pursuit of the goal (Iliesand Judge 2005). Second, emotions can be recruited as ameans of modifying goal pursuit (Tooby and Cosmides2008). For example, people can use memories or socialinteraction to build excitement (indifference) that then en-courages (discourages) continued goal pursuit, even thoughthe goal pursuit itself is not responsible for the excitement(indifference). Third, incidental emotion (i.e., emotion un-related to the goal pursuit itself) can modify goal pursuit(Tooby and Cosmides 1990). For example, incidental pridecan discourage the selection of thrifty means when a goalto save money is salient (Wilcox, Kramer, and Sen 2011).Similarly, incidental fear can discourage the pursuit of goalsthat have immediate consequences for an individual, evenif the goal pursuit is not responsible for the experience offear (Lerner and Keltner 2000).

Evidence that people are sensitive to emotions that areincidental to the pursuit of currently active goals impliesthat people have an implicit understanding of the usefulnessof emotions. If this is so, there should be situational factorsthat determine when an emotion will function as informationthat can modify goal pursuit. First, emotions should modifygoal pursuit when the subroutines that define the emotionare relevant to the goal and/or situation. To understand thisproposition, consider a person with a goal of socializingwho subsequently experiences anger. One common subrou-



SALERNO, LARAN, AND JANISZEWSKI 137

tine of anger is to lash out at others, even others who werenot responsible for the anger (Carver and Harmon-Jones2009). To the extent that people experiencing anger cananticipate the potentially negative consequences of pursuinga social goal (i.e., the negative consequences of lashing outat innocent others), they can attenuate the pursuit of thegoal.

Second, the ability of an emotion to modify goal pursuitshould depend on the timing of the emotional experiencerelative to the goal activation. If an emotion is experiencedprior to goal activation, it is possible that the emotion willinitiate its own goal-directed behavior. This proposition ispredicated on the understanding that each emotion is com-posed of motivational action tendencies (Frijda 2005). Asa result, once emotion-initiated behavior is underway, sub-sequent goal activation may be ignored or overridden if thegoal does not fit with the emotion-initiated behavior (Toobyand Cosmides 1990). For instance, a common subroutine ofhappiness is to work at tasks geared toward self-improve-ment (Trope, Igou, and Burke 2006). To the extent thathappiness initiates behavior geared toward self-improve-ment (e.g., exercising), the functional response to the sub-sequent activation of a goal that is incongruous with currentbehavior (e.g., relaxing) may not alter the emotion-initiatedbehavior. Taken together, the potential for an emotion tomodify goal pursuit should depend on its relevance to thegoal and when it is experienced relative to the activation ofthe goal.

Sadness and a Hedonic Eating Goal

Sadness is an emotion that elicits a sense of loss (Frijda2005; Lazarus 1991; Lench et al. 2011). This sense of losscan relate to any resource that affects one’s health or overallwell-being, such as the death of a loved one, an encounterwith a debilitating ailment, or a breakup with a significantother (Keller and Nesse 2006; Nesse 2006; Raghunathanand Pham 1999). A central adaptive function of sadness isto prompt personal reflection in response to the sense ofloss, allowing one to assess and potentially revise subsequentgoal-directed behavior (Bonanno, Goorin, and Coifman 2008;Welling 2003). How sadness influences indulgent consump-tion will depend on the functional response to the sense ofloss. Based on our description of the functionalist perspec-tive, we propose that there are two functional responses aconsumer can adopt: (1) prevent further loss; (2) regulatethe emotion.

A “loss-preventing response” occurs when the functionalresponse to sadness is to accept the loss and adopt a higherlevel of vigilance, in order to prevent further loss. Consistentwith this proposition, Nesse (2000, 2006) has argued that afunctional adaptation of sadness is to enhance a person’sability to detect and inhibit potentially harmful behaviorsthat could lead to further loss. Similarly, Stearns (1993) hasproposed that sadness should heighten a person’s sensitivityto harm in the presence of cues that have led to losses inthe past. In this way, sadness acts as an associative bridge,drawing from past losses to inform situations in which future

losses could occur. However, since not all loss is preventable,sadness leads people to focus on their goal pursuit becausegoal pursuit is one potential avenue for further loss that isunder the person’s control. This adaptive influence of sad-ness is consistent with previous research discussing the abil-ity of sadness to encourage a review of and/or alter goalsowing to a realization of the potential for loss (Power 1999).The implication is that the loss-preventing response of sad-ness should occur in the presence of goals that have thepotential for harm, as when one pursues a hedonic eatinggoal (Laran 2010b). Unhealthy eating is a threat to one’sfitness and is associated with feeling bad about oneself (Ticeand Bratslavsky 2000), overeating (Cochran and Tesser 1996),and health problems (Chandon and Wansink 2012). Thus, theloss-preventing response should modify a hedonic eatinggoal and decrease indulgent food consumption.

Although there is no direct evidence linking hedonic goalactivation, sadness, and a reduction in indulgent consump-tion, there are several findings suggesting indulgence is con-sidered to be a source of further loss. First, we know thatthe negative long-term consequences of indulgence impactconsumers in a number of ways, as reflected in consumeroverspending (Baumeister 2002), weight gain and obesity(Sharpe, Staelin, and Huber 2008; Wang and Beydoun 2007),superfluous materialism (Ordabayeva and Chandon 2011),and addiction (Grant et al. 2010). Second, consumers knowthey are susceptible to the negative consequences of indul-gence (Cochran and Tesser 1996; Polivy and Herman 1985).Third, there are instances in which hedonic impulses areoverridden by the activation of virtuous goals (Fishbach etal. 2003; Laran 2010a). The implication is that consumerscan learn to anticipate the negative consequences of theirindulgent behavior, as long as the context enables learning.

The foregoing predictions pertain to situations in whicha hedonic eating goal is active. When there is no active goal,or sadness is not informative to an active goal, sadnessshould initiate an “emotion regulation response” (Gross andThompson 2007; Williams et al. 2009). Emotion regulationrefers to strategies for amplifying, dampening, or sustainingan emotion (Gross and Thompson 2007). People in a neg-ative emotional state often engage in actions that will makethem feel better (Magen and Gross 2010), including indul-gent consumption (Andrade and Cohen 2007). For example,Tice et al. (2001) found that consumers ate more comfortfood when they read a story about a child who died in acar accident as opposed to a story about a child who wassaved. Garg, Wansink, and Inman (2007) showed that watch-ing a sad movie led consumers to eat more popcorn thanwatching a happy movie. In each of these cases, the presenceof a negative emotion initiated a strategy to mitigate theemotion, and the context encountered after (e.g., availableindulgent food) only facilitated, rather than inhibited, theuse of the strategy.

Summary

The following studies examine the goal-dependent influ-ence of sadness on indulgent eating behavior, the processes




responsible for this influence, and its boundary conditions.We predict that in the presence of a hedonic eating goal,sadness will exert a loss-preventing response, which willincrease one’s sensitivity to the potentially harmful conse-quences of indulgence and reduce indulgent consumption.In the presence of goals that do not have potentially harmfulconsequences, or when no goals are active, sadness willencourage emotion regulation and indulgent consumptionwill occur.

STUDY 1Study 1 investigated the influence of sadness when there

was an active hedonic eating goal (henceforth called “he-donic goal,” for brevity). A hedonic goal was primed byhaving participants create a list of indulgent activities (Fish-bach and Labroo 2007). Sadness was induced via a scenario-based writing task. Indulgent consumption was assessed byproviding participants an opportunity to consume M&M’s(Garg et al. 2007; McFerran et al. 2010). We predicted thatsad people would indulge more when no goal was activebecause of emotion regulation efforts. We predicted that sadpeople would indulge less when a hedonic goal was activebecause a loss-preventing response would encourage anavoidance of behaviors with potentially harmful long-termconsequences.

The study included additional experimental conditions in-tended to increase our confidence that the loss-preventingresponse was not characteristic of any negative emotion. Totest this possibility, anger and fear conditions were includedin the study. These emotion conditions are informative be-cause the functionalist perspective predicts that neither ofthese emotions should discourage indulgent consumptionwhen a hedonic goal is active.

Anger is an emotion that is accompanied by a sense offrustration with the external agents responsible for preventingdesirable outcomes (Carver and Harmon-Jones 2009; Steinand Levine 1990). To the extent a hedonic goal is active,the functional response to anger should be to overcome anyperceived obstacles to indulgent consumption (Izard 1993).Given an environment where there are no obstacles, as wasthe case in our procedure, anger should not influence theamount of indulgent consumption when a hedonic goal isactive. Fear is an emotion that is accompanied by an in-creased sensitivity to imminent danger (Griskevicius et al.2009; Ohman and Mineka 2003). For example, commonsources of fear are physical objects or events (e.g., heights,thunder), interpersonal situations (e.g., social exclusion),and animals (e.g., snakes, spiders; Ohman and Mineka2001). The functional response to fear is to ignore long-term consequences and focus on the present danger. Thus,the myopic focus that accompanies fear should make it non-informative to hedonic goal pursuit (i.e., it should not in-fluence the amount of indulgent consumption). It should benoted that the functionalist prediction differs from that madeby appraisal theory, which posits that fear leads to an ap-praisal of uncertainty and a lack of control (Raghunathanand Pham 1999). Uncertainty and a lack of control should

discourage goal persistence and reduce indulgent consump-tion.

Method

Participants and Design. Participants were 239 Univer-sity of Miami students (47% female, Mage p 19.82) whoparticipated in exchange for course credit. The design wasa 2 (goal prime: neutral vs. hedonic) # 4 (emotion: noemotion, anger, fear, sadness) between-subjects design. Allparticipants were asked to indicate if they were willing toconsume food. In this and all other consumption studies,participants were excluded from taking the study if theyindicated that they were unwilling to consume food.

Procedure and Stimuli. Participants entered a behaviorallab and were individually seated at desks in front of personalcomputers. Participants were then told that there were threestudies in the session, supposedly unrelated and pooled to-gether out of convenience. The first study was a goal primingtask, adapted from Fishbach and Labroo (2007). Participantswere told that the researchers were interested in undergrad-uate students’ activities. Participants in the neutral goalprime condition listed activities that they performedthroughout the course of a normal day. Participants in thehedonic goal prime condition were told that people seekpleasure and that the researchers wanted to learn about thetypes of activities that they used to indulge themselves. Theywere then asked to list indulgent activities.

After completing the priming task, all participants movedto a second study that supposedly investigated how wellpeople were able to immerse themselves in hypotheticalsituations. Unknown to participants, each person received adifferent emotion manipulation. Participants in the sadnesscondition read a scenario entitled “Breakup,” in which theperson in the passage experiences intense sadness at the endof an extended romantic relationship. Participants in theanger condition read a scenario entitled “Computer Trouble”in which the person in the passage experiences a great dealof difficulty getting technical support. Participants in thefear condition read a scenario entitled “Turbulence” in whichthe person in the passage is on a turbulent flight that beginsa rapid descent. Participants in the no-emotion control groupread a scenario entitled “Cleaning Day,” which describedan emotionally neutral scenario where the person in thepassage spends some time cleaning their home. Participantsin all four conditions were asked to read the passage andimagine, as vividly as possible, what it would feel like tobe in that situation. To strengthen the manipulation, partic-ipants were asked to spend a few minutes writing about howthey felt while reading the passage and to describe an ex-perience in which they felt the same way.

In the third study, participants were told that they wouldbe taking part in a taste test that investigated the compati-bility of foods with different videos. The instructions ex-plained that participants would have the opportunity to con-sume M&M’s while watching a YouTube video and thatthey could consume as many M&M’s as they wanted. Par-




TABLE 1

EMOTION PRETEST RESULTS

No emotion Anger Fear Sadness

Emotion ratings At baseline Post induction At baseline Post induction At baseline Post induction At baseline Post induction

Anger 3.14a1 3.19a1 3.20a1 5.40b2 3.70a1 3.40a1 2.94a1 3.39a1

(2.39) (2.68) (2.21) (2.48) (2.68) (1.85) (2.13) (2.36)Fear 3.67a1 3.48a1 3.85a1 4.15a1 4.50a1 6.60b2 4.33a1 4.39a1

(2.01) (2.48) (2.48) (2.66) (2.16) (1.57) (2.91) (2.48)Sadness 2.81a1 2.67a1 3.45a1 3.20a1 4.20a1 4.35a1 3.22a1 5.83b2

(2.23) (2.15) (2.24) (2.02) (2.90) (1.84) (2.65) (2.53)Overall mood 5.00a 5.67a 4.75a 3.40b 4.85a 3.70b 6.50a 3.83b

(1.76) (2.11) (2.29) (1.73) (2.21) (1.92) (1.98) (2.31)

NOTE.—Superscript letters denote differences (p ! .05) between means in the same emotion condition. Superscript numbers denote differences(p ! .05) between means within the same time measurement column. Standard deviations are in parentheses.

ticipants were also told that after watching the video, theywould be asked a few follow-up questions about how wellthe food went with the video. In the corner of each partic-ipant’s desk sat a bag of M&M’s placed inside a Styrofoambowl, which was covered by a napkin. Participants weretold to remove the napkin from the top of the bowl, openand empty out the entire bag of M&M’s into the bowl, andthen proceed with watching the video. Every participantwatched a video about how to make origami. The video wasapproximately 4 minutes long and selected because of itsneutral content. Participants were told that they could stopwatching the video at any point and to call over the exper-imenter if they desired to move on to the next part of thetask. Bowls were collected when participants called the ex-perimenter or the video ended. Bowls were subsequentlyweighed to determine consumption.

To maintain the cover story, participants then answereda few questions about the task itself, including how com-patible the M&M’s were with the video they watched. Par-ticipants were then told that they would also answer a fewgeneral questions about M&M’s. Our critical question askedto what extent eating foods like M&M’s have the potentialto lead to health problems later in life (1 p “strongly dis-agree,” 9 p “strongly agree”). This measure of the potentialharm from indulgence was expected to mediate the influenceof the hedonic goal priming in the sadness condition butnot in the other emotion conditions. Participants were thenextensively debriefed for suspicion, told about the real pur-pose of the experiment, thanked for their help, and dis-missed. No participant guessed the real purpose of the studyor how the goal priming and emotion task could jointly haveinfluenced their food consumption.

Pretest of Emotion Manipulations

A pretest (N p 79) assessed the efficacy of the emotionmanipulations (see table 1). Participants were asked to in-dicate how sad, angry, and afraid they felt at baseline (priorto the hypothetical scenario) and after the emotion manip-ulation on 9-point scales (1 p “not at all,” 9 p “ex-tremely”). Participants also indicated their overall mood (1p “negative mood,” 9 p “positive mood”). Participants in

the anger, fear, and sadness conditions felt significantly moreof their induced emotion after the emotion manipulation.Participants in the anger, fear, and sadness conditions didnot differ in negative mood after the emotion manipulation(F ! 1), which suggests that any observed differences be-tween angry, fearful, and sad participants is not likely aresult of differences in mood.

Results

Consumption. The means indicating the amount of in-dulgent consumption, in this and all subsequent studies, arepresented in table 2. An ANOVA on the amount eaten re-vealed an interaction between the goal priming and emotionfactors (F(3, 231) p 5.51, p ! .01; see fig. 1). In the neutralgoal prime condition, participants in the anger (M p 23.72;F(1, 231) p 8.40, p ! .01), fear (M p 19.43; F(1, 231) p3.12, p p .08), and sadness conditions (M p 24.66; F(1,231) p 9.80, p ! .01) ate more M&M’s than in the no-emotion condition (M p 13.33). These results suggest thatin the absence of an active goal, negative emotions willprompt emotion regulation. In the hedonic goal prime con-dition, participants made to feel sad consumed fewerM&M’s (M p 12.15) than participants in the no-emotion(M p 22.89; F(1, 231) p 6.94, p p .01), anger (M p26.01; F(1, 231) p 12.28, p ! .01), and fear conditions (Mp 23.25; F(1, 231) p 7.63, p ! .01). These results areconsistent with the loss-preventing response of sadness. Inthe no-emotion condition, participants consumed moreM&M’s in the hedonic goal prime condition (M p 22.89)than in the neutral goal prime condition (M p 13.33; F(1,231) p 5.84, p p .01). This result verifies the efficacy ofthe hedonic goal priming manipulation.

Comparisons of priming conditions within each emotionwere also consistent with the hypothesized loss-preventingresponse of sadness. In the sadness condition, participantsin the hedonic goal prime condition consumed fewerM&M’s (M p 12.15) than in the neutral goal prime con-dition (M p 24.66; F(1, 231) p 9.63, p ! .01). In the angercondition, participants consumed similar amounts ofM&M’s in the neutral goal prime (M p 23.72) and hedonicgoal prime conditions (M p 26.01; F ! 1). Similarly, in




TABLE 2

STUDY SUMMARY STATISTICS

Study 1

No emotion Anger Fear Sadness

Neutralprime

Hedonicprime

Neutralprime

Hedonicprime

Neutralprime

Hedonicprime

Neutralprime

Hedonicprime

Consumption 13.33a 22.89b 23.72a 26.01a 19.43a 23.25a 24.66a 12.15b

(11.34) (16.68) (16.27) (15.77) (15.14) (16.02) (15.98) (14.85)Potential harm 4.63a 3.77a 3.38a 3.43a 4.07a 3.55a 3.93a 5.57b

(2.74) (2.80) (2.47) (2.60) (2.28) (2.68) (2.14) (2.70)

Study 2

No emotion Sadness

Creativity prime Hedonic prime Creativity prime Hedonic prime

Control Felt safety Control Felt safety Control Felt safety Control Felt safety

Consumption 5.73a 5.90a 7.16a 6.70a 7.07a 6.82a 5.45a 7.33b

(3.29) (3.04) (2.87) (3.01) (2.94) (3.16) (3.30) (2.63)Potential harm 5.47a 5.63a 5.15a 5.43a 5.88a 5.17a 6.23a 4.81b

(2.29) (2.60) (2.51) (2.50) (2.16) (2.26) (2.49) (2.38)

Study 3

No emotion Sadness

Beforeemotion

Afteremotion

Beforeemotion

Afteremotion

Choice share 66.7a 61.8a 42.4a 79.2b

Potential harm 4.30a 4.80a 5.63a 4.58b

(2.79) (2.95) (2.83) (2.69)

Study 4

Neutral prime Hedonic prime Strong hedonic prime

No emotion Sadness No emotion Sadness No emotion Sadness

Choice share 46.4a 64.4b 65.5a 47.3b 82.5a 29.8b

NOTE.—In study 1, different superscripts denote differences (p ! .05) between means in the same emotion condition. In study 2,different superscripts denote differences (p ! .05) between means in the same emotion and prime condition. In study 3, differentsuperscriptsdenote differences (p ! .05) between means in the same emotion and timing condition. In study 4, different superscripts denote differences(p ! .05) between means in the same prime and emotion condition. Standard deviations are in parentheses.

the fear condition, participants consumed similar amountsof M&M’s in the neutral goal prime (M p 19.43) andhedonic goal prime conditions (M p 23.25; F ! 1). Theseresults demonstrate that the hedonic goal-based reduction inindulgence was specific to sadness. Not all negative emo-tions decreased indulgent consumption when a hedonic goalwas active.

Mediation by Potential Harm from Indulgence. The re-sults of the harm-from-indulgence measure revealed a goalpriming by emotion interaction (F(3, 231) p 2.73, p ! .05).The hedonic prime led to expectations of more potentialharm from indulgence in the sadness condition (Mneutral p3.93, Mhedonic p 5.57; F(1, 231) p 5.92, p ! .01), but notin the no-emotion condition (Mneutral p 4.63, Mhedonic p 3.77;F(1, 231) p 1.72, p 1 .10), anger (Mneutral p 3.38, Mhedonic p3.43; F(1, 231) p .01, p 1 . 10), or fear condition (Mneutral p4.07, Mhedonic p 3.55; F(1, 231) p .60, p 1 .10).

We predicted that expectations of more potential harmfrom indulgence would mediate the influence of the goal

prime on the amount of M&M’s consumed in the sadnesscondition, but not in the other emotion conditions. To testthis prediction, we used Mplus to perform a multi-groupanalysis for moderated mediation. A benefit of Mplus, overa macro such as PROCESS (Hayes 2013), is its ability toperform the simultaneous estimation, and comparison, ofthe indirect effects of a mediator (potential harm from in-dulgence) at the various levels of a multicategorical mod-erator (emotion). Supporting our predictions, the pathwayfrom goal prime to consumption through potential harm wassignificant and did not include zero in the sadness condition(indirect effect p �3.93; 95% confidence interval [CI]:�9.81 to �.52), but was not significant in the anger (indirecteffect p .05; 95% CI: �1.64 to 1.60), fear (indirect effect p.26; 95% CI: �1.13 to 2.21), or no-emotion (indirect effectp .50; 95% CI: �.73 to 2.50) conditions. Additionally, atest comparing the indirect effect in the sadness condition tothe indirect effect in the no-emotion condition was significant(difference in indirect effects p �4.43; 95% CI: �10.47 to




FIGURE 1

STUDY 1 RESULTS (ERROR BARS SHOW 95%CONFIDENCE INTERVALS)

�.87), while the same emotion-control condition comparisontests for anger (difference in indirect effects p �.45; 95%CI: �3.14 to 1.45) and fear (difference in indirect effects p�.24; 95% CI: �2.57 to 2.28) were not significant.

Ancillary Study

A common strategy for bolstering evidence about a pro-posed mediating process is to measure an individual dif-ference variable that correlates with the mediator (e.g., Hongand Sun 2012; Ramanathan and Menon 2006; Winterich andHaws 2011). Given our claim that sadness increases a per-son’s sensitivity to the potential negative consequence ofindulgent consumption (a mediator), we sought to isolatean individual difference variable that would increase theimpact of this variable, thus, providing additional supportfor its relevance. Nenkov, Inman, and Hulland (2008) showthat people become better at self-control when they focuson potentially negative outcomes. People that are able torecognize the link between their short-term behaviors andthe long-term consequences of their behaviors are less im-pulsive. We expect that people with this trait should be mostsensitive to the combined effects of hedonic goal pursuitand sadness on indulgent consumption. Therefore, we mea-sured people’s tendency to focus on the potential for neg-ative outcomes using the negative-outcome-focus subscalefrom the elaboration on potential outcomes (EPO) scale(Nenkov et al. 2008). We hypothesized that people who weremost sensitive to negative outcomes should show thegreatest decrease in indulgent consumption owing to sad-ness, provided a hedonic goal was active.

Procedure. The experiment was a two-cell design con-sisting of the neutral and hedonic goal prime conditions in

the sadness condition of study 1. The procedure was iden-tical to study 1 except for the following changes. First,participants began the session by answering questions thatwere purported to help researchers better understand thecharacteristics of the student body. Participants were askedtheir age, gender, and 10 personality questions, includingfour items from the negative-outcome-focus subscale of theEPO scale (1 p “strongly disagree,” 7 p “strongly agree”;Nenkov et al. 2008). The items were “I tend to think a lotabout the negative outcomes that might occur as a result ofmy actions,” “When thinking over my decisions I focus moreon their negative end result,” “I often worry about whatcould go wrong as a result of my decisions,” and “I amoften afraid that things might turn out badly.” Second, goalpriming was accomplished using print ads. Participants inthe neutral (hedonic) goal prime condition rated the effec-tiveness of six neutral (hedonic) print ads on a scale rangingfrom 1 (“not at all effective”) to 9 (“very effective”). Neutralads were for a refrigerator, washing machine, short filmfestival, jeans, electric car, and trading company. Hedonicads were for chocolate cake, pizza, French fries, chips, ahamburger, and cinnamon rolls. The emotion manipulationand M&M’s eating task were the same as study 1, exceptthat the video was about how to fix a showerhead.

Participants and Results. Participants were 128 Univer-sity of Miami students (52% female, Mage p 20.22) whoparticipated in exchange for course credit. Sad participantsin the hedonic goal prime condition consumed fewerM&M’s (M p 14.59) than in the neutral goal prime con-dition (M p 19.97; t(126) p 2.34, p ! .05), which replicatesthe results of study 1. Moreover, the interaction between thenegative outcome focus (a p .89) and goal prime manip-ulation was significant (b p �2.98, p ! .05). To explorethe nature of this interaction, we performed a spotlight anal-ysis (see fig. 2; Fitzsimons 2008). At 1 standard deviation(SD) below the mean of negative outcome focus (i.e., lowfocus on potential negative outcomes), participants in theneutral and hedonic goal prime conditions did not differ inthe amount of M&M’s eaten (b p �.51, p 1 .40). At 1 SDabove the mean of negative outcome focus (i.e., high focuson potential negative outcomes), participants ate fewer ofthe M&M’s in the hedonic goal prime condition than in theneutral goal prime condition (b p �9.30, p ! .01).

Discussion

Study 1 provides support for the claim that sadness en-courages emotion regulation when no goal is active but aloss-preventing response when a hedonic goal is active. Emo-tion regulation was evident in the neutral goal prime condi-tion, where sadness increased indulgent consumption relativeto the no-emotion control group. The loss-preventing responsewas evident in the hedonic goal prime condition, where sad-ness reduced indulgent consumption relative to the no-emo-tion control group. The reduced consumption in the hedonicgoal prime condition of the sadness condition depended onparticipants’ concerns about future harm (study 1). Moreover,




FIGURE 2

ANCILLARY STUDY RESULTS

participants with a negative outcome focus were most sus-ceptible to the combined influence of the goal prime andsadness (ancillary study). Finally, similar results were notobtained for anger and fear, suggesting that these emotionswere not informative with respect to the pursuit of the hedonicgoal. Instead, emotion regulation dictated the response in theanger and fear conditions.

STUDY 2

Study 2 was designed with two objectives in mind. Thefirst objective was to provide additional evidence for theloss-preventing response of sadness. We have argued thatthe loss-preventing response makes people more sensitiveto the potentially harmful consequences of engaging in in-dulgent consumption. If this is so, the influence of a loss-preventing response should be mitigated when people aremade to feel safe and secure. This hypothesis is consistentwith the tendency to seek attachment and security followinga loss (Bonanno et al. 2008; Bowlby 1969). One way inwhich people can be made to feel safe is through imageryvisualization (Gershoff and Koehler 2011). To the extentthat one is able to create a sense of safety, there should beless sensitivity to the potential for long-term harm and in-dulgent consumption should occur more freely.

The second objective was to show that the loss-preventingresponse of sadness only exerts an influence in the presenceof goals that have potentially harmful consequences associ-ated with their pursuit. If a previously activated goal is notinherently harmful, then sadness should not impact pursuit.To test this prediction, we included a condition in which acreativity goal was primed. We predicted that sadness would

not be informative for a creativity goal because the pursuitof the goal has minimal potential for harmful consequences.As a result, we predicted that the presence of this goal wouldnot stop consumers from indulging as a way to feel better(i.e., emotion regulation would take precedence).

These hypotheses were investigated using a consumptioncontext similar to study 1, except that participants wereallowed to eat chocolate chip cookies instead of M&M’s.When an emotion regulation response was active (i.e., acreativity goal was active or participants felt safe), partic-ipants should have an increased desire for the indulgent food.The desire to engage in emotion regulation should make thefood more attractive. When a loss-preventing response wasactive (i.e., when a hedonic goal was active and participantswere sad), participants should have a reduced desire for theindulgent food. This should occur because the loss-pre-venting response focuses on behaviors that have long-terminstrumental value. When participants were made to feelsafe, the loss-preventing response should be mitigated, andemotion regulation should occur.

Method

Participants and Design. Participants were 331 Univer-sity of Florida students (51% female, Mage p 20.07) whoparticipated in exchange for course credit. The design wasa 2 (goal prime: hedonic vs. creativity) # 2 (emotion: noemotion vs. sadness) # 2 (safety: control vs. felt) between-subjects design.

Procedure and Stimuli. Participants entered a behaviorallab and were individually seated in front of personal com-puters. Participants were then told that there were three stud-ies in the session, supposedly unrelated. The first study wasa goal priming task. As in study 1, participants in the hedonicgoal priming condition were told that people seek pleasureand that the researchers wanted to learn about these activ-ities. They were then asked to list their favorite indulgentactivities. Participants in the creativity goal priming con-dition were told that people seek to exhibit their creativityin a variety of ways and that the researchers wanted to learnabout these activities. They were then asked to list theirfavorite creative activities.

All participants then moved to a second study that ad-ministered the emotion manipulation. The sadness and no-emotion manipulations were identical to those used in study1. Participants in the sadness (no-emotion) condition readthe “Breakup” (“Cleaning Day”) scenario, imagined what itwould feel like to be in that situation, and then spent a fewminutes writing about how they felt while reading the pas-sage and describing a similar personal experience. After theemotion manipulations, participants in the felt safety andcontrol conditions were told, “Before you continue on withthe next study, it is important that you take a moment toclear your mind using an imagination task.” Participants inthe felt safety condition were asked to “close your eyes andthink about how family provides us with lasting securityand protection” and to “take the next minute to close your




FIGURE 3

STUDY 2 RESULTS (ERROR BARS SHOW 95% CONFIDENCE INTERVALS)

eyes and imagine how family helps us to know that every-thing will be okay.” Participants in the control conditionwere asked to “close your eyes and think about a blank,white wall in front of you” and “take the next minute toclose your eyes and imagine what this wall would look like.”

In the third study, participants were told that they wouldbe taking part in a taste test that investigated the compati-bility of foods with different videos. The instructions ex-plained that participants would have the opportunity to con-sume cookies while watching a YouTube video. As in study1, participants were told that they could consume as muchor as little as they wanted and that they would be askedfollow-up questions after the video. Participants were askedto open a 2-ounce bag of Famous Amos Chocolate ChipCookies that had been placed and covered with a napkin inthe corner of their carrel. Participants then watched a 4-minute video about how to make origami. After watchingthe entire video, participants were instructed to stop eatingand to record the number of uneaten cookies. This numberwas subtracted from 10 to calculate the total number ofcookies consumed. Afterward, participants indicated howcompatible the cookies were with the video, as well as somegeneral questions about cookies. Our critical question askedto what extent eating foods like cookies have the potentialto lead to health problems later in life (1 p “strongly dis-agree,” 9 p “strongly agree”). This measure of the potentialharm from indulgence was expected to mediate the influenceof the hedonic goal priming in the sadness condition butnot in the no-emotion condition. Participants were then ex-tensively debriefed for suspicion, told about the real purposeof the experiment, thanked for their help, and dismissed. Noparticipant guessed the real purpose of the study or how thegoal priming and emotion task could jointly have influencedtheir food consumption.

ResultsConsumption. An ANOVA revealed a three-way inter-

action of emotion, goal prime, and the safety induction (F(1,

323) p 4.22, p ! .05; see fig. 3). The analysis of thisinteraction was conducted in two parts. The first part con-firmed that there was (1) an emotion regulation effect and(2) a hedonic goal priming effect. Participants primed withthe creativity goal were used to test for the emotion regu-lation effect. Consistent with emotion regulation, sad par-ticipants (M p 6.94) ate more cookies than no-emotionparticipants (M p 5.82; F(1, 323) p 5.40, p ! .05). Theno-emotion participants were used to test for hedonic goalpriming. Consistent with hedonic goal priming, participantsin the hedonic goal prime condition (M p 6.93) ate morecookies than participants in the creativity goal prime con-dition (M p 5.82; F(1, 323) p 5.53, p ! .05).

The second part of the analysis focused on (1) the loss-preventing response that should emerge from a combinationof the hedonic goal prime and sadness and (2) the mitigationof the loss-preventing response when a participant was madeto feel safe. Both effects were investigated in the hedonicgoal prime condition, where there was an interaction ofemotion and the safety induction (F(1, 323) p 6.59, p p.01). Participants in the control condition were used to testfor the loss-preventing response. Consistent with the hy-pothesis of a loss-preventing response, sad participants atefewer cookies (M p 5.45) than no-emotion participants (Mp 7.16; F(1, 323) p 7.22, p ! .01). Participants in the sadcondition were used to assess if felt safety mitigated theinfluence of the sadness. Consistent with the hypothesis ofa mitigating influence of safety, participants in the felt safetycondition ate more cookies (M p 7.33) than those in thecontrol condition (M p 5.45; F(1, 323) p 7.31, p ! .01).

Mediation by Potential Harm from Indulgence. An ANOVAon the potential-harm-from-indulgence measure did not re-veal a significant three-way interaction of emotion, goalprime, and the safety induction (F ! 1). Our main focus,however, was on the prediction that the perceived potentialfor harm from indulgence should only be responsible forthe reduced consumption for hedonically primed sad par-




ticipants when safety was not felt. Therefore, the effect ofemotion on consumption should be moderated by the goalprime and safety induction and mediated by the potentialharm from indulgence in the hedonic goal prime, controlcondition but not in the hedonic goal prime, felt safety con-dition. We tested this prediction by conducting a moderatedmediation analysis using model 10 in the PROCESS macro(Hayes 2013), which allows for the analysis of two mod-erators (goal prime and safety induction) and their respectiveinteractions with the predictor variable (emotion). Support-ing this prediction, in the hedonic goal prime, control con-dition, the pathway from the emotion manipulation to con-sumption through potential harm (indirect effect) wassignificant and did not include zero (indirect effect p �.21;95% CI: �.52 to �.01), which supports mediation (Zhao,Lynch, and Chen 2010). In the hedonic goal prime, feltsafety condition, the indirect effect was not significant andincluded zero (indirect effect p .09; 95% CI: �.10 to .39).There was also no evidence for mediation in the creativitygoal prime, control condition (indirect effect p �.15; 95%CI: �.45 to .04) or the creativity goal prime, felt safetycondition (indirect effect p .15; 95% CI: �.03 to .46).Finally, we note that a moderated mediation analysis (model8) within the control condition—with goal prime as thepredictor variable and emotion as the moderating variable—revealed a marginally significant indirect effect of poten-tial harm on consumption in the sadness condition (indirecteffect p .59; 94% CI: �1.17 to �.01). This replicates theindirect effect of potential harm observed in the sadnesscondition of study 1, under the assumption that the creativity(study 2) and neutral (study 1) goal prime conditions areequivalent in terms of their potential for harm when pursued.

Discussion

Study 2 provides evidence that increased sensitivity tothe potentially harmful consequences of indulgence is re-sponsible for the loss-preventing response of sadness. Whena sense of safety was induced, however, people were lesssensitive to the potential harm, which led to indulgent con-sumption. In addition, sadness did not result in a loss-pre-venting response when the active goal (e.g., creativity) wasnot potentially harmful. This implies that goal states areinformative with respect to the response (e.g., emotion reg-ulation, loss-preventing) an emotion (e.g., sadness) will in-duce.

STUDY 3

As discussed earlier, when sadness is experienced in theabsence of a previously activated goal, the default programassociated with the emotion should be initiated, and peopleshould engage in emotion regulation. It is only when a he-donic goal is active prior to a person becoming sad thatsadness signals the potential for harm, as demonstrated instudies 1 and 2. Therefore, when hedonic goal activationfollows the experience of sadness, we expect that the driveto regulate emotion has already taken hold. The addition of

cues suggesting indulgence might even amplify the effectof emotion regulation, resulting in overindulgence.

Method

Participants and Design. Participants were 250 mem-bers of an online panel (Amazon Mechanical Turk, 53%female, Mage p 33.60). The design was a 2 (timing of he-donic goal activation: before emotion vs. after emotion) #2 (emotion: no emotion vs. sadness) between-subjects de-sign. All participants were primed with a hedonic goal.Twenty-three participants failed to fully read the instructionsand were excluded (Oppenheimer, Meyvis, and Davidenko2009), leaving a final sample of 227 used in subsequentanalyses.

Procedure and Stimuli. Participants were told that theywould participate in three unrelated studies. The first studypurportedly investigated cognitive processes associated withunscrambling sentences. In truth, the study was a hedonicgoal priming task. Participants were presented with 10 setsof five words and asked to form sentences by unscramblingthose words. Each sentence contained a hedonically themedword consistent with the prime (flavor, food, taste, enjoy-able, pleasure, delicious, delightful, indulge, diner, and sa-vor). For example, participants unscrambled sets of wordswith hedonic content, such as “has she taste good very.”

The second study was the emotion induction task. It usedthe same no-emotion and sadness scenarios from studies 1and 2. In accordance with the experimental design, the orderin which the hedonic goal priming and the emotion inductionstudies were presented was counterbalanced.

In the third study, participants were told that the research-ers were recruiting people to take part in taste tests thatwould involve eating food. Each participant was asked tochoose between taking part in a “raisin-eating study” or“M&M’s-eating study.” Participants were also told that eachtaste test were equivalent in time and difficulty, and that thiswas a real choice. A pretest (N p 32) confirmed that par-ticipants considered the M&M’s study (M p 5.31) to bemore indulgent (1 p “not indulgent at all,” 7 p “veryindulgent”) than the raisin study (M p 2.25; t(31) p 10.79;p ! .01), and that the M&M’s study (M p 5.75) had greaterpotential for harmful consequences of overindulgence tooccur (1 p “very little potential,” 7 p “a great deal ofpotential”) than the raisin study (M p 2.47; t(31) p 11.11;p ! .01).

After making their choice, participants were asked to in-dicate the extent to which they were “thinking about howindulgent foods can be harmful when consumed in excess”(1 p “not at all,” 9 p “a great deal”). To provide additionalevidence for the sensitivity to the potential harm from in-dulgence, we then asked participants to think about theM&M’s-eating study (the indulgent study) and provide rea-sons for why they would or would not want to take part inthe study. We were interested in how many participantswould provide reasons against the study related to the harm-ful consequences of indulgence.




FIGURE 4

STUDY 3 RESULTS

Results

Choices. A binary logistic regression revealed an inter-action between the timing of hedonic goal activation andemotion on the choice of a taste test (x2(1) p 10.28, p !

.01; see fig. 4). When the hedonic goal was activated beforethe emotion task, participants in the no-emotion conditionwere more likely to select the M&M’s-eating study (66.7%)than participants in the sadness condition (42.4%; x2(1) p7.08, p ! .01). When the hedonic goal was activated afterthe emotion task, participants in the no-emotion conditionwere less likely to select the M&M’s-eating study (61.8%)than participants in the sadness condition (79.2%; x2(1) p3.93, p ! .05).

Mediation by Potential Harm from Indulgence. An ANOVAon the perceived potential harm from indulgence revealedan interaction between the timing of hedonic goal activationand emotion (F(1, 223) p 4.24, p ! .05). When the hedonicgoal was activated before the emotion task, there was agreater potential harm in the sadness (M p 5.63) than inthe no-emotion condition (M p 4.30; F(1, 223) p 6.60, pp .01). When the hedonic goal was activated after the emo-tion task, the potential harm was similar in the sadness (Mp 4.58) and the no-emotion conditions (M p 4.80; F !

1).We used PROCESS model 8 (Hayes 2013) and predicted

that the effect of emotion on food choice would be mod-erated by the timing of the hedonic goal activation andmediated by potential harm in the sadness condition, butnot in the no-emotion condition. Supporting our predictions,when the hedonic goal was activated before the emotiontask, the pathway from emotion condition to study choicethrough potential harm (indirect effect) was significant anddid not include zero (indirect effect p �.42; 95% CI: �.82to �.10), which supports mediation. When the hedonic goalwas activated after the emotion task, the indirect effect wasnot significant and included zero (indirect effect p .07; 95%CI: �.29 to .42), which does not support mediation. Theseresults show that when a person was feeling sad, the po-tential harm from indulgence did not influence food choiceunless the hedonic goal was active prior to the emotionregulation response.

Reasons Related to Harmful Consequences of Indulgence. Toanalyze the qualitative data, we created a variable that wascoded as 0 if the participant did not list any reasons relatedto harm from indulgence and 1 if the participant did. Ex-amples of reasons related to harm from indulgence included“candy is harmful to everyone,” “I tend to over consumeon sweet foods,” “M&M’s are dangerous,” and “M&M’sare a risk to our health.” We then compared the number ofharm-from-indulgence-themed responses across conditions.A binary logistic regression showed an interaction betweenthe timing of hedonic goal activation and emotion on reasons(x2(1) p 4.39, p ! .05). When the hedonic goal was acti-vated before the emotion task, participants in the sadnesscondition were more likely to list reasons related to the harmfrom indulgence (71.2%) than participants in the no-emotion

condition (51.7%; x2(1) p 4.78, p ! .05). When the hedonicgoal was activated after the emotion task, participants in thesadness condition were as likely to list reasons related tothe harm from indulgence (39.6%) as participants in the no-emotion condition (47.3%; x2(1) p .64, p 1 .40).

Discussion

The results shed additional light on when the emotionregulation versus the loss-preventing response of sadnesswill influence indulgent consumption. When the hedonicgoal preceded the experience of sadness, the loss-preventingresponse took hold. People focused on the potentially harm-ful consequences of indulgence and indicated a reduced de-sire for indulgent consumption. When the hedonic goal didnot precede the experience of sadness, emotion regulationtook hold. People focused on emotion regulation and in-dicated an increased desire for indulgent consumption. Theresults are consistent with our conceptualization that theexperience of sadness is informative only when a previouslyactivated goal has potentially harmful consequences asso-ciated with its pursuit.

STUDY 4

If decreasing indulgent consumption is a protective re-sponse to the potentially negative consequences of indul-gence, then increasing the amount of hedonic goal activationshould increase the salience of these consequences and in-fluence indulgent consumption. One way to encourage astronger hedonic goal is to expose people to additional en-vironmental cues associated with this goal. Thus, study 4used a procedure that included both a scrambled sentencetask and exposure to pictures of tempting food, a task thathas been used by Geyskens et al. (2008) to increase indul-




FIGURE 5

STUDY 4 RESULTS

gent consumption. In the absence of sadness, this strongerhedonic goal should result in an increased desire for indul-gent goods. In the presence of sadness, a stronger hedonicgoal should encourage even less indulgent consumption. In-creased activation of the hedonic goal should result in in-creased vigilance and an increased desire to avoid indulgentconsumption.

Method

Participants and Design. Participants were 342 Univer-sity of Miami students (47% female, Mage p 19.97) whoparticipated in exchange for course credit. The design wasa 3 (goal prime: neutral, hedonic, strong hedonic) # 2 (emo-tion: no emotion vs. sadness) between-subjects design.

Procedure and Stimuli. Participants were told that theywould participate in three unrelated studies. The first studywas purportedly investigating cognitive processes associatedwith unscrambling sentences. In truth, the study was a goalpriming task. Participants were presented with 10 sets offive words and asked to form sentences by unscramblingthose words. The materials for the hedonic goal prime con-dition were the same as in study 3. In the neutral goal primecondition, participants unscrambled sets of words with neu-tral content, such as “a had trip they great.” Participants inthe strong hedonic goal prime condition completed one ad-ditional task after unscrambling the hedonically themed sen-tences. These participants were told that we were pretestingstimuli to be used at a later date. They were told that theywould receive two pictures drawn from a larger set and theyshould click on the parts of the picture that they liked most.In truth, everyone in this condition received the same twopictures containing a variety of tempting foods (Geyskenset al. 2008). One picture contained a number of bakery treatsand delectable meats, while the other picture contained slicesof cake, pies, and eclairs.

Study two was the emotion induction task. It used thesame no-emotion and sadness scenarios as the prior studies.The third study pertained to gifts. Participants were told thatthey would receive a gift card for participation in the sessionand that they could choose between a gift card to “a grocerystore” or “a trendy restaurant.” Consistent with previousresearch (Okada 2005), a pretest (N p 32) confirmed thatparticipants considered the gift card for the trendy restaurant(M p 5.25) to be more indulgent than the one for groceries(M p 2.94; 1 p “not indulgent at all,” 7 p “very indul-gent”; t(31) p 6.45; p ! .01). Participants also indicatedthat the trendy restaurant (M p 5.53) had greater potentialfor harmful consequences of overindulgence to occur (1 p“very little potential,” 7 p “a great deal of potential”) thanthe grocery store (M p 4.41; t(31) p 2.60; p p .01).Participants were told that both gift cards were worth thesame amount of money. After participants made their choice,they were extensively debriefed for suspicion.

Results

A binary logistic regression shows an interaction betweenthe goal prime and emotion on gift card choice (x2(1) p28.43, p ! .01; see fig. 5). Within the no-emotion conditions,participants in the hedonic goal prime condition were morelikely to select the trendy restaurant gift card (65.5%) thanparticipants in the neutral goal prime condition (46.4%; x2(1)p 4.22, p ! .05), and participants in the strong hedonicgoal condition were more likely to select the trendy restau-rant gift card (82.5%) than participants in the hedonic con-dition (65.5%; x2(1) p 4.28, p ! .05). These results confirmthat the hedonic and strong hedonic goal prime manipula-tions worked as intended. Within the sadness conditions,however, participants in the hedonic goal prime conditionwere less likely to select the trendy restaurant gift card(47.3%) than participants in the neutral goal prime condition(64.4%; x2(1) p 3.39, p ! .05), and participants in the stronghedonic goal condition were less likely to select the trendyrestaurant gift card (29.8%) than participants in the hedonicgoal prime condition (47.3%; x2(1) p 3.60, p ! .05). Sup-porting the strength of the effect, the interaction betweenthe hedonic and strong hedonic goal prime conditions andemotion on gift card choice remained significant even afterremoving the neutral goal prime condition from the analysis(x2(1) p 7.70, p ! .01). Finally, sadness increased the choiceof the trendy restaurant in the neutral goal prime condition,as would be predicted by emotion regulation (x2(1) p 3.76,p p .05).

Discussion

The results provide further support for the claim that sad-ness can have goal-dependent influences on indulgent con-




sumption. When no emotion was present, people becamemore indulgent as the hedonic goal became stronger (i.e., astronger goal prime led to increased indulgence). When sad-ness was present, people became less indulgent as the he-donic goal became stronger (i.e., loss-preventing led to de-creased indulgence). The results again support the idea thatsadness can exert a protective function. This loss-preventingresponse increased in intensity as the drive to engage in thepursuit of a potentially harmful goal increased.

GENERAL DISCUSSION

This research demonstrates that emotions can generatedistinct functional responses depending on the presence orabsence of certain consumer goals. Study 1 compared sad-ness to other negative emotions (anger and fear) and foundthat only sadness reduced indulgent consumption in the pres-ence of an active hedonic eating goal. The effect of sadnesson decreasing indulgent consumption was driven by an in-creased sensitivity to the potential harm from indulgence.Study 2 found that (1) sadness only attenuated the pursuitof goals that had potentially harmful consequences and (2)inducing a sense of safety attenuated a consumer’s sensi-tivity to these consequences. These findings suggest that theloss-preventing response of sadness only becomes germanein certain contexts. Study 3 showed that sadness only re-duced indulgent consumption when it was experienced after,but not prior to, the activation of a hedonic eating goal. Thisis further evidence that contextual factors determine thefunctional response to emotions. Finally, study 4 showedthat as the drive to indulge increased, sadness became evenmore effective at limiting indulgence. This suggests that theeffectiveness of emotion, as a regulatory agent, is propor-tional to the strength of the drive it counteracts.

This article contributes to research showing that sadnessis an evolutionarily adaptive emotion (Andrews and Thom-son 2009; Keller and Nesse 2006). Research has shown thatone function of sadness is to encourage an increased re-sponsiveness to the environment (Nesse 1990, 2000). Con-sistent with this idea, our findings show that, in certaincontexts, sadness encourages people to identify behaviorsthat are potentially harmful to their long-term fitness. Thismay occur because sadness induces consumers to link pastharmful experiences to current predicaments in order to pre-vent similar losses from occurring in the future (Andrewsand Thomson 2009). In our research, we demonstrate a spe-cific instance in which sadness encourages vigilance anddecreases indulgent consumption, as this is a behavior thatcan lead to further loss and harm. This sensitivity to thepotentially harmful consequences of specific types of goalpursuit should become more important as a person ages, asexperience should help a person learn how to prevent re-peated losses. Related to this idea, Kunzmann and Gruhn(2005) find that the sense of loss that accompanies sadnessbecomes progressively more intense as people grow older.Thus, it is possible that the loss-preventing response of sad-ness becomes stronger as people age.

Individual Differences and Goal/EmotionCombinations

We found that the tendency to focus on negative outcomesincreases the loss-preventing response of sadness. Theremay be other individual difference factors that have similarinfluences. For example, impulsivity might be an individualdifference factor that has an impact on the loss-preventingresponse of sadness in the presence of a hedonic goal. Im-pulsive consumers struggle with resisting indulgent temp-tations (Mukhopadhyay, Sengupta, and Ramanathan 2008;Ramanathan and Williams 2007); consequently, the poten-tially harmful consequences of indulgent consumption maybe more accessible in impulsive consumers’ memories. Al-ternatively, it may be the case that impulsive consumershave chronically active hedonic goals, which could also leadto the loss-preventing response of sadness.

How consumers respond to self-control lapses may alsoprove to be an important individual difference factor to con-sider (Laran and Janiszewski 2011; Zemack-Rugar, Corus,and Brinberg 2012). Consumers who tend to regroup fol-lowing a self-control lapse could exhibit a stronger loss-preventing response because they are more in touch withthe benefits of avoiding indulgent consumption. In contrast,consumers who tend to further undermine themselves fol-lowing a lapse in self-control might display a weaker loss-preventing response because they fail to account for thenegative aspects of indulgent consumption.

In addition to the consideration of individual differencefactors that may influence indulgent consumption, researchexamining how goals and emotions shape other aspects ofbehavior remains largely uncharted. Most research on goalsand emotions has focused on how behaviors resulting ingoal attainment (nonattainment) can lead to the experienceof positive (negative) emotions (Carver and Scheier 1998;Soman and Cheema 2004). Only recently have researchersstarted to examine how emotions can be informative to goal-directed behaviors (Louro et al. 2007). In this regard, thefunctionalist perspective is an ideal theoretical frameworkfor investigating this issue (Keltner and Gross 1999). Forinstance, pride is an emotion that is typically associated withresisting temptations (Mukhopadhyay and Johar 2007).However, in the presence of a goal to be healthy, the ex-perience of pride can result in a sense of accomplishmentand redirect behavior toward indulgence (Wilcox et al.2011). Happiness is an emotion that typically facilitates be-havior geared toward self-improvement by increasing one’sreceptiveness to feedback (Raghunathan and Trope 2002).However, when a goal to manage one’s mood is active,happiness can redirect behavior away from self-improve-ment (Fishbach and Labroo 2007). Envy, an emotion elicitedfrom making unfavorable social comparisons with superiorothers, can increase impulsive behavior (Crusius and Muss-weiler 2012). Yet, when a goal to perform is salient, envycan increase pursuit of the goal by improving task perse-verance (van de Ven, Zeelenberg, and Pieters 2011). In eachof these instances, discrete emotions are shown to have cer-




tain action tendencies that can become modified when pairedwith an active goal.

Limitations and Future Directions

Our studies were not without limitations. In studies 1–3,the mediator was assessed after the measure of indulgentconsumption. The issue of when the mediator should be mea-sured is a challenge for many researchers. To illustrate thispoint, Iacobucci, Saldanha, and Deng (2007) examined twoarticles that reviewed the topic (Holland 1986; James, Mulaik,and Brett 1982) and found that 71.1% of the articles thatmeasured an independent variable, mediator, and dependentvariable took these measures simultaneously or out of order(e.g., mediator measured after the dependent measure). Theprimary concern for researchers, which was also the case forour research, is the potential for contamination effects re-sulting from the measurement of the mediator before the de-pendent variable. The mere measurement of the potential forharm of indulgence could have had an influence on indulgentconsumption and overridden our prior manipulations. Futureresearch could possibly circumvent this issue by measuringthe mediator at an implicit level, thereby capturing the processin its intended temporal sequence while maintaining the in-tegrity of the goal and emotion manipulations.

Our research could also be used as a starting point forfurther exploring other aspects of the interaction betweengoals and emotions. For instance, in our studies, it is possiblethat the hedonic goal priming manipulations encouraged anaffective experience. Related to this possibility, research hasshown that priming emotions can influence people’s behav-ior. For instance, Goldsmith, Cho, and Dhar (2012) showedthat the semantic priming of guilt led to increased pleasure.While it is uncertain whether this increased pleasure was aresult of the actual experience of guilt or the semantic as-sociation people have with guilt, it speaks to the possibilityof certain emotion primes being affectively relevant.

In addition, the current research focused primarily on he-donic goal pursuit, which, in the context of indulgent foodconsumption, always has a mix of both beneficial (i.e., it istasty) and harmful (i.e., it is unhealthy) consequences. Onepossible extension of this research would be to explorewhether the loss-preventing response of sadness would oc-cur for the pursuit of goals that only have beneficial (harm-ful) consequences depending on the success (failure) of thegoal. For instance, when one has an active impression man-agement goal in a social interaction, there are beneficial (i.e.,being well received by others) and harmful (i.e., being dis-liked by others) consequences of goal pursuit. It is hard toknow if sadness would still create a loss-preventing re-sponse, and result in avoidance of the social interaction.

Finally, while our research shows that the temporal se-quence of the hedonic goal and sadness interaction is criticalto determining when sadness initiates versus modifies goalpursuit, the examination of the temporal interplay betweenother goals and emotions is largely unexamined. Through theinvestigation of how temporal dynamics influence other goalsand emotions, our understanding of how these factors link

together would improve and lead to a more systematic frame-work for understanding goal and emotion interactions. Futureresearch could investigate these possibilities more closely.

Conclusion

To conclude, we replicate prior findings that hedonic en-vironmental cues can promote indulgence (Wadhwa, Shiv,and Nowlis 2008) and that sadness can create the need toregulate emotion through indulgent consumption (Garg andLerner 2013; Tice et al. 2001). However, we extend thesefindings by showing that hedonic environmental cues andsadness may not always encourage indulgent consumption.Rather, sadness can discourage the pursuit of a hedonic eat-ing goal. By decreasing indulgent consumption, one adap-tive consequence of sadness is to help consumers becomehealthier, which is consistent with functionalism and the ideathat all emotions, regardless of valence, are useful (Keltnerand Gross 1999). It would be naive to suggest that consum-ers should actively seek out ways to persistently feel sad.Sadness is only relevant to a subset of goals, and it has othernegative consequences. However, we do believe that neg-ative emotions are too often stigmatized as something thatshould not be experienced. Research has shown that bothpositive and negative emotions encourage adaptive re-sponses, wherein each emotion serves a functional purposein our day-to-day lives (Andrews and Thomson 2009; Toobyand Cosmides 2008). Thus, it is worth exploring conditionsunder which negative emotional experiences are beneficial.

DATA COLLECTION INFORMATION

The first author supervised the collection of data by re-search assistants using participants from the University ofMiami (study 1, ancillary study, and study 4), Universityof Florida (study 2), and the Amazon Mechanical Turk(study 3) between autumn 2011 and spring 2013. The firstauthor was primarily responsible for the data analysis withsupervision and input from the second and third authors.Data were discussed throughout the entire research agendaby all authors.

REFERENCESAndrade, Eduardo B., and Joel B. Cohen (2007), “Affect-Based

Evaluation and Regulation as Mediators of Behavior: TheRole of Affect in Risk-Taking, Helping and Eating Patterns,”in Do Emotions Help or Hurt Decision Making: A Hedge-foxian Perspective, ed. Kathleen Vohs, Roy Baumeister, andGeorge Loewenstein, New York: Russell Sage, 35–68.

Andrews, Paul W., and J. Anderson Thomson Jr. (2009), “TheBright Side of Being Blue: Depression as an Adaptation forAnalyzing Complex Problems,” Psychological Review, 116(July), 620–54.

Baumeister, Roy F. (2002), “Yielding to Temptation: Self-ControlFailure, Impulsive Purchasing, and Consumer Behavior,”Journal of Consumer Research, 28 (March), 670–76.

Bonanno, George A., Laura Goorin, and Karin G. Coifman (2008),“Sadness and Grief,” in Handbook of Emotions, Vol. 3, ed.




Michael Lewis, Jeannette M. Haviland-Jones, and Lisa Feld-man Barrett, New York: Guilford, 797–810.

Bowlby, John (1969), Attachment and Loss, Vol. 1, New York:Basic.

Campos, Joseph J., Donna L. Mumme, Rosanne Kermoian, andRosemary G. Campos (1994), “A Functionalist Perspectiveon the Nature of Emotion,” Monographs of the Society forResearch in Child Development, 59 (February), 284–303.

Carver, Charles S., and Eddie Harmon-Jones (2009), “Anger Is anApproach-Related Affect: Evidence and Implications,” Psy-chological Bulletin, 135 (March), 183–204.

Carver, Charles S., and Michael F. Scheier (1998), On the Self-Regulation of Behavior, New York: Cambridge UniversityPress.

Chandon, Pierre, and Brian Wansink (2012), “Does Food Mar-keting Need to Make Us Fat? A Review and Solutions,” Nu-trition Reviews, 70 (October), 571–93.

Cochran, Winona, and Abraham Tesser (1996), “The ‘What theHell’ Effect: Some Effects of Goal Proximity and Goal Fram-ing on Performance,” in Striving and Feeling: Interactionsamong Goals, Affect, and Self-Regulation, ed. Leonard L.Martin and Abraham Tesser, Hillsdale, NJ: Erlbaum, 99–120.

Cosmides, Leda, and John Tooby (2000), “Evolutionary Psychol-ogy and the Emotions,” in Handbook of Emotions, Vol. 2, ed.Michael Lewis and Jeannette M. Haviland, New York: Guil-ford, 91–115.

Crowe, Ellen, and E. Tory Higgins (1997), “Regulatory Focus andStrategic Inclinations: Promotion and Prevention in DecisionMaking,” Organizational Behavior and Human Decision Pro-cesses, 69 (February), 117–32.

Crusius, Jan, and Thomas Mussweiler (2012), “When People WantWhat Others Have: The Impulsive Side of Envious Desire,”Emotion, 12 (February), 142–53.

DeSteno, David, Piercarlo Valdesolo, and Monica Y. Bartlett(2006), “Jealousy and the Threatened Self: Getting to theHeart of the Green-Eyed Monster,” Journal of Personalityand Social Psychology, 91 (October), 626–41.

Fishbach, Ayelet, and Ravi Dhar (2005), “Goals as Excuses orGuides: The Liberating Effect of Perceived Goal Progress onChoice,” Journal of Consumer Research, 32 (December),370–77.

Fishbach, Ayelet, Ronald S. Friedman, and Arie W. Kruglanski(2003), “Leading Us Not Unto Temptation: Momentary Al-lurements Elicit Overriding Goal Activation,” Journal of Per-sonality and Social Psychology, 84 (February), 296–309.

Fishbach, Ayelet, and Aparna A. Labroo (2007), “Be Better or BeMerry: How Mood Affects Self-Control,” Journal of Per-sonality and Social Psychology, 93 (August), 158–73.

Fitzsimons, Gavan J. (2008), “Death to Dichotomizing,” Journalof Consumer Research, 35 (June), 5–8.

Frijda, Nico H. (2005), “Emotion Experience,” Cognition andEmotion, 19 (June), 473–97.

Garg, Nitika, and Jennifer S. Lerner (2013), “Sadness and Con-sumption,” Journal of Consumer Psychology, 23 (January),106–13.

Garg, Nitika, Brian Wansink, and J. Jeffrey Inman (2007), “TheInfluence of Incidental Affect on Consumers’ Food Intake,”Journal of Marketing, 71 (January), 194–206.

Gershoff, Andrew D., and Jonathan J. Koehler (2011), “SafetyFirst? The Role of Emotion in Safety Product Betrayal Aver-sion,” Journal of Consumer Research, 38 (June), 140–50.

Geyskens, Kelly, Siegfried Dewitte, Mario Pandelaere, and LukWarlop (2008), “Tempt Me Just a Little Bit More: The Effect

of Prior Food Temptation Actionability on Goal Activationand Consumption,” Journal of Consumer Research, 35 (De-cember), 600–610.

Goldsmith, Kelly, Eunice Kim Cho, and Ravi Dhar (2012), “WhenGuilt Begets Pleasure: The Positive Effect of a Negative Emo-tion,” Journal of Marketing Research, 49 (December), 872–81.

Gollwitzer, Peter M., and Paschal Sheeran (2006), “ImplementationIntentions and Goal Achievement: A Meta-Analysis of Effectsand Processes,” Advances in Experimental Social Psychology,38, 69–119.

Grant, Jon E., Marc N. Potenza, Aviv Weinstein, and David A.Gorelick (2010), “Introduction to Behavioral Addictions,”American Journal of Drug and Alcohol Abuse, 36 (August),233–41.

Griskevicius, Vladas, Noah J. Goldstein, Chad R. Mortensen, JillM. Sundie, Robert B. Cialdini, and Douglas T. Kenrick (2009),“Fear and Loving in Las Vegas: Evolution, Emotion, and Per-suasion,” Journal of Marketing Research, 46 (June), 385–95.

Griskevicius, Vladas, Michelle N. Shiota, and Samantha L. Neufeld(2010), “Influence of Different Positive Emotions on PersuasionProcessing: A Functional Evolutionary Approach,” Journal ofPersonality and Social Psychology, 10 (April), 190–206.

Griskevicius, Vladas, Michelle N. Shiota, and Stephen M. Nowlis(2010), “The Many Shades of Rose-Colored Glasses: An Evo-lutionary Approach to the Influence of Different Positive Emo-tions,” Journal of Consumer Research, 37 (August), 238–50.

Gross, James J., and Ross A. Thompson (2007), “Emotion Reg-ulation: Conceptual Foundations,” in Handbook of EmotionRegulation, ed. James J. Gross, New York: Guilford, 3–24.

Hayes, Andrew F. (2013), An Introduction to Mediation, Moder-ation, and Conditional Process Analysis: A Regression-BasedApproach, New York: Guilford.

Higgins, E. Tory, James Y. Shah, and Ronald S. Friedman (1997),“Emotional Responses to Goal Attainment: Strength of Reg-ulatory Focus as Moderator,” Journal of Personality and So-cial Psychology, 72 (March), 515–25.

Holland, Paul W. (1986), “Statistics and Causal Inference,” Journalof the American Statistical Association, 81 (December),945–60.

Hong, Jiewen, and Yacheng Sun (2012), “Warm It Up with Love:The Effect of Physical Coldness on Liking of Romance Mov-ies,” Journal of Consumer Research, 39 (August), 293–306.

Iacobucci, Dawn, Neela Saldanha, and Xiaoyan Deng (2007), “AMeditation on Mediation: Evidence That Structural EquationsModels Perform Better Than Regressions,” Journal of Con-sumer Psychology, 17 (April), 140–54.

Ilies, Remus, and Timothy A. Judge (2005), “Goal Regulationacross Time: The Effects of Feedback and Affect,” Journalof Applied Psychology, 90 (May), 453–67.

Izard, Carroll E. (1993), “Organizational and Motivational Func-tions of Discrete Emotions,” in Handbook of Emotions, ed.Michael Lewis and Jeannette M. Haviland, New York: Guil-ford, 631–41.

James, Lawrence R., Stanley A. Mulaik, and Jeanne M. Brett(1982), Causal Analysis: Assumptions, Models, and Data,Beverly Hills, CA: Sage.

Keller, Matthew C., and Randolph M. Nesse (2006), “The Evo-lutionary Significance of Depressive Symptoms: DifferentAdverse Situations Lead to Different Depressive SymptomPatterns,” Journal of Personality and Social Psychology, 91(August), 316–30.

Keltner, Dacher, and James J. Gross (1999), “Functional Accounts




of Emotions,” Cognition and Emotion, 13 (September),467–80.

Keltner, Dacher, Jonathan Haidt, and Michelle N. Shiota (2006),“Social Functionalism and the Evolution of Emotions,” inEvolution and Social Psychology, ed. Mark Schaller, DouglasT. Kenrick, and Jeffry A. Simpson, New York: PsychologyPress, 115–42.

Kruglanski, Arie W. (1996), “Goals as Knowledge Structures,” inThe Psychology of Action: Linking Cognition and Motivationto Behavior, ed. Peter M. Gollwitzer, New York: Guilford,599–618.

Kruglanski, Arie W., James Y. Shah, Ayelet Fishbach, Ronald S.Friedman, Woo Y. Chun, and David Sleeth-Keppler (2002),“A Theory of Goal Systems,” in Advances in ExperimentalSocial Psychology, ed. Mark P. Zanna, San Diego, CA: Ac-ademic Press, 331–78.

Kunzmann, Ute, and Daniel Gruhn (2005), “Age Differences inEmotional Reactivity: The Sample Case of Sadness,” Psy-chology and Aging, 20 (March), 47–59.

Labroo, Aparna A., and Vanessa M. Patrick (2009), “PsychologicalDistancing: Why Happiness Helps You See the Big Picture,”Journal of Consumer Research, 35 (February), 800–809.

Labroo, Aparna A., and Derek D. Rucker (2010), “The Orientation-Matching Hypothesis: An Emotion-Specificity Approach toAffect Regulation,” Journal of Marketing Research, 47 (Oc-tober), 955–66.

Laran, Juliano (2010a), “Choosing Your Future: Temporal Distanceand the Balance between Self-Control and Indulgence,” Jour-nal of Consumer Research, 36 (April), 1002–15.

——— (2010b), “Goal Management in Sequential Choices: Con-sumer Choices for Others Are More Indulgent than PersonalChoices,” Journal of Consumer Research, 37 (August),304–14.

Laran, Juliano, and Chris Janiszewski (2009), “Behavioral Con-sistency and Inconsistency in the Resolution of Goal Con-flict,” Journal of Consumer Research, 35 (April), 967–84.

——— (2011), “Work or Fun? How Task Construal and Com-pletion Influence Regulatory Behavior,” Journal of ConsumerResearch, 37 (April), 967–83.

Lazarus, Richard S. (1991), Emotion and Adaptation, New York:Oxford University Press.

Lench, Heather C., Sarah A. Flores, and Shane W. Bench (2011),“Discrete Emotions Predict Changes in Cognition, Judgment,Experience, Behavior, and Physiology: A Meta-Analysis ofExperimental Emotion Elicitations,” Psychological Bulletin,137 (September), 834–55.

Lerner, Jennifer S., and Dacher Keltner (2000), “Beyond Valence:Toward a Model of Emotion-Specific Influences on Judgmentand Choice,” Cognition and Emotion, 14 (July), 473–93.

Locke, Edwin A., and Gary P. Latham (1990), A Theory of GoalSetting and Task Performance, Englewood Cliffs, NJ: Pren-tice Hall.

Louro, Maria J., Rik Pieters, and Marcel Zeelenberg (2007), “Dy-namics of Multiple-Goal Pursuit,” Journal of Personality andSocial Psychology, 93 (August), 174–94.

Magen, Eran, and James J. Gross (2010), “Getting Our Act To-gether: Toward a General Model of Self-Control,” in SelfControl in Society, Mind, and Brain, ed. Ran R. Hassin, KevinN. Ochsner, and Yaacov Trope, New York: Oxford UniversityPress, 335–53.

McFerran, Brent, Darren W. Dahl, Gavan J. Fitzsimons, and AndreaC. Morales (2010), “I’ll Have What She’s Having: Effects of

Social Influence and Body Type on the Food Choices of Oth-ers,” Journal of Consumer Research, 36 (April), 915–29.

Miller, George A., Eugene Galanter, and Karl H. Pribram (1960),Plans and the Structure of Behavior, New York: Holt.

Mukhopadhyay, Anirban, and Gita V. Johar (2007), “Tempted orNot? The Effect of Recent Purchase History on Responses toAffective Advertising,” Journal of Consumer Research, 33(March), 445–53.

Mukhopadhyay, Anirban, Jaideep Sengupta, and Suresh Raman-athan (2008), “Recalling Past Temptations: An Information-Processing Perspective on the Dynamics of Self-Control,”Journal of Consumer Research, 35 (December), 586–99.

Nenkov, Gergana Y., J. Jeffrey Inman, and John Hulland (2008),“Considering the Future: The Conceptualization and Mea-surement of Elaboration on Potential Outcomes,” Journal ofConsumer Research, 35 (June), 126–41.

Nesse, Randolph M. (1990), “Evolutionary Explanations of Emo-tions,” Human Nature, 1 (March), 261–89.

——— (2000), “Is Depression an Adaptation?” Archives of Gen-eral Psychiatry, 57 (January), 14–20.

——— (2006), “Evolutionary Explanations for Mood and MoodDisorders,” in The American Psychiatric Publishing Textbookof Mood Disorders, ed. Daniel J. Stein, David J. Kupfer, andAlan F. Schatzberg, Washington, DC: American PsychiatricPublishing, 159–75.

Ohman, Arne, and Susan Mineka (2001), “Fears, Phobias, andPreparedness: Toward an Evolved Module of Fear and FearLearning,” Psychological Review, 108 (July), 483–522.

——— (2003), “The Malicious Serpent: Snakes as a PrototypicalStimulus for an Evolved Module of Fear,” Current Directionsin Psychological Science, 12 (February), 5–9.

Okada, Erica (2005), “Justification Effects on Consumer Choiceof Hedonic and Utilitarian Goods,” Journal of Marketing Re-search, 42 (February), 43–53.

Oppenheimer, Daniel M., Tom Meyvis, and Nicolas Davidenko(2009), “Instructional Manipulation Checks: Detecting Sa-tisficing to Increase Statistical Power,” Journal of Experi-mental Social Psychology, 45 (July), 867–72.

Ordabayeva, Nailya, and Pierre Chandon (2011), “Getting Aheadof the Joneses: When Equality Increases Conspicuous Con-sumption among Bottom-Tier Consumers,” Journal of Con-sumer Research, 38 (June), 27–41.

Polivy, Janet, and Peter C. Herman (1985), “Dieting as a Problemin Behavioral Medicine,” in Advances in Behavioral Medi-cine, ed. Edward S. Katkin and Stephen B. Manuck, NewYork: JAI, 1–37.

Power, Mick J. (1999), “Sadness and Its Disorders,” in Handbookof Cognition and Emotion, ed. Tim Dalgleish and Mick J.Power, West Sussex: Wiley, 497–519.

Raghunathan, Rajagopal, and Michel T. Pham (1999), “All Neg-ative Moods Are Not Equal: Motivational Influences of Anx-iety and Sadness on Decision Making,” Organizational Be-havior and Human Decision Processes, 79 (July), 56–77.

Raghunathan, Rajagopal, and Yaacov Trope (2002), “Walking theTightrope between Feeling Good and Being Accurate: Moodas a Resource in Processing Persuasive Messages,” Journalof Personality and Social Psychology, 83 (September),510–25.

Ramanathan, Suresh, and Geeta Menon (2006), “Time-VaryingEffects of Chronic Hedonic Goals on Impulsive Behavior,”Journal of Marketing Research, 43 (November), 628–41.

Ramanathan, Suresh, and Patti Williams (2007), “Immediate andDelayed Emotional Consequences of Indulgence: The Mod-




erating Influence of Personality Type on Mixed Emotions,”Journal of Consumer Research, 34 (August), 212–23.

Saarni, Carolyn, Joseph J. Campos, Linda Camras, and DavidWitherington (2006), “Emotional Development: Action, Com-munication, and Understanding,” in Handbook of Child Psy-chology, Vol. 3, Social, Emotional and Personality Devel-opment, ed. William Damon, Richard M. Lerner, and NancyEisenberg, New York: Wiley, 226–99.

Sharpe, Kathryn M., Richard Staelin, and Joel Huber (2008), “Us-ing Extremeness Aversion to Fight Obesity: Policy Implica-tions of Context Dependent Demand,” Journal of ConsumerResearch, 35 (October), 406–22.

Soman, Dilip, and Amar Cheema (2004), “When Goals Are Coun-terproductive: The Effects of Violation of a Behavioral Goalon Subsequent Performance,” Journal of Consumer Research,31 (June), 52–62.

Stearns, Carol Z. (1993), “Sadness,” in Handbook of Emotions,ed. Michael Lewis and Jeannette M. Haviland, New York:Guilford, 547–61.

Stein, Nancy L., and Linda J. Levine (1990), “Making Sense Outof Emotion: The Representation and Use of Goal-StructuredKnowledge,” in Psychological Approaches to Emotion, ed.Nancy L. Stein, Bennett Leventhal, and Tom Trabasso, Hills-dale, NJ: Erlbaum, 45–73.

Stein, Nancy L., Maria D. Liwag, and Elizabeth Wade (1996), “AGoal-Based Approach to Memory for Emotional Events: Im-plications for Theories of Understanding and Socialization,”in Emotion: Interdisciplinary Perspectives, ed. Robert D. Ka-vanaugh, Betty Zimmerberg, and Steven Fein, Hillsdale, NJ:Erlbaum, 91–118.

Thompson, Ross A., and Kristin Lagatutta (2006), “Feeling andUnderstanding: Early Emotional Development,” in The Black-well Handbook of Early Childhood Development, ed. Kath-leen McCartney and Deborah Phillips, Oxford: Blackwell,317–37.

Tice, Dianne M., and Ellen Bratslavsky (2000), “Giving in to FeelGood: The Place of Emotion Regulation in the Context ofGeneral Self-Control,” Psychological Inquiry, 11 (3), 149–59.

Tice, Dianne M., Ellen Bratslavsky, and Roy F. Baumeister (2001),“Emotional Distress Regulation takes Precedence over Im-pulse Control: If You Feel Bad, Do It!” Journal of Personalityand Social Psychology, 80 (January), 53–67.

Tiedens, Larissa Z., and Susan Linton (2001), “Judgment underEmotional Certainty and Uncertainty: The Effects SpecificEmotions on Information Processing,” Journal of Personalityand Social Psychology, 81 (December), 279–90.

Tooby, John, and Leda Cosmides (1990), “The Past Explains thePresent: Emotional Adaptations and the Structure of AncestralEnvironments,” Ethology and Sociobiology, 11 (4–5), 375–424.

——— (2008), “The Evolutionary Psychology of the Emotionsand Their Relationship to Internal Regulatory Variables,” inHandbook of Emotions, Vol. 3, ed. Michael Lewis, JeannetteM. Haviland-Jones, and Lisa F. Barrett, New York: Guilford,114–37.

Trope, Yaacov, Eric R. Igou, and Chritopher T. Burke (2006),“Mood as a Resource in Structuring Goal Pursuit,” in Affectin Social Thinking and Behavior, ed. Joseph P. Forgas, NewYork: Psychology Press, 217–34.

van de Ven, Niels, Marcel Zeelenberg, and Rik Pieters (2011),“Why Envy Outperforms Admiration,” Personality and SocialPsychology Bulletin, 37 (June), 784–95.

van Osselaer, Stijn M. J., and Chris Janiszewski (2012), “A Goal-Based Model of Product Evaluation and Choice,” Journal ofConsumer Research, 39 (August), 260–92.

Wadhwa, Monica, Baba Shiv, and Stephen M. Nowlis (2008), “ABite to Whet the Reward Appetite: The Influence of Samplingon Reward-Seeking Behaviors,” Journal of Marketing Re-search, 45 (August), 403–13.

Wang, Youfa, and May A. Beydoun (2007), “The Obesity Epidemicin the United States—Gender, Age, Socioeconomic, Racial/Ethnic, and Geographic Characteristics: A Systematic Reviewand Meta-Regression Analysis,” Epidemiologic Reviews, 29(July), 6–28.

Welling, Hans (2003), “An Evolutionary Function of the Depres-sive Reaction: The Cognitive Map Hypothesis,” New Ideasin Psychology, 21 (August), 147–56.

Wilcox, Keith, Thomas Kramer, and Sankar Sen (2011), “Indul-gence or Self-Control: A Dual Process Model of the Effectof Incidental Pride on Indulgent Choice,” Journal of Con-sumer Research, 38 (June), 151–63.

Williams, Lawrence E., John A. Bargh, Christopher C. Nocera,and Jeremy R. Gray (2009), “The Unconscious Regulation ofEmotion: Nonconscious Reappraisal Goals Modulate Emo-tional Reactivity,” Emotion, 9 (December), 847–54.

Winterich, Karen Page, and Kelly L. Haws (2011), “Helpful Hope-fulness: The Effect of Future Positive Emotions on Con-sumption,” Journal of Consumer Research, 38 (October),505–24.

Witherington, David C., and Jennifer A. Crichton (2007), “Frame-works for Understanding Emotions and Their Development:Functionalist and Dynamic Systems Approaches,” Emotion,7 (August), 628–37.

Zemack-Rugar, Yael, Canan Corus, and David Brinberg (2012),“The ‘Response-to-Failure’ Scale: Predicting Behavior Fol-lowing Initial Self-Control Failure,” Journal of Marketing Re-search, 49 (December), 996–1014.

Zhao, Xinshu, John G. Lynch Jr., and Qimei Chen (2010), “Re-considering Baron and Kenny: Myths and Truths about Me-diation Analysis,” Journal of Consumer Research, 37 (Au-gust), 197–206.



Documents

Journal of Consumer Research, Inc....marketing, and Juliano Laran ([email protected]) is associate professor of marketing, University of Miami, Coral Gables, FL 33124-6524. Chris Jan-iszewski