EFFECTS OF EMPLOYEES POSITIVE AFFECTIVE DISPLAYS ON ... · customer. ― Ronnie Clotfelter, owner and operator of Chick-fil-A, Sharpsburg, Georgia Focus on winning one customer at

r Academy of Management Journal2017, Vol. 60, No. 1, 109–129.https://doi.org/10.5465/amj.2014.0367

EFFECTS OF EMPLOYEES’ POSITIVE AFFECTIVE DISPLAYSON CUSTOMER LOYALTY INTENTIONS: AN EMOTIONS-AS-

SOCIAL-INFORMATION PERSPECTIVE

ZE WANGUniversity of Central Florida

SURENDRA N. SINGHUniversity of Kansas

YEXIN JESSICA LIUniversity of Kansas

SANJAY MISHRAUniversity of Kansas

MAUREEN AMBROSEUniversity of Central Florida

MONICA BIERNATUniversity of Kansas

Employees’ positive affective displays have been widely used as a strategic tool to en-hance service experience and strengthen customer relationships. Companies have pri-marily focused their employee training programs on two dimensions of display:intensity and authenticity. Yet there is limited research on when, how, and why thesetwo dimensions affect customer reactions. Drawing on the emotions as social informa-tion (EASI) framework (Van Kleef, 2009), we develop a conceptual model in whichdisplay intensity and display authenticity differentially influence customer loyalty bychanging customers’ affective reactions and cognitive appraisals. Further, we proposethat the relative impact of either dimension depends on customers’ motivation to un-derstand the environment deeply and accurately (i.e., their epistemic motivation). Wetested our model in one field study and one laboratory study. Results across these twostudies provide consistent support for the proposed model and advance our under-standing about how different dimensions of employees’ positive affective displays en-hance customer reactions. Thus, findings of this research contribute to knowledge on theinterpersonal effects of emotions in customer–employee interactions.

We call these traits ‘The Core Four’—eye contact,speaking enthusiastically, smiling, and engaging thecustomer.

― Ronnie Clotfelter, owner and operator ofChick-fil-A, Sharpsburg, Georgia

Focus on winning one customer at a time. Be honestand sincere.

― Robert Spector, The Nordstrom Way toCustomer Service Excellence

In many services, employees’ displayed emotionstoward customers are crucial determinants of cus-tomer satisfaction and loyalty. Policy statementsexemplified by the two quotes above demonstratethe importance of employees’ affective displays, aswell as the fact that they are trained to emphasizedifferent dimensions of display. Someemployees areobliged to follow scripted requirements, which de-mand that positive emotions be displayed frequentlyduring customer interactions, regardless of whether

We are grateful to associate editor Gerben Van der Vegtand three anonymous reviewers for their extremelyhelpfulguidance throughout the review process. We also thankour research assistants for their earnest efforts.

109

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s expresswritten permission. Users may print, download, or email articles for individual use only.

https://doi.org/10.5465/amj.2014.0367

the emotions are authentically felt (Pugh, 2001; Rafaeli& Sutton, 1990). Others, however, are trained to focuson authenticity rather than intensity, e.g., “be honestand sincere” (the Nordstrom Way) and apply “theprinciples of trust, honesty, respect, integrity, andcommitment” (Ritz-Carlton Gold Standards).

Policies that emphasize authenticity differ fromthose that highlight intensity in critical ways, in-cluding time and financial commitments on the partof the company and employees (Hochschild, 1983;Richard, 2006). When making important decisionsabout display policies, managers should have a goodgrasp of which dimension of affective display is bestfor their needs. For instance, what type of customerresponds more favorably to intense positive expres-sions than to authentic displays? How and why dothese dimensions of employees’ displayed emotions(intensity and authenticity) impact customer re-sponses in service deliveries?

The research reported here is designed to addressthese questions and provide a better understandingofwhen, how, andwhy the intensity and authenticityof employees’ displayed emotions differentially im-pact customer responses in servicedeliveries. To thisend, we draw on the emotions as social information(EASI) model (Van Kleef, 2009) to simultaneouslyexamine the influence of display authenticity anddisplay intensity in customer–employee interactions.We contend that: (a) these two dimensions influencecustomer responses through different mechanisms(emotions, cognitions); and (b) customers’ epistemicmotivation, or the motivation to develop and main-tain a rich and accurate understanding of one’s cur-rent situation (Kruglanski, 1989; Van Kleef, De Dreu,& Manstead, 2004), alters the effect of employees’display intensity and authenticity on customer re-sponses.Wedevelop and test amoderatedmediationmodel in which display intensity and display au-thenticity influence customer loyalty through theireffects on customers’ psychological responses, andthese indirect effects are moderated by epistemicmotivation.

This research extends the extant literature on therole of employees’ displayed emotion in service de-liveries. Despite the importance of employees’ pos-itive affective displays, prior research has not fullyaddressed what we need to know about customers’reactions to different dimensions of emotional dis-play (Groth, Hennig-Thurau, & Walsh, 2009). Re-searchers have recognized the importance of intensityand authenticity (Barger & Grandey, 2006; Grandey,Fisk, Mattila, Jansen, & Sideman, 2005; Pugh, 2001;Wang & Groth, 2014), but these facets have typically

been studied separately. This is problematic, for emo-tional displays vary along both dimensions and theunmeasured dimension may contribute or even drivethe obtained effects. The only study (Hennig-Thurau,Groth, Paul, & Gremler, 2006) that simultaneously ex-amined the effects of intensity and authenticity oncustomer reactions found that the effectiveness of em-ployees’ emotional labor depends on authenticity butnot intensity, raising the questions ofwhether intensityhas any unique effects on customer reactions andwhether these findings would hold in different con-texts. We argue that both intensity and authenticityplay a role in customer reactions, but their influencedepends on the customer’s motivation to process in-formation. In the current research, we consider in-tensity and authenticity simultaneously, explore themechanisms that underlie their influence on customerreactions, and illustrate moderating factors that altertheir effectiveness. We provide answers to previouslyunexplored questions such as: which customers re-spond favorably to intense versus authentic positivedisplays, and why? What guidelines should managersuse when determiningwhether to emphasize intensityor authenticity in employee training programs?

This research also contributes to the literature onthe EASI model. Previous studies using this modelhave so far examineddiscrete emotions suchas anger(Van Kleef, Anastasopoulou, & Nijstad, 2010), hap-piness (Van Kleef et al., 2004), disappointment andregret (Van Kleef, De Dreu, & Manstead, 2006) (seeVanKleef et al., 2010 for a recent review).Drawingonthe EASI model, this research compares two ex-pressive dimensions of the same emotion (happi-ness): intensity and authenticity. We contend thatthese dimensions offer nuanced information aboutthe expresser, and that receivers selectively use thisinformation to determine how to react. Moving be-yond valence-based effects as well as the differentialimpact of discrete emotions, we propose that differentdimensions of the same emotion provide functionalinformation that is relevant to social coordination andinteraction.

This study is also the first attempt to test the EASImodel in the context of employee–customer inter-actions. Previously, the EASImodel has largely beenlimited to experimental studies on conflict and ne-gotiation (DeDreu&VanKleef, 2004), leadership andteam performance (Van Kleef et al., 2010), and on-going personal relationships (Martini & Busseri,2012). In such relationships, the outcomes of theexpresser and perceiver are heavily intertwined. Byexamining the tenets of the EASI model in theemployee–customer domain, we test the applicability

110 FebruaryAcademy of Management Journal

of themodel to situations inwhich the expresser andreceiver do not necessarily share outcomes or havea lasting relationship with each other. Extending theEASI model to these situations is important becauseprevious research has found that qualitatively dif-ferent types of relationshipshave their owndecision-rules and interaction norms (Fiske, 1992).

Our findings also have important practical impli-cations. Organizations train, monitor, and managetheir frontline employees’ positive affective displaysvery carefully (Grandey, 2003). These practices con-sume a significant portion of spending on employeelearning and development (Richard, 2006), whichwas estimated at a staggering $164.2 billion in 2012 inthe United States alone (ASTD, 2013). But it is notalways clear what kind of affective display is mosteffective. Our research provides a framework forcompanies to understand how and why the effec-tiveness of employees’ positive affective displaysmay differ depending on the customers’ or re-cipients’ characteristics.

Below, we develop an EASI-based model for ex-amining the relationship between employees’ posi-tive affective displays and customer loyalty.Webeginwith an overview of research on displayed emotion,followed by a description of the EASI model and itsrelevance to research on emotional displays. We testour hypotheses in both a field study and a laboratoryexperiment. Theoretical and practical implications ofthe findings are discussed.

THEORETICAL DEVELOPMENT ANDHYPOTHESES

Dimensions of Positive Affective Displays:Intensity and Authenticity

Prior research on employees’ displayed emotionshas concentrated primarily on positive affectivedisplays or employees’ expression of socially de-sired emotions in customer interactions (e.g., Pugh,2001; Rafaeli & Sutton, 1990). This is also the focusof our investigation. We examine two dimensions ofdisplay: intensity and authenticity. The intensity ofemotional displays refers to mechanical aspects,such as the strength and frequency of smiling (Barger& Grandey, 2006), or the aggregated occurrence ofpositive display behaviors, such as establishing eyecontact, greeting the customer, and changing in-tonation (Tsai & Huang, 2002). The stronger or morefrequent these behaviors, the more intense the dis-play. Studies that highlight the importance of theintensity of employees’ display find that frequent

positive emotional expressions, such as broad smiles,greetings, and frequent eye contact typically elicitpositive customer outcomes (e.g., Barger & Grandey,2006; Pugh, 2001).

The authenticity of emotional displays, on theother hand, reflects the extent to which employees’expressions are consistent with their internal affec-tive states (Grandey, 2003). Authenticity can beviewed from two perspectives: the actor’s and theobserver’s. From the actor’s perspective, “emotionalauthenticity refers to a person’s ability to acknowl-edge, express, and be sincere about his or her feel-ings” (Huy, 1999: 339). In the current research, weare interested in the observer’s perception of au-thenticity. Since the observer is not privy to the ac-tor’s internal feelings, authenticity from his/herviewpoint refers to the extent to which the actor ap-pears to be expressing his/her true emotions versusattempting to “mislead another person into thinkingan emotion is felt when it is not” (Ekman, 1993: 390).Prior research has found that perceivers can recog-nize sincere from fake emotional expressions (Ekman& Friesen, 1982; Ekman, O’Sullivan, & Frank, 1999)and generally respond more positively to authentic(versus inauthentic) displays due to the importanceof honesty and trustworthiness in social interac-tions (Hennig-Thurau et al., 2006). In support of this,studies that stress the importance of authenticityhavefound that employees’ deep (versus surface) acting,which results in more authentic affective displays,generally leads to more positive customer reactions,such as satisfaction, rapport, and loyalty intentions(e.g., Grandey et al., 2005; Groth et al., 2009).

Although extant research has provided importantinsights on emotional displays, there are some gapsin our knowledge. For example, researchers havetypically examined either intensity or authenticity;they have rarely considered the two simultaneously.Because positive expressions usually vary in bothintensity and authenticity, there may be possibleconfounds inherent in focusing on a single dimen-sion (e.g., the unmeasured facet may contribute to oreven drive the observed effects). In addition, it isunclear whether contexts that reduce or enhance theeffectiveness of one dimension produce similar ef-fects on the other dimension. In other words, thequestion of what kind of customers and services re-spond best to authentic versus intense displays re-mains unanswered. The broader research literatureon emotions demonstrates that how observers pro-cess emotions and respond to them depends ona predictable set of individual and situational char-acteristics (Van Kleef, 2009). Building on this work,

2017 111Wang, Singh, Li, Mishra, Ambrose, and Biernat

we argue that customers’ motivation to thoroughlyunderstand their environment will influence howdisplay intensity and display authenticity affectcustomer responses.

The Emotions as Social Information Model: TwoPaths to Emotional Influence

Research on workplace emotions has increased inthe last decade.Much of this research has focused onthe intrapersonal effects of emotions. However,emotions are not simply reflections of one’s internalaffective states; they are also social informationalcues that communicate important messages to ob-servers, e.g., the displayer’s personality or interper-sonal intentions (Fridlund, 1992). The EASI modelprovides a comprehensive framework to understandhow emotional expressions are interpreted and usedby observers (Van Kleef, 2009). This model proposesthat emotional expressions can affect observers’ be-haviors via two distinct routes: (1) affective reactionsand (2) cognitive appraisals.

Affective reactions refer to the emotions elicitedin the observer. They generally arise when the dis-player’s emotional expression elicits correspondingemotions in the observer through primitive conta-gion (e.g., Hatfield, Cacioppo, & Rapson, 1992; Pugh,2001). Primitive emotional contagion or “catching”another’s emotions (Hatfield et al., 1992) occursspontaneously and requires relatively little cognitiveprocessing. Cognitive appraisals, on the other hand,are inferences that an individual draws about theexpresser’s true feelings and intentions. Unlike af-fective reactions, cognitive appraisals require ob-servers to make inferences about the meaning of theexpresser’s emotional displays,which guide attitudesand behaviors by providing contextually relevant in-formation about the expresser.

Critically, the EASI model stipulates that the ex-tent to which an observer is influenced by affectivereactions and cognitive appraisals is determined byhis/her epistemic motivation. Epistemic motivationrefers to howmotivated the observer is to understandthe environment deeply and accurately (Kruglanski,1989; Van Kleef et al., 2004). Observers with highepistemicmotivation aremore likely to pay attentionand use all available information before making adecision, and thus are more susceptible to the in-fluence of cognitive appraisals (Van Kleef, Homan,Beersma, Van Knippenerg, Van Knippenberg, &Damen, 2009). Observers with low epistemic moti-vation, on the other hand, have been shown to bemore likely to rely on their feelings when making

judgments (Albarracin & Wyer, 2001) and shouldthus be more influenced by their affective reactions.

Prior research using the EASI model has investi-gated the impact of discrete emotions (e.g., sadnessversus happiness) on these two routes of influence(Van Kleef et al., 2010). Research on emotional laborand affective displays, however, suggests that oneemotion can be comprised of several distinct di-mensions (Grandey et al., 2005; Hennig-Thurau et al.,2006). We thus extend the existing EASI literatureby investigating how two dimensions of the sameemotion—intensity and authenticity—differentiallyaffect affective reactions and cognitive appraisals toinfluence customer attitudes and behaviors. We sug-gest intensity influences customer reactions throughthe affective route, whereas authenticity influencescustomer reactions through the cognitive route.Asweexplain inmore detail later, this assertion is based onreasoning and evidence that intense displays ofpositive affect influence customer reactions mainlythrough emotional contagion, while authentic dis-plays of positive affect influence customer reactionsthrough processing of the emotion’s meaning. Ad-ditionally, in line with the EASI model, we proposethat the effect of intensity and authenticity on cus-tomers’ responses is determined by the customer’sepistemic motivation. The conceptual framework ofthe current research is presented in Figure 1.

Display Intensity and Customers’ AffectiveReactions

The first route by which emotional displays in-fluence customer reactions is the affective route.Prior work on display intensity suggests employees’intense positive displays, such as broad smiles,greeting, intonation, and frequent eye contact, in-fluence customer outcomes by changing or enhanc-ing customer emotions (Pugh, 2001). To explain thiseffect, several streams of intensity research have uti-lizedemotional contagion theory (Hatfield et al., 1992).Emotional contagion theory contends that humansinnately and unconsciously mimic the nonverbal be-haviors of others and, in doing so, converge emotion-ally with them through physiological feedback (Laird&Bresler, 1992). Hatfield, Rapson, and Le (2009: 26)provide a description of how emotional contagionoperates:

In theory, the process of emotional contagion consistsof three stages: Mimicry, Feedback, and Contagion.People tend: (a) to automatically mimic the facial ex-pressions, vocal expressions, postures, and instrumental


behaviors of those around them, and thereby (b) to feela pale reflection of others’ emotions as a consequence ofsuch feedback. (c) The result is that people tend to catchone another’s emotions.

Presumably, when people automatically mimic theircompanions’ fleeting facial, vocal, and postural ex-pressions of emotion, they often come to feel a palereflection of their companions’ actual emotions. Byattending to this stream of tiny moment-to-momentreactions, people are able to “feel themselves into” theemotional lives of others.

We posit that observers are more likely to pick upon and thus successfully mimic behaviors that aremore intense or frequent, a prediction that is sup-ported by previous research (Barger & Grandey,2006; Pugh, 2001; Tsai & Huang, 2002). Barger andGrandey (2006), for instance, find that employees’smile strength (a common measure of intensity)predicts customers’ smile strength, which enhancescustomers’ positive mood. Importantly, even fakedisplays of positive emotions in the expresser mayengender positive emotions in the observer. Ina field test involving service interactions between220 customers and 191 bank tellers, Pugh (2001)found that intense positive affective displays inemployees increased customers’ positive affect, butthat these emotional displays were not significantlyrelated to employees’ internal feelings, leading tothe suggestion that “A professional acts as theymust, not as they feel” (Pugh, 2001: 1018). Based onthe aforementioned research, we propose that dis-play intensity affects observers primarily through

spontaneous contagion processes that result in emo-tional convergence with the displayer:

Hypothesis 1. The intensity of employees’ posi-tive affective displays will produce positiveaffective reactions in customers.

Display Authenticity and Customers’ CognitiveAppraisals

The second route identified in the EASI model isthrough cognitive appraisals. Observersmake context-specific cognitive inferences from an expresser’s af-fectivedisplay that provide relevant information aboutthe situation or the expresser, such as whether he/shewants to affiliate (displays a smile) or be left alone(displays a frown) (Van Kleef, 2009). Consistent withthe EASI model, research suggests that observers de-rivemeaning fromtheauthenticityofaffectivedisplays(Van Kleef, 2009; Wang & Groth, 2014). Display au-thenticity may be an especially important cue fromwhich tomake judgments about the expresser becauseit provides information about his/her personality andintentions (Ekman, Friesen, & O’Sullivan, 1988). Ingeneral, an authentic smile conveys that the expresseris trustworthy, kind, sociable, and cooperative, whilean inauthentic smile signals that the expresser is put-ting on a “mask” because of social norms or as an at-tempt to deceive the observer (Krumhuber, Manstead,Cosker, Marshall, Rosin, & Kappas, 2007).

In the context of service delivery, the relevantcognitive appraisals customers make are about

FIGURE 1A Conceptual Model of Customers’ Responses to Employees’ Positive Affective Displays in Service Encounters

EAD Intensity

EAD Authenticity

Affective Reactions

Appraisals aboutEmployee Service

Performance

Customer LoyaltyIntentions

Customers’ EpistemicMotivation

H3

H4

H5

H1

H2

H6Employees’ Positive

Affective Displays (EAD)Customers’ Psychological

ResponsesCustomer-Related

Outcomes

Customer Individual-LevelDifferences


employee service performance (Gabriel, Acosta, &Grandey, 2013), or the degree to which employeessuccessfully perform “behaviors that are relevant toorganizational goals and that are under the control ofindividual employees” (Liao & Chuang, 2007: 42).Employee service performance encompasses factorssuch as the extent to which the employee is trust-worthy, cares about customers’ needs, and is willingand able to provide service (Brady & Cronin, 2001).

We propose that the authenticity of employees’positive displays influences customers’ appraisalsof employee service performance in several ways.First, authenticity is related to trustworthiness, amain component of service performance. Second,employees’ authentically positive displays signaltheir motivation to move beyond the “required” roleand that there is genuine intention to be helpful tocustomers (Grandey et al., 2005). This enhances cus-tomers’ perceptions of the employee as customer-focused, and results in more favorable evaluations(Chi, Grandey, Diamond, &Krimmel, 2011; Groth et al.,2009). Inauthentic emotional displays, on the otherhand, can be interpreted as a lack of interest in de-veloping a close relationship with the customer or evi-dence that the employee is not trustworthy (Cote, 2005;Grandey et al., 2005), leading customers to infer that theemployee is unwilling or unable to provide a qualityinteraction. In sum, we hypothesize that authentic(versus inauthentic) affective displays will lead to pos-itive appraisals about employee service performance.

Hypothesis 2. The authenticity of employees’positive affective displayswill lead customers tomake positive appraisals of employee serviceperformance.

Effect of Affective Reactions and CognitiveAppraisals on Customer Loyalty Intentions

While eliciting positive mood and appraisals incustomers is desirable, this is often not the end goalfor managers and organizations. The hope is thatpositive mood and appraisals lead to downstreamconsequences such as customer loyalty. Customerloyalty is a behavioral construct that is influenced bycustomers’ feelings of attachment and affection to-ward an organization (Payne & Webber, 2006). Loy-alty intentions are demonstrated through customers’intentions to continue business relations or tospread positive word-of-mouth (Swan &Oliver, 1989).Customer loyalty is a strongpredictor of organizationalprofit and growth (Reichheld & Teal, 2001; Rust &Zahorik, 1993).

Customers’ positive affective reactions can en-hance loyalty intentions in several ways. When in-dividuals are in a positive mood, they tend to look atthe world through rose-colored glasses and makeglobal positive evaluations (Sinclair, 1988). In addi-tion, people generally strive for positive experiencesand seek encounters that have induced positiveemotions in the past (Skinner, 1953). Indeed, cus-tomers who experiencemore positive feelings duringa service transaction are more likely to stay longerin a store (Tsai & Huang, 2002), spread positiveword-of-mouth (Swan &Oliver, 1989), and expresshigher intentions to revisit the store (Hennig-Thurauet al., 2006). These findings lead to the followinghypothesis:

Hypothesis 3. Positive affective reactions posi-tively relate to customer loyalty intentions.

Positive appraisals of employee service perfor-mance have also been shown to affect customerloyalty intentions (Salanova, Agut, & Peiro, 2005).When frontline employees extend their respect andcourtesy to customers, theyhelp the firm formabondwith customers, which leads to more favorableevaluations of the service experience (Keh, Ren, Hill,& Li, 2013). Furthermore, customers’ perception ofemployees’ congenial attitudes and behaviors drivecustomers’ intentions to stay loyal to a firm (Liao &Chuang, 2007; Payne &Webber, 2006; Salanova et al.,2005). A survey of 2,170,004 patients from 1,506 hos-pitals, for example, found that employee serviceperformance (as measured by responsiveness tocustomers’ concerns and complaints and sensitivityto the inconvenience of medical service) was highlycorrelated with patients/customers’ likelihood ofrecommending the hospital to others (Berry, Wall, &Carbone, 2006). This leads us to predict:

Hypothesis 4. Positive appraisals of employeeservice performance positively relate to cus-tomer loyalty intentions.

The Moderating Role of Customers’ EpistemicMotivation

Thus far, we have argued that the intensity andauthenticity of employees’ affective displays influ-ence customers’ affective reactions and cognitiveappraisals, respectively, which in turn affect loyaltyintentions. Although both affective reactions andcognitive appraisals can lead to greater loyalty in-tentions, it is theoretically and managerially impor-tant to determine when each process is likely to


produce positive outcomes. According to the EASImodel, the impact of these two processes is de-termined by the observer’s epistemic motivation(Kruglanski, 1989; Van Kleef et al., 2004). Epistemicmotivation is an individual’s motivation to developa rich and accurate understanding of one’s currentsituation. It is strongly tied to information processingstyle. Heightened epistemic motivation has beenlinked to focused search of diagnostic information(Kruglanski & Webster, 1996), decreased use ofheuristics and stereotypes (Fiske & Neuberg, 1990),and increased systematic processing (Mayseless &Kruglanski, 1987).

Prior research has determined that heightenedepistemic motivation accentuates the effect of cog-nitive appraisals and attenuates the influence of af-fective reactions. In a study of leader emotions onfollower performance, Van Kleef et al. (2009) foundthat followers who had high epistemic motivationwere heavily influenced by their cognitive infer-ences of the leader’s emotions, while those who hadlow epistemic motivation were more influenced bytheir affective reactions to the leader’s emotions. VanKleef et al. (2009) suggest this is because low epi-stemic motivation leads individuals to base theirbehavior on their emotions, whereas high epistemicmotivation leads individuals to “rely less on theiraffective state but instead use other sources of in-formation to guide their behavior, such as in-formation provided by others’ emotional displays”(p. 565). Similarly, other research has shown thatwhile emotions are easily accessible heuristic guides,individuals with high epistemic motivation are morelikely to discount their affective reactions as inaccu-rate or irrelevant, leading them to rely more oncognitive appraisals (Albarracin & Kumkale, 2003;Greifeneder & Bless, 2007).

The EASImodel thus suggests that the influence ofemotional display intensity on customer loyalty in-tentions via affective reactionswill be accentuated ascustomers’ epistemicmotivation decreases, whereasthe impact of emotional display authenticity oncustomer loyalty intentions via cognitive appraisalswill be enhancedas customers’ epistemicmotivationincreases. Formally, we hypothesize:

Hypothesis 5. Customers’ epistemic motivationmoderates the relationship between the in-tensity of employees’ positive affective displaysand customer loyalty intentions through cus-tomers’ affective reactions, such that the medi-ated relationship is weakened when customershave higher levels of epistemic motivation.

Hypothesis 6. Customers’ epistemic motivationmoderates the relationship between the au-thenticity of employees’ positive affective dis-plays and customer loyalty intentions throughcustomers’ appraisals about employee serviceperformance, such that the mediated relation-ship is strengthened when customers have higherlevels of epistemic motivation.

We test the proposed hypotheses in two studies. InStudy 1, a field study, we examine the effects of em-ployees’ intensityandauthenticityoncustomer loyaltyintentions through affective reactions and appraisalsabout employee service performance, and investigatethe moderating role of self-reported motivation toprocess the employee’s behavior. Study 2 providesa replicationofStudy1,usingabetter controlledvideo-simulation methodology and a more established indi-vidual difference measure of epistemic motivation.

STUDY 1: FIELD STUDY WITH IN-SITUOBSERVATIONS AND SURVEYS

In this field study,weexamine theproposition thattwo dimensions of employees’ displayed emotions(intensity versus authenticity) differentially impactcustomers’ responses, and that customers’ epistemicmotivation moderates these effects. The current re-search focuses on customers’ reactions to employees’affective displays. Therefore, Study 1 investigates themoderating role of one specific type of epistemicmotivation—customers’ motivation to develop anaccurate understanding of employees’ displayed emo-tions (Cote & Hideg, 2011), which should determinethe relative influence of affective reactions andcognitive appraisals on loyalty intentions. Accord-ing to the EASI model (Van Kleef, 2009), customersare more likely to base their behavioral intentionson their emotions when they have low (versus high)epistemic motivation. Therefore, the effects of af-fective display intensity on customer outcomes viaaffective reactions should be stronger when episte-mic motivation is low. On the other hand, the effectof authenticity via cognitive appraisals should in-crease as epistemic motivation increases.

Procedure

In this study, we observed customer–employeedyadic interactions in 36 stores, which representa variety of services, such as fashion retailers, diningservices, and grocery stores. The first author trainedfive research assistants for two weeks by using


relevant literature on emotion detection (e.g., Ekman& Friesen, 1982) to familiarize the assistants with theprocess and to minimize error variance caused byobserver heterogeneity.

The field study was conducted during regularbusiness hours by teams of three research assistantsover the course of three months. Two research assis-tants independently rated the intensity and authen-ticity of employees’ emotional displays (inter-coderreliability 5 0.78 and 0.84, respectively) after un-obtrusively observing customer–employee dyadic in-teractions. The third research assistant interviewed thecustomers after they exited the store. We used RArather than customer ratings of authenticity for severalreasons. Following Pugh (2001), unobtrusive observa-tionsweremade to reduce thepossibility of sensitizingthe employees under direct observation, and to mini-mize disturbance to other customers and employees.We use independent coders’ ratings on intensity andauthenticity rather than the customers’ retrospectiveevaluations of employees’ emotional expressions so asto capture emotional expressions at the time of thedisplay rather than when the interaction was over. Inaddition, using research assistants rather than cus-tomers to assess authenticity and intensity reduceserrors from common method bias. Finally, previousresearch indicates people are reasonably accurate atdetecting real from fake expressions (Ekman&Friesen,1982; Ekman et al., 1988, 1999), which suggests dif-ferent observers should have similar perceptions ofauthenticity, and that RA ratings are an acceptableproxy for customer ratings in this context.1

Three-hundred and twenty-six transactions with122 service employees were observed. A customerresponse rate of 63% resulted in 208 sets of codingsheets and customer surveys. Seven pairs had exten-sivemissingdata.Theeffective samplewas, therefore,201 customer–employee dyads (customer sample:Mage5 30.4, 18 to 79; 47% female; employee sample:Mage 5 28.2, 18 to 66; 56% female). The customers’demographics were self-reported, and the researchassistants estimated the employees’ information.

Measures

Employees’ emotional displays of intensity andauthenticity. In line with previous research, we con-ceptualized intensity in a reflective manner and

measured the intensity of emotional displays withtwo items that assess the strength and frequency ofemployees’ smiling (Barger & Grandey, 2006; Hennig-Thurau et al., 2006). See Table 1 for themeasures. Tworesearch assistants independently rated smile strengthona scale of 0–100%in20%increments,with anchorsat 0% (no smile) and 100% (maximal smile); the fre-quency of smiles was measured by smiles per minute:15Not smilingat all, 25Less than1smileperminute,35 1–2 smiles perminute, 45 3–4 smiles perminute,55 5 or more smiles per minute, and 65 continuoussmiling. We used four items or both coders’ ratings ofsmile strength and the frequency as indicators becauseusingcoders’average ratingswouldhavegivendisplayintensity only two indicators, which can lead to anunder-identifiedmodel. Cronbach’s a for the intensityscale was 0.88.

Employees’ authenticity has been measured bycoders’ or customers’ perceived authenticity (Grandeyet al., 2005; Hennig-Thurau et al., 2006) and by anassessment of employees’ alternative acting strate-gies (Chiet al., 2011).Weuse the formermethod in thisfield study and adopt the latter approach in the labo-ratory experiment. In the field test, we assessedauthenticity using coders’ ratings on a three-item,seven-point scale (1 5 “strongly disagree” to 7 5“strongly agree”) adapted from Grandey et al. (2005):“The employee faked how he felt in this interaction”;“This service provider seemed to put on an act in thisinteraction”; “the service provider actually experi-enced theemotions (s)hehad to showto thecustomer.”The authenticity scores were the average of the twocoders’ ratings. Cronbach’s a for this scale was 0.95.

Customers’ epistemic motivation. In this study,we measured customers’ epistemic motivation byassessing their motivation to process employees’emotional displays accurately. Following Cote andHideg (2011), we created a three-item scale to mea-sure customers’ motivation to process employees’affective displays: “I paid attention to the serviceprovider’s emotional display”; “I was interested inthe employee’s emotional expressions”; “I paid closeattention to employees’ gestures when talking tothem” (15 “stronglydisagree” to55 “stronglyagree”).These items were framed in the context of customer–employee interactions in service transactions.Cronbach’s a for this scale was 0.83.

Customers’affective reactions to serviceexperience.As in previous research, affective reactions weremeasured using items that assessed emotional re-actions (Van Kleef et al., 2009). The items, adaptedfrom Tsai and Huang (2002) and Hennig-Thurau et al.(2006), were on a five-point Likert scale (1 5 strongly

1 A supplementary study revealed that RAs’ and cus-tomers’ ratings of employee authenticity are significantlycorrelated, and that usingRAratings is amore conservativetest of our hypotheses.


disagree, 5 5 strongly agree) and include: “I feelcontented after the service transaction”; “I feel pleasedafter the service transaction”; “I enjoyed interactingwith this employee.” Cronbach’s a for this scalewas 0.82.

Customers’ cognitive appraisals of employeeserviceperformance.Customers’ cognitive appraisalsof the employee’s service performanceweremeasuredby customers’ assessment of the employee’s helpful-ness (one item adapted from the service performancescale, Liao & Chuang, 2007) and their interactive ex-periencewith theserviceemployee (two itemsadaptedfrom the interaction quality scale, Brady & Cronin,2001). These two facets of employee service perfor-mance were chosen because they are closely linkedto employees’ affective delivery, whereas other di-mensions of customers’ cognitive appraisals, such asreliability, are more dependent on employees’ task-based service delivery (Brady & Cronin, 2001; Tsai &Huang, 2002). The three-item, five-point Likert scale(1 5 strongly disagree, 55 strongly agree) includesthe following items: “I’d say the quality of my in-teraction with this employee is high”; “This serviceemployee was willing and able to provide the ser-vice I need”; “I’d say the quality of my interaction

with this employee is excellent.” Cronbach’s awas 0.90.

Customer loyalty intentions. We adapted thecustomer loyalty intention scale from Groth et al.(2009). The items in this five-point Likert scale (1 5strongly disagree, 5 5 strongly agree) are: “The nexttime I need a similar type of service, I will choose thisservice provider again”; “I will say positive thingsabout this service provider to others”; “I will recom-mend this service provider to others.” Cronbach’s afor this scale was 0.94.

Preliminary Analyses

Given the nested structure of our data (201 cus-tomer surveys from interactionswith 122 employeesat 36 stores), we evaluated the factor structure of themeasures through a multilevel confirmatory factoranalysis (MCFA) of the latent variables in ourmodel:display intensity, display authenticity, affective re-actions, cognitive appraisals, epistemic motivation,and loyalty intentions. The standardized loadings inthe measurement model are high and load on theirrespective factors (see Table 1). The hypothesizedsix-factor model, where individual scale items

TABLE 1Results of the Confirmatory Factor Analysis in the Field Study: Study 1

Items Standardized Loadings SE

Display Intensity (Adapted from Barger & Grandey, 2006; Hennig-Thurau et al., 2006)1. The intensity of smiles (coder1/coder2) 0.85/0.86 0.04/0.042. The frequency of smiling (coder1/coder2) 0.81/0.73 0.05/0.08Display Authenticity (Adapted from Grandey, 2003; Grandey et al., 2005)1. This service provider seemed to be faking how he/she felt in this interaction. 0.98 0.012. This service provider seemed to put on act in this interaction. 0.98 0.013. The service provider actually experiences the emotions he had to display to the customer. 0.99 0.00Affective Reactions (Adapted from Hennig-Thurau et al., 2006; Tsai & Huang, 2002)1. I felt contented after the service transaction. 0.82 0.022. I felt pleased after the service transaction. 0.88 0.05

0.0213. I enjoyed interacting with this employee. 0.63 0.05Cognitive Appraisals of Employee Service Performance (Adapted from Brady & Cronin,

2001, Liao & Chuang, 2007)1. I’d say the quality of my interaction with this employee is excellent. 0.87 0.042. This service employee was willing and able to provide the service I need. 0.79 0.043. I’d say that the quality of my interaction with this employee is high. 0.84 0.05Customers’ Epistemic Motivation (Based on Cote & Hideg, 2011)1. I was interested in the employee’s emotional expressions. 0.68 0.062. I paid attention to the employee’s emotional display. 0.79 0.073. I paid close attention to the employees’ gestures when talking to them. 0.71 0.07Customer Loyalty Intentions (Adapted from Hennig-Thurau et al., 2006)1. The next time I need similar type of services, I will choose this service provider again. 0.85 0.042. I will say positive things about this service provider to others. 0.94 0.013. I will recommend this service provider to others. 0.96 0.02


loaded on separate first-order latent factors, dis-played good fit (x2 (137)5 185.57; RMSEA5 0.04;SRMR5 0.04; CFI 5 0.98; Hu & Bentler, 1999).

To rule out the possibility that common methodbias accounts for these results, we tested an alterna-tive two-factor model that included one latent factorcomposed of those items rated by research assistants(the intensity and authenticity of employees’ af-fective displays) and a second factor based on thevariables that were rated by customers (affectivereactions, cognitive appraisals, epistemic motiva-tion, and loyalty intentions). The fit of this modelwas poor x2 (151)5 792.41; RMSEA5 0.15; SRMR50.14; CFI 5 0.75) and was significantly worse thanthatof the firstmodel (Δx2 (14)5606.84,p,0.01).Wealso conducted two MCFAs to directly assess thedistinctiveness of display intensity and display au-thenticity (x2 (13) 5 22.73; RMSEA 5 0.06; SRMR 50.05; CFI5 0.98) and between affective reactions andcognitive appraisals (x2 (8) 5 12.24; RMSEA 5 0.05;SRMR 5 0.04; CFI 5 0.99). In both cases, the two-factor model provides a good fit to the data, therebyincreasing our confidence in the distinctiveness ofthese constructs.

Hypotheses Testing

The means, standard deviations, reliability esti-mates, and correlations of the variables are shown inTable 2.Wepredicted that the intensity of employees’positive affective displays indirectly impacts cus-tomer loyalty intentions by evoking positive affectiveresponses in customers (as predicted byHypothesis 1and Hypothesis 3), whereas the authenticity of em-ployees’ positive affective displays enhances cus-tomer loyalty intentions by changing customers’appraisals of employee service performance (as pre-dictedbyHypothesis2andHypothesis4).Further,wepredicted in Hypothesis 5 and Hypothesis 6 that theindirect effects of intensity and authenticity on cus-tomer loyalty intentions are conditional on customers’epistemic motivation.

We tested these hypotheses with multilevel struc-tural equation modeling (MSEM) using Mplus 7(Muthen & Muthen, 2012). MSEM enabled us toisolate the customer-level effects of interest and avoidinaccurate standard errors or biased estimationdue to non-independent observations (Muthen &Asparouhov, 2011; Preacher, Zyphur, & Zhang,2010). We used all scale items as indicators for theintensity and authenticity of employees’ affectivedisplays and customers’ future loyalty intentions.Following recommendations for testing structural

equation models including latent variable inter-actions, we modeled the latent factors for affectivereactions, cognitive appraisals, and epistemic moti-vation using standardized scale scores as single in-dicators (Cortina, Chen, & Dunlap, 2001). Employeeswithin stores were the two clustering variables. Weused the recommended analysis approach (Complexand Twolevel in Mplus), which computes standarderrors and x2 statistics taking into account non-independence of observations due to both clusterlevels (Muthen & Muthen, 2012). We employed ro-bust maximum likelihood estimation (MLR), whichprovides estimates of standard errors and x2 test ofmodel fit that controls for the nested nature of data.The resulting model provided good fit to the datax2 (83) 5 151.06; RMSEA 5 0.06; SRMR 5 0.10;CFI 5 0.96).

The results support our hypotheses. For displayintensity, the direct effects indicate that employees’emotional display intensity positively relates tocustomers’ affective reactions2 (b 5 0.13, p , 0.05),supporting Hypothesis 1. Additionally, in line withHypothesis 3, customers’ affective reactions posi-tively predicted their loyalty intentions (b 5 0.39,p , 0.01). For display authenticity, the effect of au-thenticity on customers’ appraisals of employeeservice performance was significant (b 5 0.14, p ,0.05), in support of Hypothesis 2. The effect of ap-praisals on customer loyalty intentions was alsosignificant (b 5 0.51, p , 0.01), supporting Hypoth-esis 4. However, these results should be evaluated inlight of the conditional indirect effects. Hypotheses 5and 6 predict that the indirect effects of intensity andauthenticity on customer loyalty intentions dependupon the customer’s epistemic motivation. The re-sults support these predictions. The interaction ofaffective reactions and epistemic motivation on cus-tomer loyalty intentions was negative and significant(b 5 20.23, p , 0.05), whereas the interaction effectof cognitive appraisal and epistemic motivation waspositive and significant (b5 0.28, p, 0.01).

We tested the significance of the indirect effectsusing bias-corrected bootstrapping confidence inter-vals (CI) (10,000 re-samples), which does not requireany assumptions about the sampling distribution un-derlying the moderated mediation model (Preacher &Hayes, 2008). As predicted, the indirect path of

2 In supplementary analyses,we also tested the effects ofthe intensity X authenticity interaction on customers’ af-fective reactions and cognitive appraisals. Neither of theseinteractions was significant (b 5 0.02 and 0.01, respec-tively; p . 0.10).


intensity on customer loyalty intentions throughcustomers’ affective reactions is significant forcustomers with low epistemic motivation (effect 50.14, SE 5 0.08, 90% CI 5 0.01, 0.24), but not forcustomers with high epistemic motivation (effect 50.04, SE 5 0.03, 90% CI 5 0.00, 0.10). The indirectpath of authenticity on customer loyalty intentionsthrough appraisals of employee service performancewas significant for customers with high epistemicmotivation (effect 5 0.21, SE 5 0.09, 90% CI 5 0.07,0.37), but not for customers with low epistemic moti-vation (effect5 0.06,SE5 0.04, 90%CI520.01, 0.15).

We also probed the significant interactions withthe simple effects approach, andplotted the intensityX motivation and authenticity X motivation inter-actions using 1/2 1 SD of the predictor and moder-ator variables (Preacher, Curran, & Bauer, 2006). Asshown in Figure 2, the effects of intensity and au-thenticity on customer loyalty intentionswere in thepredicted directions. That is, the effect of intensityon loyalty intentions was greater for customers withlow (versus high) epistemic motivation, while theeffect of authenticity was greater for customers withhigh (versus low) epistemic motivation.

Discussion

Study 1 extends prior research on the impact ofemployees’ positive affective displays on customerreactions by distinguishing between the effects ofintensity and authenticity and how they influenceaffective reactions and cognitive appraisals. We findsupport for the hypothesis that intensity and authen-ticity affect loyalty intentions via different pathways,i.e., affective reactions and cognitive appraisals, re-spectively. Additionally, we find that the impact ofintensity and authenticity on customer outcomeslargely depends on customers’ epistemic motivation.

When customers have low motivation to develop anaccurate understanding of employees’ positive affec-tive displays, display intensity enhances customerloyalty intentions. When customers have high moti-vation to interpret employees’ displayed emotions,display authenticity enhances intentions to stay loyalto the service provider. These results shed light on thedifferent ways in which affective displays influencecustomer reactions, and suggest customers respond indistinct, yet predictable, ways to display intensity andauthenticity depending on their epistemicmotivation.

This field study is an important first step in creat-ing a greater understanding of the relationship be-tween employee emotional displays and customerreactions. However, this approach is not without itslimitations. First, although in-situ interviews fromthe field study contribute to external validity, we donot have control of certain situational influences,such as the dynamics of employee–customer inter-actions, distractions by other employees or customers,and ambient conditions (e.g., noise, odor). Second,surveys do not enable us to establish causal relation-shipsamongconstructs.Toaddress theseconcerns,weconducted a well-controlled, video-simulated labora-tory experiment to supplement the field study andprovide convergent evidence for our hypotheses.

STUDY 2: EXPERIMENT WITH MEASUREDEPISTEMIC MOTIVATION

In Study 2, we attempt to replicate our findings fromStudy 1 by showing all participants the same recordedservice interaction and measuring epistemic motiva-tion with a well-established measure—the 11-itemneed for structure scale (Neuberg & Newsom, 1993).The personal need for structural scale is a reliablemeasure of individual differences in information pro-cessing motivation (Van Kleef et al., 2010). This

TABLE 2Descriptive Statistics, Correlations, and Reliabilities for All Measures in Study 1a

Variables Mean SD 1 2 3 4 5 6

1. Display Intensity 2.66 1.25 (0.88)2. Display Authenticity 3.08 1.98 0.31** (0.95)3. Affective reactions 3.81 1.02 0.24** 0.23** (0.82)4. Cognitive appraisals of employee serviceperformance 4.04 0.98 0.19* 0.29** 0.65** (0.90)5. Customers’ epistemic motivation 2.68 1.06 –0.08 0.13 0.04 –0.09 (0.83)6. Customer loyalty intentions 3.98 1.12 0.21** 0.26** 0.65** 0.78** –0.07 (0.94)

a n 5 201.*p , 0.05

**p , 0.01Note: Internal reliabilities (Cronbach’s alphas) for the constructs are provided in parentheses on the diagonal.


laboratory replication is motivated by two purposes.Complementary studies with different research ap-proaches can overcome weaknesses of one researchmethod. Compared to Study 1, the laboratory ex-periment enhances internal validity by controllingpotential confounding factors (e.g., ambient condi-tions, store busyness). Further, measuring epistemicmotivation with the need for structure scale allowsus to cross-validate our findings (Lykken, 1968).

Experiment Preparation

Video production. We used videotaped serviceinteractions as stimuli because of the flexibilityand realism that videotapes offer in managing

manipulations and controlling confounds (Grandeyet al., 2005). We recruited six female actors froma pool of part-time employees of local stores thatagreed to allow filming on their premises. After ini-tial auditions, two actors were selected. They prac-ticed for three weeks to operationalize variousexperimental conditions. We assessed the actors onthe basis of their acting skills and their comfort levelsfor performing deep acting and surface acting strat-egies (Hennig-Thurau et al., 2006). Based on thesescreenings, one actorwas chosen to act as the serviceemployee, and the other actor as the customer. Thepractice videos were taken and pretested for multi-ple rounds until the actors and experimenters weresatisfied with the performance. Final versions of the

FIGURE 2Interaction Effects of Display Dimensions and Motivation on Customer Loyalty Intentions

Panel A: Moderating effect of epistemic motivation on the relationship between display intensity and customer loyalty intentions

Panel B: Moderating effect of epistemic motivation on the relationship between display authenticity and customer loyalty intentions

4.20

4.00

3.80

3.60

1.50 2.00 2.50 3.00 3.50 4.00Customers’ Epistemic Motivation

Cu

stom

er F

utu

re L

oyal

ty I

nte

nti

ons

4.40

4.20

4.00

3.80

3.60

3.40

1.50 2.00 2.50 3.00 3.50 4.00Customers’ Epistemic Motivation

Cu

stom

er F

utu

re L

oyal

ty I

nte

nti

ons

Panel A: Moderating effect of epistemic motivation on the relationship between display intensity and customer loyalty intentions.Panel B: Moderating effect of epistemic motivation on the relationship between display authenticity and customer loyalty intentions


videos were filmed in situ in a local store. All videoswere professionally produced. They were approxi-mately the same length (two and a half minutes) andhad similar levels of interpersonal contact. To min-imize the influence of on-screen customers on theaudience’s perceptions, the customers could be seenonly from the back. The scripts and scenes wereedited to be identical across conditions.

Manipulation of employees’ emotional intensityand authenticity. The actor who played the role ofthe employee displayed a smile, as defined by a no-ticeable upward twist of the lips (Barger & Grandey,2006). To manipulate intensity, the actor: (a) usedvoice intonation, gesticulated, and showed broadand frequent smiles in the high-intensity conditions;and (b) spoke calmly, avoided gesticulating, anddisplayed slight and less frequent smiles in the low-intensity conditions.

Following Hennig-Thurau et al. (2006), we oper-ationalized authenticity by training the actor in deepand surface acting techniques. When employees use“deep acting” strategies, they modify their internalfeelings to match the displayed emotion, leading toauthentic affective displays; when they use “surfaceacting” strategies, they fake manifest expressionsand suppress felt emotions, resulting in inauthenticemotional expressions (Grandey, 2003). The trainingincluded readingpertinentmaterials (e.g.,Hochschild,1983), discussing the concepts of emotional labor, andpracticing acting strategies. For example, the actorwastrained in deep acting to engender genuine positiveinner feelings and displays, and she used Stanislavski’s(1965) acting technique to recall pleasant emotionalmemories and think of the upcoming customer en-counters as an opportunity to help someone andmake her feel good.

Pretest. We conducted a pretest to ascertain thatintended authenticity and intensity levels weresuccessfully manipulated in the videos. A total of131 participants were randomly assigned to watchone of four experimental videos and to rate the au-thenticity and intensity of the employees’ displayedemotions. We adopted the same scale for perceivedintensity and authenticity used in Study 1. The au-thenticity manipulation was successful (Mauth 52.64,Minauth5 1.63;F (1, 127)5 44.09,p, 0.01). Theintensity manipulation did not influence perceivedauthenticity. The intensity manipulation was simi-larly successful (Mhigh5 3.09,Mlow5 2.07;F (1, 127)5142.36, p, 0.01). However, the authenticity manipu-lation also had an effect on perceived display intensity(Mauth 5 3.30, Minauth 5 2.69; F (1, 127) 5 15.73, p ,0.01), suggesting authentic displays of positive affect

appear more intense than inauthentic displays of pos-itive affect. Previous research has also observed thisspill-over impact of authenticity manipulation on in-tensity (see Hennig-Thurau et al., 2006). Importantly,as in past research, the intended intensity manipula-tion had a significantly greater effect on perceived in-tensity (h2 5 0.53) than the unintended authenticitymanipulation (h250.11).This largedifference ineffectsize supports the validity of the manipulation.

Participants and Procedure for the MainExperiment

Participants in this study were recruited from anonline research panel (N 5 343, 51% female, 18–75years of age, Mage 5 34.2, SD 5 12.0). Participantswere randomly assigned to conditions in a 2 (displayauthenticity: authentic vs. inauthentic) 3 2 (displayintensity: high vs. low) between-subjects design.Theywatched the service encounter video from theperspective of the customer in the scene. Immedi-ately after watching the video, participants completeda questionnaire containing the dependent measuresand ancillary questions (e.g., demographics).

Measurement of customers’ epistemicmotivation.Customers’ epistemicmotivationwasmeasured by the11-item need for structure scale (Neuberg & Newsom,1993). Prior researchhasvalidated this scale’s ability toassess individuals’ chronic levels of motivation to en-gage in systematic information processing (Van Kleefet al., 2009). Sample items from this scale include: “Ienjoy having a clear and structured mode of life”; “Itupsetsme togo intoa situationwithoutknowingwhat Ican expect from it.”Participants in our study indicatedhow much they agree with each statement on a scaleranging from 15 “strongly disagree,” to 75 “stronglyagree” (a 5 0.89).

Measurement of customers’ responses. We usedthe same scales from Study 1 to measure customer’saffective reactions (a 5 0.86) and loyalty intentions(a 5 0.97). Similar to Study 1, we measured cus-tomers’ cognitive appraisals about employee serviceperformance using a three-item scale. The scaleitems include: “This service employee was willingand able to provide the service I need;” “the em-ployee kept the best interests of the customer inmind;” “the employee seemed to achieve his/herown goals by satisfying me.” The first item was fromStudy 1, the second and third items were adoptedfrom Groth et al. (2009). We adopted the latter twoitems for this study as they specifically assess cus-tomers’ appraisals of employees’behaviors andserviceperformance. In addition, we measured customers’


assessment of display intensity and display authen-ticity using the same scale items as in Study 1.

Results and Discussions

To test stage-twomoderatedmediation as outlinedin our theoretical diagram, we again used structuralequation modeling with Mplus (Muthen & Muthen,2012). Similar to Study 1, latent factors (affectivereactions, cognitive appraisals, epistemicmotivation)and their interactions were modeled using standard-ized scale scores following recommendedproceduresby Cortina et al. (2001). We used all three items asindicators for customers’ future loyalty intention. Themeans, standard deviations, and correlation matricesof customer response variables are presented inTable 3. The model had good fit to the data (x2 (24)550.91; RMSEA5 0.06; SRMR 5 0.04; CFI5 0.99). Ingeneral, the results were consistent with our hypoth-eses. Intensity had a positive effect on customers’ af-fective reactions (b 5 0.06, p 5 0.10), although thisrelationship did not reach standard levels of signifi-cance. Authenticity also positively predicted cus-tomers’ cognitive appraisals (b 5 0.30, p , 0.01).Thus, Hypothesis 1 and Hypothesis 2 were sup-ported. Customers’ positive affective reactions (b 50.50, p , 0.01) and favorable cognitive appraisals(b 5 0.44, p , 0.01) further enhanced their loyaltyintentions, supportingHypothesis 3andHypothesis4.Moreover, in linewithHypothesis 5 andHypothesis 6,findings revealed significant two-way interactions ofaffective reactions X epistemic motivation (b5 –0.09,p , 0.05) and cognitive appraisals X epistemic moti-vation on loyalty intentions (b5 0.09, p, 0.05).3

Similar to Study 1, we used a bootstrapping-basedapproach with 10,000 samples to estimate the in-direct effects (Preacher & Hayes, 2008). The indirecteffect of intensity on loyalty intentions via affectivereactions was weaker for customers with high epi-stemic motivation (effect 5 0.09, 90% CI 5 0.00,

0.22) than for customers with low epistemic moti-vation (effect 5 0.14, 90% CI 5 0.00, 0.30). The in-direct effect of authenticity on loyalty intentions asmediated by appraisals of employee service perfor-mance was stronger for customers with higher epi-stemic motivation (effect 5 0.62, 90% CI 5 0.41,0.87) than for customers with lower epistemic mo-tivation (effect 5 0.40, 90% CI 5 0.24, 0.61).

This study provides further support for our pro-posed theoretical paradigm. In this experimentalstudy, as well as the field study, we measured cus-tomers’ epistemicmotivation andprovidedevidencefor its role in moderating the relationship betweenemployees’ affective displays and customers’ responses.

GENERAL DISCUSSION

The current research was conducted to better un-derstand the effects of two dimensions of em-ployees’ positive affective displays—intensity andauthenticity—in service contexts. One field studyand one laboratory experiment were conducted totest our theoretical framework, which utilized andextended the EASI model of emotion communica-tion. The results consistently demonstrated that:(1) affective display intensity primarily influencescustomer outcomes through affective reactions whileauthenticity does so mainly through cognitive ap-praisals; and (2) the impact of employees’ affectivedisplay intensity and authenticity on customer out-comes varies as a function of the customer’s epistemicmotivation. In other words, employees’ display in-tensity enhances loyalty intentions through cus-tomers’ affective reactions, especially for customerswith low epistemicmotivation. Employees’ displayauthenticity affects loyalty intentions through cus-tomers’ appraisals about employee service perfor-mance, and this pathway is more influential forcustomers with high epistemic motivation. Next,we consider the implications of these findings forresearch and practice.

Implications for Research

Our EASI-based model clarifies the distinct rolesof authenticity and intensity in influencing customerresponses and contributes to the literature on theeffectiveness of employees’ affective displays. Ourresearch provides an overarching framework forunderstanding the influence of different dimensionsof employees’ emotional displays on customer re-sponses. Further, this framework is consistent withand enhances the understanding of previous research.

3 We also conducted supplemental analyses in whichwe use the measured versions of the intensity and au-thenticity variables. Similar patterns of results were ob-served. Intensity positively predicted customers’ affectivereactions (b 5 0.16, p , 0.01), authenticity positivelyinfluenced customers’ cognitive appraisals (b 5 0.32, p ,0.01). Affective reactions (b5 0.49,p, 0.01) and cognitiveappraisals (b5 0.45, p, 0.01) enhanced customer loyaltyintentions. Epistemic motivation moderated such effectsas both affective reactions X epistemic motivation (b5–0.09, p , 0.05) and cognitive appraisals X epistemicmotivation ( b 50.10, p , 0.05) were significant.


Below we discuss our results in the context of thisprevious work, consider the implications of our re-search in this context, and identify future researchdirections.

Although extant literature has highlighted the im-portance of both intensity and authenticity of em-ployees’ positive affective displays in service delivery,most studieshavenotdistinguishedbetween these twodimensionsorhave focusedononedimensionwithoutmeasuring the other. Several papers concluded that allsmiles are “not created equal” or that “service witha smile” is not enough (Grandey et al., 2005; Hennig-Thurau et al., 2006), yet the ways in which differenttypes of positive emotional expressions affect ob-servers have not been explicated. For some customers,“service with a smile”may be the best strategy while,for others only authentic positive expressions will do.In the current research, we propose a theoreticalframework that can elucidate which customers re-spond better to intense versus authentic displays,and why.

We know of only one paper (Hennig-Thurau et al.,2006) that has explored the simultaneous roles ofdisplay intensity and authenticity. Hennig-Thurauet al. (2006) showed that after accounting for theeffects of authenticity, the effects of intensity are nolonger significant. They conclude authenticity isthe driving force influencing customer responses toemployees’ emotional displays. We propose anotherpossibility—that both dimensions of display influ-ence customer responses, but in different situationsand through different mechanisms. Indeed, theirresearch context—amovie consulting service—wasdescribed as a personalized service in which par-ticipants would answer a series of questions abouttheir movie-viewing habits and preferences ina one-to-one consultation. This type of service islikely to be perceived as highly personally in-volving, thereby increasing customers’ epistemicmotivation (Van Kleef et al., 2009). Their findings

are consistentwith our EASI-basedmodel, such thatwhen customers’ epistemic motivation is higher, au-thenticity is more influential than intensity. Our re-search demonstrates the importance of consideringdimensions of display and the context in which theyoccur.

Previous findings on the positive effects of em-ployees’ emotional displays on customer reactionsare also consistent with our model—and can be in-formed by it. For example, Wang and Groth (2014)found that the effect of authenticity was accentuatedwhen customers had a weak relationship with theemployee, or when they received highly personal-ized services. Gabriel, Acosta, and Grandey (2013)examined how service familiarity moderates the ef-fectiveness of employees’positive affective displays.They found that such displays had a greater impacton customers’ judgments of employee service per-formance when customers were relatively unfamil-iar with the service. Our conceptual model andempirical findings complement and extend thesefindings. Both unfamiliar service contexts, whichpromote more accurate emotion information pro-cessing, and highly personalized services, whichtend to be personally involving, should increaseepistemicmotivation (VanKleef et al., 2009). If this isthe case, our model suggests that employees’ au-thentic displays are more likely to result in positivecustomer outcomes than intense displays in thesecontexts, and that this is due to heightened cognitiveappraisals of employees’ service performance.

In all, our model provides a general framework forunderstanding when and why emotional displaysinfluence customer reactions. It offers a comprehen-sive and textured theoretical approach that is uniquefrom extant work—and it provides a framework toguide future research. For example, the factors thatinfluence epistemic motivation are rich and varied.Variations in customers’ epistemic motivation arisefrom both individual and situational differences

TABLE 3Descriptive Statistics, Correlations, and Reliabilities for All Measures in Study 2a

Variables Mean SD 1 2 3 4

1. Affective reactions 4.53 1.46 (0.96)2. Cognitive appraisals 5.49 1.20 0.66** (0.87)3. Epistemic motivation 3.27 1.11 –0.11* 0.13* (0.89)4. Loyalty intentions 5.23 1.65 0.77** 0.77** –0.08 (0.97)

a n 5 343.*p , 0.05

**p , 0.01Note: Internal reliabilities (Cronbach’s alphas) for the overall constructs are provided in parentheses on the diagonal.


(De Dreu & Van Kleef, 2004; Van Kleef, 2009).Thus, identifying epistemic motivation as a criticalmoderator opens up future research paths on bothindividual and situational factors that affect customerreactions to employees’ emotional displays.

Our work also enriches emerging research on theEASI model. Prior studies testing the EASI modelhave exclusively focused on how observers react todisplayers’ expressions of different discrete emo-tions, such as anger (Van Kleef, Anastasopoulou, &Nijstad, 2010), happiness (VanKleef et al., 2004), anddisappointment and regret (Van Kleef, De Dreu, &Manstead, 2006). These studies did not distinguishbetween different dimensions of emotions, such asthe intensity and authenticity of displayers’ expres-sions. In this research, we integrate the EASI litera-ture with research on employees’ affective displays,which has considered intensity and authenticity asdistinct dimensions of positive expression. In gen-eral, our findings suggest that judgments about dis-crete emotions as social information also apply tocomparing different dimensions of a single emotion.We specifically investigated the dimensions of in-tensity and authenticity, but future research mayidentify other affective dimensions that can be studiedthrough the lens of this model. The results of our re-search also support the applicability of theEASImodelto employee–customer interactions, which are differ-ent from the situations examined in prior researchinvolving shared outcomes between observers anddisplayers (De Dreu & Van Kleef, 2004; Van Kleefet al., 2010).

Implications for Practice

Our research provides practical implications byilluminating which customers are most likely to re-spond positively to intense and authentic emotionaldisplays. Firms invest considerable resources intraining and monitoring their employees’ affectivedisplays to enhance service experience and im-prove customer relationships.Managers often focuson either the intensity or authenticity dimensionof employees’ emotional displays. For example,Chick-fil-A’s“theCoreFour” ruleasks thatemployeesmake eye contact, speak enthusiastically, smilebroadly, and engage the customer. Such behavior,although less genuine or authentic by its very nature(Gountas & Ewing, 2003), has several benefits. Em-ployees who are required to display intense (but notnecessarily authentic) positive emotions are easier torecruit, train, and monitor than those trained to beauthentic (Richard, 2006).

However, such scripted emotional displays—thatsomehave characterized as pseudo-care “lacking theessential humane attribute of genuine care” (e.g.,Lynch, 1992: 379)—maynot always be desirable. Forcertain customers, only authentic displays of posi-tive emotion will result in positive outcomes. Whileprior research supports the idea that customers gen-erally appreciate authentic positive displays (Hennig-Thurauet al., 2006), it is expensive and time-intensiveto recruit employees who exude natural warmth, ortrain them to display genuine positive emotions ona consistent basis. In addition, there are some situa-tions in which the advantage of authentic displaysdisappears (e.g., in busy service contexts; Grandeyet al., 2005), resulting in awaste of training resources.

Tomake themost efficient and informeddecisionsabout emotional display rules, managers need tounderstand when and why intensity and authentic-ity positively affect customer outcomes. Our re-search suggests that employees’ emotional displayrequirements should be contingent on customers’epistemic motivation. Our study focused on episte-mic motivation as a characteristic of the individual.Yet, as we note above, service context is importantas well. For example, we have suggested unfamiliarservices and personalized services would increasecustomers’ epistemic motivation. Similarly, routineservice encounters and services that are inherentlyless involving may lower customers’ epistemic moti-vation. When customers have low epistemic motiva-tion, emotional rules without finer-grain requirementsmay be better. For instance, failing to deliver positivedisplays can be detrimental in some contexts (e.g.,McDonald’s admits that it is unsustainable for thecompany to have 20% of customers’ complaints being“unfriendly employees”), but emphasis on authentic-ity may be an expensive investment that goes un-rewarded (e.g., customersmay appreciateMcDonald’semployees’ smiling, but are less likely to make in-ferencesabout theemployees’genuineness).Further, itis sometimes difficult for managers to empower thefrontline (as services are often standardized or scrip-ted) and invest extensively on employees’ genuineenthusiasm (as these positions often carry low sal-aries). Concrete requirements of frequent positive ex-pressions such as Walmart’s “10-foot attitude” orChick-Fil-A’s “the Core Four” would be easier toimplement and monitor than the requirement ofauthenticity.

On the other hand, when customers have highepistemic motivation, authentic positive emotionaldisplays are needed to maximize positive customeroutcomes. In such cases, employees’ genuine care


and enthusiasm should be emphasized. Recruitingemployees who are naturally high on positive emo-tionality or temperament (Larsen & Ketelaar, 1991),or are adept at sustaining higher emotional laborthrough reappraisal or deep acting—adjusting bothfelt and expressed emotions (Gross & John, 2003)—isrecommended. For training purposes,managersmayneed to adjust their techniques to include empow-erment or aspirations such as “the ‘wow stories’ ofRitz-Carlton personnel going above and beyond thecall of duty that are shared globally every week”(DeRose & Tichy, 2013). In certain service contexts,emphasizing affective intensity and the mechanicaldisplay of positive expressive behaviors may not besufficient to produce customer loyalty. Scripted in-tense positive displays may even backfire undercertain circumstances because high demand for fre-quent smiling may result in faking or inauthenticexpressions that ultimately undermine customeroutcomes.

Limitations and Future Research

All studies have limitations and ours are no ex-ception. For example, in the field study, we relied onobservers’ coding of emotional displays. Althoughthis is an established and conservative approachto test our propositions, with the added benefits ofminimized same-source bias (Barger & Grandey, 2006;Pugh, 2001; Tsai & Huang, 2002), future research maybenefit fromdirectlymeasuring customers’perceptionof the emotional display to further explore its down-streamconsequences.Wealsoused a context-specificmeasure of epistemic motivation in Study 1. Thismeasure has the benefit of directly assessing mo-tivation to interpret employees’ emotional display,but it had not been validated as a measure of episte-mic motivation. To alleviate this concern, we con-ducted a follow-up study in which we examined thecorrelation between the Study 1 measure of episte-mic motivation and the commonly used measure ofepistemic motivation that we employed in Study 2.The results of this follow-up demonstrate that theStudy 1 measure is strongly correlated with a stan-dard measure of epistemic motivation (r5 0.58, p,0.01). These results and the consistency of the find-ings from the two studies give us confidence in thecontext-specific measure created for Study 1. In ad-dition, laboratory experiments often suffer from lowecological validity. We designed the experiment toapproximate actual service experiences while con-trolling for potential confounds such as employeefamiliarity, but such experiments nevertheless seem

artificial when compared to real customer–employeeinteractions. Of course, one of the benefits of utilizingdifferent methodological approaches is that thestrengths of one (e.g., the richness of the field;control in the laboratory) complement the weak-nesses of the other. The consistency of the resultsacross the studies increases our confidence in therobustness of the findings.

The practical implications of our findings arequalified by two caveats. First, extant emotionallabor research, including the studies reportedherein, has focused on a limited assortment ofemployees’ affective displays (e.g., smiling, greet-ing, thanking the customer). However, Gremler andGwinner (2008) indicate that other aspects of em-ployees’ nonverbal behaviors, including attentivebehavior (e.g., back-channel responses), imitativebehavior (e.g., matching posture and voice tone), andcommon grounding behavior (e.g., discovering mu-tual interests), also merit attention. Second, althoughmany companies can infer the extent to which theirservices induce high or lowmotivation in customers,there isnodatabase thatmeasurescustomers’epistemicmotivation. Companies in service industries, suchas hotels, in which extended customer contact oc-curs, could create their own databases that infercustomers’ motivation from past interactions; otherfirmsmayneedtorelyonsituationalvariables (e.g., levelof involvement, the duration of the transaction, servicefamiliarity, relationship with the service provider) orpast customer feedback.

It is also worth noting that findings in bothstudies suggest that the intensity and authenticityfacets of employees’ affective displays are dis-tinctive, yet correlated (r 5 0.31 in Study 1). In-deed, in the laboratory, the intensity manipulationinfluenced perceptions of authenticity. This maybe due to several reasons. First, customers mightinfer slight smiles or infrequent expressions assimply courteous behaviors to comply with socialnorms, rather than the expression of genuine pos-itive internal feelings, as smiling and other positiveaffective displays are considered normative inmany service encounters (Pugh, 2001; Rafaeli &Sutton, 1990). In addition, extant research onDuchenne smiles suggests that authentic smilescompared to inauthentic smiles are more likely toinvolve the activation of the Orbicularis oculi (themuscle group surrounding the eyes), also known asthe Duchenne marker (Ekman & Friesen, 1982).Without a certain level of smile intensity or acti-vation of the zygomatic major muscle, it is hard fora displayer to contract the Obicularis oculi or show


the Duchenne marker (Hess et al., 2002). Thus, ourfindings are consistent with observations that authen-tic expressions may sometimes be perceived to bemore intense than inauthentic ones.

Two alternativemodelsmight also be considered.4

First, although the EASI model and previous re-search (van Kleef et al., 2004) indicate that epistemicmotivation should act as a second-stage moderator,one might wonder if epistemic motivation influencesthe relationship between intensity and authenticityand affective reactions and cognitive appraisal. Toexplore this possibility, we conducted supplemen-tary analyses utilizing the data from Study 1. Theresults of the supplementary analyses reveal no sig-nificant first stage moderation. Further, when episte-mic motivation is included as both a first-stageand second-stage moderator, the pattern of resultsremains unchanged. Second, it is possible that in-tensity influences cognitive appraisal and authentic-ity influences affective reactions. It is conceivable, insomecontexts, that intensity directly affects cognitiveappraisals and authenticity directly affects affectivereactions. Future research might consider contextualvariables that enhance the relationship between in-tensity and cognitive appraisal and between authen-ticity and affective reactions.

In this paper, we considered customers’ epistemicmotivation as individual-level differences. Fu-ture research may further examine whether situ-ational factors may affect customers’ epistemicmotivation, which in turn shape the effects ofdisplay intensity and authenticity. This knowl-edge may provide actionable guidelines for com-panies to determine whether to emphasizeintensity or authenticity in employee trainingprograms. For instance, prior research has shownthat one situational factor that influences episte-mic motivation is power (Keltner, Gruenfeld, &Anderson, 2003), such that high power decreasesmotivation while low power enhances it (De Dreu& Van Kleef, 2004). Along this line, companiesthat adhere to the policy “the customer is king”are likely to benefit from employees who displayintense positive emotions, whereas those whoseemployees assume a more powerful role (e.g., inthe health care industry) may benefit more fromauthentic displays. Future research might alsoexplore other contextual factors that influenceepistemic motivation such as purchase involvement(Shao, Baker, & Wagner, 2004) and time pressure(Suri & Monroe, 2003).

In addition, extant literature, including the cur-rent research, has not addressed how culturalvariables influence the interplay of employeeemotional displays and customer responses inservice deliveries. It is possible that customers’epistemic motivation is contingent on their cul-tural group membership. For instance, customersshould be more motivated to process emotionalcues from an in-group versus an out-group em-ployee (Young & Hugenberg, 2010), as peopleperceive in-group members to be more importantsocial sources and are more concerned abouttreatment from them (Correll & Park, 2005). Normsabout emotion-communication also vary acrosscultures. Culture-specific encoding and decodingrules may lead customers from different culturalgroups to process employees’ displayed emotionsdifferently. For instance, one core personality traitof Hispanic culture is expressive sociability(Triandis, 1989). Thus, the effects of intensity maybe strengthened, as positive expressions (e.g., smileand laugh) and interpersonally engaging behaviors(e.g., hugging, touching) are valued andemphasizedin the Hispanic culture (Martin, Hammer, &Bradford, 1994).

Conclusion

In conclusion, our findings support prior researchthat indicates that employees’ affective displays arean effective tool to enhance customer experience inservices. However, our conceptualization highlightstheneed toconsider the impactofdifferentdimensionsof emotional displays on customer responses. Ourstudies reveal that the effectiveness of intensity andauthenticity is determined by customers’ epistemicmotivation in different yet theoretically predictableways. The results of this research thus contribute totheoretical knowledge about emotional labor strate-gies, suggest directions for future research, and pro-vide implications and recommendations aboutemployee training strategies for managers.

REFERENCES

Albarracin, D., & Kumkale, G. T. 2003. Affect as informa-tion inpersuasion:Amodel of affect identification anddiscounting. Journal of Personality and Social Psy-chology, 84: 453–469.

Albarracin, D., & Wyer, R. S. 2001. Elaborative and non-elaborative processing of a behavior-related communi-cation. Personality and Social Psychology Bulletin, 27:691–705.4 We thank the anonymous reviewers for these suggestions.


American Society for Training and Development 2013.The American Society for Training and Develop-ment (ASTD) state of the industry annual report.Available at: http://www.td.org/Publications/Blogs/ATD-Blog/2013/ASTD-Release-2013-State-of-the-Industry-Report. Accessed June 39, 2015.

Barger, P. B., & Grandey, A. A. 2006. Service with a smileand encounter satisfaction: Emotional contagion andappraisal mechanisms. Academy of ManagementJournal, 49: 1229–1238.

Berry, L. L.,Wall, E.A., &Carbone, L. P. 2006. Service cluesand customer assessment of the service experience:Lessons from marketing. The Academy of Manage-ment Perspectives, 20: 43–57.

Brady, M. K., & Cronin, J. J., Jr.. 2001. Some new thoughtson conceptualizing perceived service quality: A hier-archical approach. Journal of Marketing, 65: 34–49.

Chi,N.W.,Grandey,A.A.,Diamond, J.A., &Krimmel,K.R.2011. Want a tip? Service performance as a functionof emotion regulation and extraversion. The Journalof Applied Psychology, 96: 1337–1346.

Correll, J., & Park, B. 2005. A model of the ingroup as a so-cial resource. Personality and Social PsychologyReview, 9: 341–359.

Cortina, J. M., Chen, G., & Dunlap, W. P. 2001. Testinginteraction effects in LISREL: Examination and illus-tration of available procedures. Organizational Re-search Methods, 4: 324–360.

Cote, S. 2005. A social interaction model of the effectsof emotion regulation on work strain. Academy ofManagement Review, 30: 509–530.

Cote, S., &Hideg, I. 2011. The ability to influence others viaemotion displays: A new dimension of emotional in-telligence.OrganizationalPsychologyReview, 1: 53–71.

DeRose, C., & Tichy, N. 2013. How about some “happy”with that happy meal. Forbes, 23 April.

DeDreu, C. K.W., &VanKleef, G. A. 2004. The influence ofpower on the information search, impression forma-tion, and demands in negotiation. Journal of Experi-mental Social Psychology, 40: 303–319.

Ekman, P. 1993. Facial expression and emotion. TheAmerican Psychologist, 48: 384–392.

Ekman, P., & Friesen,W. V. 1982. Felt, false, andmiserablesmiles. Journal of Nonverbal Behavior, 6: 238–252.

Ekman, P., Friesen, W. V., & O’Sullivan, M. 1988. Smileswhen lying. Journal of Personality and Social Psy-chology, 54: 414–420.

Ekman, P., O’Sullivan, M., & Frank, M. G. 1999. A few cancatch a liar. Psychological Science, 10: 263–266.

Fiske, A. P. 1992. The four elementary forms of sociality:Framework for a unified theory of social relations.Psychological Review, 99: 689–723.

Fiske, S. T., & Neuberg, S. L. 1990. A continuum model ofimpression formation: From category-based to indi-viduating processes as a function of information, mo-tivation, and attention. Advances in ExperimentalSocial Psychology, 23: 1–74.

Fridlund, A. J. 1992. The behavioral ecology and socialityof human faces. InM. S. Clark (Ed.), Emotion: 90–121.Newbury Park, CA: Sage.

Gabriel, A. S., Acosta, J. D., & Grandey, A. A. 2013. Thevalue of a smile: Does emotional performance mattermore in familiar or unfamiliar exchanges? Journalof Business and Psychology, 30: 37–50.

Gountas, S., & Ewing, M. 2003. The softer side of services:exploring the effect of sincerity and emotional con-tagion on satisfaction. ANZMAC 2003 ConferenceProceedings.

Grandey,A.A. 2003.When “the showmust go on”: Surfaceand deep acting as determinants of emotional ex-haustion and peer-rated service delivery.Academy ofManagement Journal, 46: 86–96.

Grandey, A. A., Fisk, G. M., Mattila, A. S., Jansen, K. J., &Sideman, L.A. 2005. Is “servicewith a smile” enough?Authenticity of positive displays during service en-counters. Organizational Behavior and HumanDecision Processes, 96: 38–55.

Greifeneder, R., & Bless, H. 2007. Relying on accessiblecontent versus accessibility experiences: The caseof processing capacity. Social Cognition, 25:853–881.

Gremler, D. D., & Gwinner, K. P. 2008. Rapport-buildingbehaviors used by retail employees. Journal of Re-tailing, 84: 308–324.

Gross, J., & John, O. P. 2003. Individual differences in twoemotion regulation processes: Implications for affect,relationships and well-being. Journal of Personalityand Social Psychology, 85: 348–362.

Groth, M., Hennig-Thurau, T., &Walsh, G. 2009. Customerreactions to emotional labor: The roles of employeeacting strategies and customer detection accuracy.Academy of Management Journal, 52: 958–974.

Hatfield, E., Cacioppo, J. T., & Rapson, R. L. 1992. Primitiveemotional contagion.Review of personality and socialpsychology In M. S. Clark (Ed.), Emotion and socialbehavior, vol. 14: 151–177. Thousand Oaks, CA: SagePublications.

Hatfield, E., Rapson, R. L., & Le, Y. C. L. 2009. Emotionalcontagion and empathy. In J. Decety &W. Ickes (Eds.),The social neuroscience of empathy: 19–30. Boston,MA: MIT Press.

Hennig-Thurau, T., Groth, M., Paul, M., & Gremler, D. D.2006. Are all smiles created equal? How emotionalcontagion and emotional labor affect service rela-tionships. Journal of Marketing, 70: 58–73.


http://www.td.org/Publications/Blogs/ATD-Blog/2013/ASTD-Release-2013-State-of-the-Industry-Report



Hess, U., Beaupre, M. G., & Cheung, N. 2002. Who towhom and why-cultural differences and similaritiesin the function of smiles. In M. H. Abel (Ed), An em-pirical reflection on the smile. 187-216. Lewiston,NY: The Edwin Mellen Press.

Hochschild, A. R. 1983. The managed heart: Commer-cialization of human feeling. Berkeley, CA: Univer-sity of California Press.

Hu, L. T., & Bentler, P. M. 1999. Cutoff criteria for fit in-dexes in covariance structure analysis: Conventionalcriteria versus new alternatives. Structural EquationModeling, 6: 1–55.

Huy, Q. N. 1999. Emotional capability, emotional intelli-gence, and radical change.Academy ofManagementReview, 24: 325–345.

Keh, H. T., Ren, R., Hill, S. R., & Li, X. 2013. The beautiful,the cheerful, and the helpful: The effects of serviceemployee attributes on customer satisfaction. Psy-chology and Marketing, 30: 211–226.

Keltner, D., Gruenfeld, D. H., & Anderson, C. 2003. Power,approach, and inhibition.Psychological Review, 110:265–284.

Kruglanski, A. W. 1989. Lay epistemics and human knowl-edge:Cognitiveandmotivational bases. NewYork,NY:Plenum.

Kruglanski, A. W., & Webster, D. M. 1996. Motivatedclosing of the mind: “Seizing” and “freezing.” Psy-chological Review, 103: 263–283.

Krumhuber, E., Manstead, A. S., Cosker, D., Marshall, D.,Rosin, P. L., & Kappas, A. 2007. Facial dynamics asindicators of trustworthiness and cooperative behav-ior. Emotion (Washington, D.C.), 7: 730–735.

Laird, J. D., & Bresler, C. 1992. The process of emotionalexperience: A self-perception theory. Review of per-sonality and social psychology In M. S. Clark (Ed.),Emotion, vol. 13: 213–234. Thousand Oaks, CA: SagePublications.

Larsen, R. J., & Ketelaar, T. 1991. Personality and suscepti-bility to positive and negative emotional states. Journalof Personality and Social Psychology, 61: 132–140.

Liao, H., & Chuang, A. 2007. A multilevel investigation offactors influencing employee service performance andcustomer outcomes.Academy ofManagement Journal,47: 41–58.

Lynch, J. 1992. Hear it from the heart. Managing ServiceQuality, 3: 379–383.

Lykken, D. T. 1968. Statistical significance in psychologi-cal research. Psychological Bulletin, 70: 151–159.

Martin, J. N., Hammer, M. R., & Bradford, L. 1994. Theinfluence of cultural and situational contexts on His-panic and non-Hispanic communication competencebehaviors. Communication Quarterly, 42: 160–179.

Martini, T. S., & Busseri, M. A. 2012. Emotion regulationand relationship quality in mother–young adult childdyads. Journal of Social andPersonal Relationships,29: 185–205.

Mayseless, O., & Kruglanski, A. W. 1987. What makes youso sure? Effects of epistemic motivations on judgmentalconfidence. Organizational Behavior and HumanDecision Processes, 39: 162–183.

Muthen, B., & Asparouhov, T. (2011). Beyond multilevelregression modeling: Multilevel analysis in a generallatent variable framework. In J. Hox & J. K. Roberts(Eds), Handbook of advanced multilevel analysis,15-40. New York, NY: Taylor & Francis.

Muthen, L. K., & Muthen, B. O. 2012. Mplus: Statisticalanalysis with latent variables (Version 7). Los Angeles,CA: Muthen & Muthen.

Neuberg, S. L., & Newsom, J. T. 1993. Personal need forstructure: Individualdifferences in thedesire for simplerstructure. Journal of Personality and Social Psychol-ogy, 65: 113–131.

Payne, S. C., & Webber, S. S. 2006. Effects of service pro-vider attitudes and employment status on citizenshipbehaviors and customers’ attitudes and loyalty be-havior. The Journal of Applied Psychology, 91:365–378.

Preacher, K. J., Curran, P. J., & Bauer, D. J. 2006. Compu-tational tools for probing interactions in multiple lin-ear regression, multilevel modeling, and latent curveanalysis. Journal of Educational and BehavioralStatistics, 31: 437–448.

Preacher, K. J., & Hayes, A. F. 2008. Asymptotic andresampling strategies for assessing and comparingindirect effects in multiple mediator models. Behav-ior Research Methods, 40: 879–891.

Preacher, K. J., Zyphur, M. J., & Zhang, Z. 2010. A generalmultilevel SEM framework for assessing multilevelmediation. Psychological Methods, 15: 209–233.

Pugh, S. D. 2001. Service with a smile: Emotional conta-gion in the service encounter. Academy of Manage-ment Journal, 44: 1018–1027.

Rafaeli, A., & Sutton, R. I. 1990. Busy stores and de-manding customers: How do they affect the displayof positive emotion? Academy of ManagementJournal, 33: 623–637.

Reichheld, F. F., & Teal, T. 2001. The loyalty effect: Thehidden force behind growth, profits, and lastingvalue. Boston, MA: Harvard Business Press.

Richard, E. M. 2006. Applying appraisal theories of emo-tion to the concept of emotional labor. DissertationAbstract International, 67, UMI No. AAT 3208194.

Rust, R. T., & Zahorik, A. J. 1993. Customer satisfaction,customer retention, and market share. Journal of Re-tailing, 69: 193–215.


Salanova, M., Agut, S., & Peiro, J. M. 2005. Linking orga-nizational resources andwork engagement to employeeperformance and customer loyalty: The mediation ofservice climate. The Journal of Applied Psychology,90: 1217–1227.

Shao, C. Y., Baker, J. A., & Wagner, J. 2004. The effects ofappropriateness of service contact personnel dress oncustomer expectations of service quality andpurchaseintention: Themoderating influences of involvement andgender. Journal of Business Research, 57: 1164–1176.

Sinclair, R. C. 1988. Mood, categorization breadth, and per-formance appraisal: The effects of order of informationacquisition and affective state on halo, accuracy, infor-mation retrieval, and evaluations. Organizational Be-havior and Human Decision Processes, 42: 22–46.

Skinner, B. F. 1953. Science and human behavior. NewYork, NY: Macmillan.

Suri, R., & Monroe, K. B. 2003. The effects of time con-straints on consumers’ judgments of prices and prod-ucts.The Journal of ConsumerResearch, 30: 92–104.

Stanislavski, C. 1965. An actor prepares. New York, NY:Theatre Art Books.

Swan, J. E., & Oliver, R. L. 1989. Postpurchase communica-tions by consumers. Journal of Retailing, 65: 516–533.

Triandis, H. C. 1989. The self and social behavior in differingcultural contexts. Psychological Review, 96: 506–520.

Tsai, W. C., & Huang, Y. M. 2002. Mechanisms linkingemployee affective delivery and customer behavioralintentions. The Journal of Applied Psychology, 87:1001–1008.

Van Kleef, G. A. 2009. How emotions regulate social lifethe emotions as social information (EASI) model. Cur-rent Directions in Psychological Science, 1: 184–188.

VanKleef, G. A., Anastasopoulou, C., &Nijstad, B. A. 2010.Can expressions of anger enhance creativity? A test ofthe emotions as social information (EASI)model. Journalof Experimental Social Psychology, 46: 1042–1048.

Van Kleef, G. A., De Dreu, C. K., & Manstead, A. S. 2004.The interpersonal effects of emotions in negotiations:a motivated information processing approach. Journalof Personality and Social Psychology, 87: 510–528.

Van Kleef, G. A., De Dreu, C. K., & Manstead, A. S. 2006.Supplication and appeasement in conflict and nego-tiation: The interpersonal effects of disappointment,worry, guilt, and regret. Journal of Personality andSocial Psychology, 91: 124–142.

VanKleef,G.A.,Homan,A.C.,Beersma,B.,VanKnippenerg,D., Van Knippenberg, B., & Damen, F. 2009. Searingsentiment or cold calculation? The effects of leaderemotional displays on team performance depend onfollower epistemic motivation. Academy of Manage-ment Journal, 52: 562–580.

Wang, K. L., & Groth, M. 2014. Buffering the negative ef-fects of employee surface acting: The moderating roleof employee–customer relationship strength and per-sonalized services. The Journal of Applied Psychol-ogy, 99: 341–350.

Young, S. G., & Hugenberg, K. 2010. Mere social categori-zation modulates identification of facial expressionsof emotion. Journal of Personality and Social Psy-chology, 99: 964–977.

Ze Wang ([email protected]) is an assistant professor ofmarketing at the College of Business Administration,University of Central Florida. She received her PhD fromthe University of Kansas. Her research examines cus-tomers’ service experience fromboth customers’ and firms’perspectives, particularly with respect to frontline man-agement, impression formation, social judgments, andconsumer-generated content analysis.

Surendra N. Singh ([email protected]) is AT&T FoundationProfessor of Business, and Area Director: Marketing and Busi-ness Law, School of Business and Professor of Health PolicyandManagement, School of Medicine, University of Kansas.

Yexin Jessica Li ([email protected]) is an assistant pro-fessor of marketing at the School of Business, University ofKansas. She receivedher PhD fromArizonaStateUniversity.Her research interests include social cognition and the role ofmotivations and emotions on consumer decision-making.

Sanjay Mishra ([email protected]) is associate professor ofMarketing at the School of Business, University of Kansas.He received his PhD from Washington State University.His research focuses on consumers’ response to informa-tion, e.g., new introductions, brands, web sites, environ-mental cues.

Maureen Ambrose ([email protected]) is the Gordon J.Barnett Professor of Business Ethics at the University ofCentral Florida. She received her PhD from the Universityof Illinois at Urbana-Champaign. Her research interestsinclude organizational fairness, ethics, and workplacedeviance.

Monica Biernat ([email protected]) is Professor of Psychol-ogy and Associate Chair at the University of Kansas whereshe directs the PhD program in social psychology. Herprimary research examines the processes of stereotypingand prejudice, and focuses on how stereotypes affectjudgments of individual members of stereotyped groups.She is currently Editor of Personality and Social Psychol-ogy Review and Executive Officer of the Society for Ex-perimental Social Psychology.


mailto:[email protected]






Copyright of Academy of Management Journal is the property of Academy of Managementand its content may not be copied or emailed to multiple sites or posted to a listserv withoutthe copyright holder's express written permission. However, users may print, download, oremail articles for individual use.

Documents

EFFECTS OF EMPLOYEES POSITIVE AFFECTIVE DISPLAYS ON ... · customer. ― Ronnie Clotfelter, owner and operator of Chick-fil-A, Sharpsburg, Georgia Focus on winning one customer at