CUSTOMER REACTIONS TO EMOTIONAL LABOR: THE ROLES
OF EMPLOYEE ACTING STRATEGIES AND CUSTOMER
DETECTION ACCURACY
MARKUS GROTH
University of New South Wales
THORSTEN HENNIG-THURAU
Bauhaus-University of Weimar and City University London
GIANFRANCO WALSH
University of Koblenz-Landau
In this research, we extend emotional labor theories to the customer domain by
developing and testing a theoretical model of the effects of employee emotional labor
on customer outcomes. Dyadic survey data from 285 service interactions between
employees and customers show that employees’ emotional labor strategies of deep and
surface acting differentially influence customers’ service evaluations and that custom-
ers’ accuracy in detecting employees’ strategies can intensify this impact. We also
investigate the potential moderating effects of service type on the relationship between
emotional labor and customer outcomes but find no support for such an effect.
Recent service management research increas-
ingly focuses on the role of emotions in service
delivery, particularly the emotional labor per-
formed by service employees. Frontline workers
are expected to display certain emotions (e.g., hap-
piness) and suppress others (e.g., anger) in their
daily interactions with customers to comply with
their job requirements and organizational expecta-
tions. Against this background, the concept of emo-
tional labor—the “process of regulating both feel-
ings and expressions for the organizational goals”
(Grandey, 2000: 97)—has received ample attention
in existing research in an effort to understand how
service organizations can better deliver “service
with a smile” to their customers by effectively man-
aging their employees’ emotional display (for a re-
cent review, see Grandey [2008]).
Since Hochschild (1983) introduced the concept,
most research on emotional labor has focused on its
dimensionality and its effects on employee well-
being. With regard to emotional labor dimensions,
deep acting (attempting to modify felt emotions so
that a genuine emotional display follows) and sur-
face acting (faking or amplifying emotions by dis-
playing emotions not actually felt) represent two
main strategies of regulating emotion that employ-
ees use to comply with expectations of emotional
display (Hochschild, 1983; Kruml & Geddes, 2000).
Research has highlighted several negative conse-
quences of emotional labor on employees, includ-
ing psychological health problems such as stress,
burnout, and emotional exhaustion (Brotheridge &
Grandey, 2002; Hochschild, 1983; Morris & Feld-
man, 1997), but has only sparingly addressed the
question whether emotional labor also influences
customers. This gap is surprising, because the per-
sonal interaction between a service employee and
customer comprises an essential part of the service
experience (Bitner, 1990; Bowen, 1990), and the
role of emotions in influencing social processes
(Hatfield, Cacioppo, & Rapson, 1994; Hochschild,
1979) provides reason to believe that employees’
emotion regulation during service interactions af-
fects customer outcomes, such as service quality
and customer loyalty, that are critical for service
success.
This research is an attempt to narrow the existing
research gap by examining the link between em-
ployees’ emotional labor and customers’ percep-
tions of the service experience. Specifically, we
consider whether employees’ emotional labor strat-
egies—either genuine emotional display (deep act-
ing) or fake emotional display (surface acting)—
affect customers’ evaluations of service experiences
This research was funded by grants from the (U.K.)
Academy of Marketing and the University of Strathclyde
Business School. We wish to thank Hillary Elfenbein and
Raja Chattapadhyay for providing valuable comments on
drafts. We also thank our three anonymous reviewers and
Elizabeth Morrison, to whom we are particularly grateful,
for their constructive help in improving this article.

Academy of Management Journal
2009, Vol. 52, No. 5, 958–974.
958
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 express
written permission. Users may print, download or email articles for individual use only.

and their long-term relationships with the service
providers. Do customers react positively to all dis-
plays of friendly emotions, even feigned ones, or do
they have a less positive service experience when
employees fake their emotional display? To what
extent are customers able to detect fake and genu-
ine emotions accurately, and how does this influ-
ence their assessments of service experience?
We addressed these questions by developing a
theoretical model of the links between employees’
emotional labor, customers’ accuracy in detecting
different emotional labor strategies, and the result-
ing customer outcomes, drawing on social psy-
chological research into emotion regulation and
emotion recognition. We tested the model empiri-
cally by analyzing the survey data of 285 matched
employee-customer dyads with partial least squares
(PLS) structural equation modeling. Our analysis of
dyads offers a unique contribution to research, as it
allowed us to focus on emotional labor performed
in a specific and immediate service transaction
rather than ask employees about their retrospective
behavioral patterns in general, as is often done in
emotional labor research.
THEORETICAL BACKGROUND
Role of Emotions in Service Interactions
Although some prior research has examined the
role of emotions in service interactions between
employees and customers, most of this research
focuses on employees’ outer emotional display
rather than their internal emotion regulation. For
example, some studies examine the relationship
between employees’ displayed emotions and cus-
tomers’ emotions by asking customers or indepen-
dent observers to rate the emotional display of ser-
vice employees and then linking these ratings to
service evaluations. Tsai (2001) and Tsai and
Huang (2002) uncovered a link between indepen-
dent observers’ assessments of employee affective
delivery and self-reported customer mood and loy-
alty intentions, and Mattila and Enz (2002) simi-
larly found a link between observational data on
hotel clerks’ emotional displays and customers’
service encounter evaluations, as well as positive
moods after the encounter. Tan, Foo, and Kwek
(2004) reported a link between the extent of em-
ployees’ positive emotions (measured by greeting,
eye contact, and so forth) and customer satisfac-
tion. These studies focused on outwardly displayed
emotions and primarily addressed whether service
friendliness and related observable behaviors lead
to positive customer evaluations. However, they
leave unexplored the underlying cognitive emotion
management processes and, in particular, specific
emotional labor strategies.
To our knowledge, only two studies have inves-
tigated how emotional labor strategies might influ-
ence service delivery outcomes. Hennig-Thurau,
Groth, Paul, and Gremler (2006), in a study of the
emotional contagion process, found a significant
impact of employees’ emotional authenticity on
customers’ emotions in a simulated service en-
counter. Grandey (2003) focused instead on the
concept of “affective delivery”—which she defined
as service delivery perceived as friendly and warm
by customers—as an outcome of emotional labor
strategies and found a positive relationship with
deep acting but a negative relationship with surface
acting. However, in her study coworkers of the
observed employees, rather than customers, as-
sessed the affective delivery. Although these stud-
ies have revealed important insights into related
phenomena, they have not revealed how emotional
labor strategies affect the customer experience, nor
the effects of whether customers are able to accu-
rately detect emotional authenticity.
Emotional Labor as an Emotion
Regulation Process
Emotional labor research focuses specifically on
the self-regulatory processes that employees use to
display emotions in compliance with organization-
al expectations. Service organizations usually have
explicit or implicit emotional display rules: that is,
norms and standards of behavior that indicate
which emotions are appropriate and should be
publicly expressed toward customers and which
should be suppressed (Hochschild, 1983; Rafaeli &
Sutton, 1987). For example, all employees at Ritz-
Carlton hotels need to follow “the Ritz-Carlton Ba-
sics,” service rules for dealing with customers,
spelled out on pocket-sized cards issued to all em-
ployees. One service rule reads: “Smile—We are on
stage. Always maintain positive eye contact.”
Researchers have identified deep and surface act-
ing as the two most commonly used emotional
labor strategies for coping with display rule re-
quirements (Hochschild, 1983; Kruml & Geddes,
2000; Zapf, 2002). In deep acting, employees at-
tempt to modify their felt emotions so that a genu-
ine, organizationally desired emotional display can
follow. Hochschild (1983) exemplified deep acting
by citing flight attendants who cope with angry and
annoying passengers by thinking of them as fright-
ened first-time fliers, therefore changing their inner
feelings from annoyance to pity and empathy.
When deep acting, employees endeavor to express
authentic emotions, and though not every attempt
2009 959Groth, Hennig-Thurau, and Walsh

succeeds, emotions expressed as a result of deep
acting are more likely to be authentic than those
expressed through surface acting, which occurs
when employees only change their outward emo-
tional display without genuinely altering how they
actually feel (i.e., they are faking). In surface acting,
frustrated employees may suppress their frustra-
tion and simply smile at an annoying customer,
thus “putting on a mask” without actually changing
their feelings and expressing feigned rather than
genuine emotions (Grandey, 2003).
An important social psychological theoretical
underpinning of deep and surface acting strategies
comes from the concept of emotion regulation
(Coˆte´, 2005; Grandey, 2000; Gross, 1998a). Emotion
regulation—the “process by which individuals in-
fluence which emotions they have, when they have
them, and how they experience and express these
emotions” (Gross, 1998a: 275)—encompasses a
broader set of behaviors, whereas emotional labor
represents a specific type of emotion regulation
(Coˆte´, 2005). Research differentiates between two
kinds of emotion regulation that closely correspond
with the emotional labor strategies of deep and
surface acting: In antecedent-focused emotion reg-
ulation, people modify their perceptions of a situ-
ation through cognitive reappraisal or by drawing
on emotional memories before the emotion is fully
developed (Gross, 1998b), which mirrors Hoch-
schild’s (1983) strategy of deep acting. In response-
focused emotion regulation, people change their
depiction of a given emotion after experiencing that
emotion rather than adjust their perception of the
situation (Gross, 1998a; Totterdell & Holman,
2003), which is similar to surface acting (Grandey,
2000).
Emotion Recognition
Because we focus on emotion detection during
employee-customer interactions, research on peo-
ple’s ability to recognize other people’s emotions is
of pivotal interest. Emotion recognition has been
widely studied in the social psychological litera-
ture and remains perhaps the most reliably vali-
dated dimension of emotional intelligence
(Elfenbein, Marsh, & Ambady, 2002). It is well doc-
umented that people’s ability to understand the
nonverbal behavior of others and detect deception
predicts a range of important work outcomes, in-
cluding job performance (Elfenbein, Foo, White,
Tan, & Aik, 2007; Elfenbein et al., 2002).
People can generally distinguish sincere from de-
ceptive emotional display, though their judgments
are prone to error (Ekman, 2001, 2003; Ekman &
O’Sullivan, 1991; Ekman, O’Sullivan, & Frank,
1999). One difficulty in accurately detecting emo-
tions is that no universally applicable cues can
differentiate deception from truth telling (Ekman &
O’Sullivan, 1991). Research shows some success
with the use of the facial action coding system
(Ekman & Friesen, 1978), which systematically at-
tempts to categorize the physical expression of
emotions. Although using this coding scheme
improves detection accuracy of truth telling and
deception, the system relies on an extremely
time-consuming, frame-by-frame video analysis
performed by trained experts and is thus not appli-
cable to real-time interactions in social life. In the
context of service interactions, research shows that
employees can detect customers’ emotions (Scherer
& Ceschi, 2000), but we are unaware of any research
that examines customers’ accuracy in detecting the
emotions of service employees or its influence on
customers’ evaluations of a service transaction.
THEORETICAL MODEL OF THE EFFECTS OF
EMOTIONAL LABOR STRATEGIES AND
DETECTION ACCURACY ON
CUSTOMER OUTCOMES
Model Overview
We provide the theoretical model for our re-
search in Figure 1. Building on deep and surface
acting as two key emotional labor strategies
(Grandey, 2003; Gross, 1998b; Hochschild, 1983),
we predict that service employees’ use of deep and
surface acting influences the important customer
outcomes of perceived customer orientation and
service quality and that the accuracy with which
customers detect the emotional labor strategies
moderates this relationship. In addition, comparing
high- and moderate-contact services, we predict
that customers’ reactions to emotional labor differ
depending on the type of service offered.
Customer Outcomes of Emotional
Labor Strategies
Managing frontline employees’ emotions has
been recognized as an important facet of maintain-
ing loyal customers (Albrecht & Zemke, 1985;
Schneider & Bowen, 1985), as the experience and
perception of emotional cues during service deliv-
ery strongly influence customers’ evaluations of a
service encounter (Oliver, 1997; Schmit & Allsc-
heid, 1995). We build on Grandey (2000) and pos-
tulate that the emotional labor strategies used by
service employees—that is, deep or surface act-
ing—differentially impact perceived customer ori-
entation and service quality, two well-established
960 OctoberAcademy of Management Journal

customer perceptions that feature established links
with customer loyalty toward a service firm.
Effect of employee emotional labor on per-
ceived customer orientation. Service employees’
customer orientation reflects the extent to which
their behavior during personal interactions with
customers meets the customers’ needs (Hennig-
Thurau, 2004). Employees’ interest in, and ability
to fulfill, customers’ service-related needs repre-
sent the central elements of customer orientation
(Brady & Cronin, 2001; Brown, Mowen, Donavan, &
Licata, 2002). Customers’ perceptions of employ-
ees’ customer orientation is a key element of a
service company’s value creation process, as cus-
tomer orientation strongly influences customers’
perceptions of the quality of the service provided,
as well as behavioral outcomes such as customer
loyalty (Brady & Cronin, 2001; Hennig-Thurau,
2004; Walsh & Beatty, 2007).
With regard to the impact of emotional labor
strategies on perceived customer orientation, we
expect that employee deep acting has a positive
impact. Efforts to engage in authentic positive dis-
plays should signal to customers that employees
are interested in their needs and motivated to meet
them. This perception, which can occur at either a
conscious or an unconscious level (Barsade, 2002),
is fostered by the authenticity of the displayed
emotions, because authenticity signals that the em-
ployees’ displayed interest in the customers is sin-
cere and genuine (Hennig-Thurau et al., 2006).
However, for employee surface acting, we expect a
negative impact on perceived customer orientation,
because faking positive emotions may lead custom-
ers to question whether the employees are truly
interested in their needs and sufficiently motivated
to work hard to satisfy them. In addition, Richards
and Gross (2000) suggested that surface acting may
require employees to invest more cognitive re-
sources, which could impair their cognitive perfor-
mance and affect customers’ evaluations even if the
customers are not aware that the employees are
engaging in surface acting (Richards & Gross, 2000).
Thus:
Hypothesis 1a. Employee deep acting relates
positively to perceived customer orientation.
FIGURE 1
Theoretical Model
Customer
Loyalty
Intentions
Employee
Surface
Acting
Employee
Deep Acting
Perceived
Customer
Orientation
Perceived
Service Quality
Postulated moderating paths
H1a(+)
H3a(+)
H7(+)
H8(+)
H9(+)
Customer
Deep Acting
Detection
Accuracy
H1b(–)
H2a(+)
H2b(–)
Customer
Surface Acting
Detection
Accuracy
H3b(+)
H4b(–)
H4a(–)
H5
Service Type
H6
2009 961Groth, Hennig-Thurau, and Walsh

Hypothesis 1b. Employee surface acting relates
negatively to perceived customer orientation.
Effect of employee emotional labor on service
quality. Service quality, one of the most researched
constructs in service management, refers to a cus-
tomer’s overall impression of the relative superior-
ity of a service (Bitner & Hubbert, 1994). Service
quality bridges frontline employees’ performance
with customers’ loyalty to a service (Heskett, Jones,
Loveman, Sasser, & Schlesinger, 1994; Zeithaml,
Berry, & Parasuraman, 1996).
We argue that emotional labor relates to key di-
mensions of service quality such as reliability
(“employees show a sincere interest”), responsive-
ness (“employees are willing to help you”), and
assurance (“employees instill confidence”) (Para-
suraman, Berry, & Zeithaml, 1991; Parasuraman,
Zeithaml, & Berry, 1988). Specifically, we expect
employee deep acting to have a positive influence
on customers’ service quality perceptions, because
the greater authenticity of such displays should
suggest a sincere interest (i.e., increase service
quality reliability) and result in higher customer
confidence (i.e., increase service quality assur-
ance). Employee authenticity should also stimulate
customers’ beliefs that the employees serving them
are truly willing to help (i.e., increase service qual-
ity responsiveness). In contrast, the lack of au-
thenticity associated with surface acting may
lead customers to question, either consciously or
unconsciously, the employees’ reliability and re-
sponsiveness and should reduce customers’ con-
fidence in the service firm, which in turn de-
creases service quality. Therefore, we propose:
Hypothesis 2a. Employee deep acting relates
positively to perceived service quality.
Hypothesis 2b. Employee surface acting relates
negatively to perceived service quality.
Moderating Role of Customers’ Emotional Labor
Detection Accuracy
We propose that the hypothesized link between
service employees’ emotional labor strategies and
the customers’ evaluation of the service experience
is moderated by the customers’ level of detection
accuracy regarding whether employees are engag-
ing in deep or surface acting. This argument is
based on evidence from social psychological re-
search pertaining to emotion recognition that sug-
gests that employees will be able to conceal their
true emotions from customers in an interactive ser-
vice encounter only to a certain degree and that
customers’ emotion recognition accuracy will dif-
fer across individuals. Specifically, Ekman and col-
leagues demonstrated that though people can gen-
erally monitor and control some aspects of their
behavior according to display requirements, true
feelings sometimes “leak out” through behavioral
channels that are less controllable and often be-
yond the conscious awareness of social actors
(Ekman, 2001; Ekman & Friesen, 1969).
Given that the recognition of employees’ emo-
tional labor strategies varies, we expect the effects
of deep and surface acting on customer outcomes to
be stronger if customers accurately detect the emo-
tional labor strategy used by employees. If employ-
ees strive to display authentic emotions, but their
efforts go unnoticed by the customers, the positive
effects of deep acting should be weaker (Ekman,
2001; Ekman et al., 1999). Similarly, employee sur-
face acting that goes unnoticed by customers may
not have a strong negative impact on customer out-
comes, but we expect the negative effects of surface
acting on customer outcomes to be stronger if sur-
face acting is accurately perceived as such by
customers. Thus:
Hypothesis 3a. The greater customers’ deep
acting detection accuracy, themore strongly pos-
itive the relationship between employee deep
acting and perceived customer orientation.
Hypothesis 3b. The greater customers’ deep
acting detection accuracy, the more strongly
positive the relationship between employee
deep acting and perceived service quality.
Hypothesis 4a. The greater customers’ surface
acting detection accuracy, the more strongly
negative the relationship between employee
surface acting and perceived customer
orientation.
Hypothesis 4b. The greater customers’ surface
acting detection accuracy, the more strongly
negative the relationship between employee
surface acting and perceived service quality.
Moderating Effects of Service Type
Services are highly heterogeneous and diverse in
nature (Cook, Go, & Chung, 1999). For this research,
it is particularly relevant that services differ in re-
gard to their level of employee-customer contact.
Drawing on Bowen (1990), we distinguish between
high-contact services and moderate-contact ser-
vices and expect the type of service offered (high
vs. moderate contact) to moderate the effect of em-
ployees’ emotional labor strategies on customers’
service evaluations. High-contact services involve
extensive contact with employees, a high level of
962 OctoberAcademy of Management Journal

customization, and strong differences between al-
ternative service offerings (e.g., dental services).
Moderate-contact services, in contrast, entail less
contact and a less-important role of employees,
as well as less customization (e.g., dry-cleaning
services).
Specifically, we argue that customers might not
value genuine emotional displays from employees
as much in moderate-contact services as they do in
high-contact services (Grayson, 1998). In situations
where customers do not place importance on gen-
uine display (i.e., moderate contact services), even
if they correctly recognize employees’ surface act-
ing, it may not negatively influence their service
experience (see, for example, Rafaeli, 1989; Sutton
& Rafaeli, 1988). In contrast, in high-contact ser-
vices, employees’ use of deep versus surface acting
may have a more pronounced influence on custom-
ers’ evaluations of service experiences. Thus:
Hypothesis 5. The relationship between em-
ployee deep (surface) acting and perceived
customer orientation is stronger for high-
contact services than for moderate-contact
services.
Hypothesis 6. The relationship between deep
(surface) acting and perceived service quality
is stronger for high-contact services than for
moderate-contact services.
Relationships among Customer Outcomes
Our theoretical model includes relationships
among the customer outcomes of perceived cus-
tomer orientation and service quality and links
them with customer loyalty intentions, which play
a pivotal role in stable, long-term relationships be-
tween service businesses and their customers
(Zeithaml et al., 1996). Service quality perceptions
form during service delivery in such a way that
customer-oriented attitudes and corresponding be-
haviors of frontline personnel have a significant
effect on service quality perceptions (Brady & Cro-
nin, 2001; Crosby, Evans, & Cowles, 1990). Brady
and Cronin (2001) found that employees’ customer
orientation influences service quality, and Hennig-
Thurau (2004) reported a strong effect of perceived
customer orientation on attitudinal customer satis-
faction, a construct closely related to service qual-
ity. Furthermore, empirical evidence indicates that
customer-oriented employee attitudes and behav-
iors drive customers’ intentions to stay loyal to a
service firm (DeWitt & Liu, 2002; Hennig-Thurau,
2004), as does service quality, an established deter-
minant of customer loyalty intentions (Zeithaml et
al., 1996). Therefore:
Hypothesis 7. Perceived customer orientation
relates positively to perceived service quality.
Hypothesis 8. Perceived customer orientation re-
lates positively to customer loyalty intentions.
Hypothesis 9. Perceived service quality relates
positively to customer loyalty intentions.
METHODS
Procedures and Sample
We surveyed dyads of customers and service em-
ployees from a variety of service industries imme-
diately after a customer and a frontline employee
completed a service transaction. Thus, our unit of
analysis is a distinct service interaction rather than
general, retrospective patterns of behavior, which
is the focus of most emotional labor research. Grad-
uate students of a major university distributed the
questionnaires to service customers. Each partici-
pating student received a packet with up to five
pairs of matching customer and employee ques-
tionnaires, as well as a cover letter that explained
the study and provided instructions about data col-
lection. We instructed participating students to use
one pair of questionnaires for themselves and dis-
tribute the remaining pairs using a snowballing
technique, in which they recruited friends or rela-
tives to participate (Zinkhan, Burton, & Wallendorf,
1983). They were to distribute at least three of the
remaining pairs of questionnaires to working
adults. We informed neither participating custom-
ers nor employees about the research topic; the
information provided to all participants suggested
the study was about “satisfaction with services.”
Customers took both the customer survey and the
employee survey with them to their next service
encounter and asked the service employee who
served them immediately after the service transac-
tion whether he or she could fill out the employee
survey. If the employee agreed, the customer simul-
taneously filled out the customer survey.1 In addi-
tion to the survey itself, each employee received a
short letter explaining the study and assuring him
or her of the confidentiality of responses, as well as
an envelope with a unique seal. We instructed the
employees to put completed surveys in the enve-
lopes and seal them, so that they were assured that
the customers would not see their responses. The
1 Although the nature of the research design made it
impossible to determine the exact response rate for ser-
vice employees, information obtained during debriefing
sessions suggested that the 299 dyads obtained represent
approximately a 40 percent response rate.
2009 963Groth, Hennig-Thurau, and Walsh

employees then handed the sealed envelopes back
to the customers, who returned the completed sur-
vey pairs to us (we informed them that breaking the
seal would invalidate a questionnaire). All survey
pairs contained identifying codes so that we could
subsequently identify the employee-customer dy-
ads. To ensure a sufficient variety in services, we
assigned the student customers randomly to one of
two experimental conditions, either high-contact
service or moderate-contact service; the cover letter
listed services of each type and instructed the cus-
tomers to take both surveys with them on their next
visit to a listed service. We took all services listed
in the cover letters directly from Bowen (1990).
To ensure the legitimacy of the collected data, we
incorporated several quality checks. Most impor-
tant, both the customer and the employee survey
asked for the date and time of the focal service
transaction, the name of the business, and the name
of the employee. We informed the students prior to
the study that we would conduct quality checks to
verify the information provided and would only
pay for completed survey pairs containing valid
data. On receiving the completed survey pairs, we
performed random checks by calling service busi-
nesses to verify that the transactions had taken
place. We uncovered no inconsistencies. Next, we
compared the handwriting on all questionnaires to
ensure that no customer had filled out the em-
ployee questionnaire or multiple customer ques-
tionnaires. As a result of these quality checks, we
deemed 14 pairs of questionnaires either to be
questionable or to contain too much missing data
and removed them from further analysis.
The final sample therefore contained 285 employee-
customer dyads. The customers in the final sample
ranged in age from 17 to 63 years, with a mean age
of 26.7 years (s.d.
10.5). Fifty-eight percent of
customers were female. Service employees in the
final sample had a mean age of 27.8 years (s.d.

9.6), over a range from 16 to 66 years. Their average
job tenure was 3.1 years (s.d.
4.9), over a range
from 10 months to 42 years. Sixty-three percent
of employees were female. Finally, 24 percent of
the dyads involved high-contact services, and the
remaining dyads represented moderate-contact
services.
Measures
Employee measures. The employee question-
naire contained reflective, multi-item measures of
the two emotional labor strategies (deep and sur-
face acting), as well as several demographic and
identifying variables (day and time of service inter-
action, business name and type, employee name).
In addition, we assessed positive and negative af-
fectivity as control variables, which enabled us to
rule out alternative explanations for a link between
emotional labor strategies and customer outcomes.
To assess employee deep and surface acting, we
used two three-item measures from Grandey
(2003), originally developed by Brotheridge and
Lee (2003). The focus of these items was the par-
ticular service interaction an employee had just
completed with a customer, not the employee’s
preferred emotional labor strategy in general. Spe-
cifically, all items were preceded by the stem,
“During today’s interaction with the customer who
handed you the questionnaire . . .,” and used a
response scale ranging from 1, “strongly disagree,”
to 7, “strongly agree.” All construct-measuring
items from our model appear in the Appendix. The
control variables of positive and negative affectiv-
ity were assessed with the ten-item Positive Affect
Negative Affect Scale (PANAS; Watson, Clark, &
Tellegen, 1988). This scale assesses trait affectivity,
asking respondents to report how they “feel on
average” on a scale ranging from 1, “not at all” to 5,
“extremely.”
Customer measures. The customer question-
naire contained reflective, multi-item measures of
perceived customer orientation, service quality,
customer loyalty intentions, and perceptions of em-
ployees’ emotional labor strategies, as well as de-
mographic and identifying variables. We measured
perceived customer orientation with a six-item
scale adapted from Brown et al. (2002) but re-
worded the items slightly to capture the customer
perspective (e.g., “I try to help customers achieve
their goals” was changed to “The employee tried to
help me achieve my goals”). One item was dropped
because of poor factor loadings. To measure per-
ceived service quality, we used a two-item scale
assessing overall service quality developed by
Brady and Cronin (2001). We measured customer
loyalty intentions with four items, three taken from
Zeithaml et al. (1996) and one from Taylor and
Baker (1994). We report all items in the Appendix.
Detection accuracy. To assess emotional labor
detection accuracy, it was necessary to assess cus-
tomers’ perceptions of employees’ emotional labor
strategies. We did so by using the same items from
the employee questionnaire, adapted to reflect the
customer perspective (e.g., “I just pretended to
have the emotions I needed to display to this cus-
tomer” was changed to “I believe the employee just
pretended to have the emotions he/she needed to
display to me”). Pretesting indicated that one sur-
face acting item provoked some misunderstanding
among customers, so we paraphrased that item. As
in the employee questionnaire, the stem of these
964 OctoberAcademy of Management Journal

questions focused on the interaction a customer
had just completed with a service employee. The
response scale was the same as it was for the em-
ployee questionnaire (1
“strongly disagree” to
7
“strongly agree”).
We then created the emotional labor detection ac-
curacy measures using Lance’s (1988) residual cen-
tering regression approach. The approach involves a
two-step procedure in which an interaction term is
first regressed on its two components via ordinary
least squares and then the residuals of this regression
are used instead of the respective interaction term in
tests of the structural model. We chose residual cen-
tering over alternatives such as mean centering be-
cause the former minimizes multicollinearity that
might result from the usual high correlations of re-
gression variables with their product terms and also
provides a “straightforward means to assess the pre-
dictability of some criterion from the interaction
among predictors” (Lance, 1988: 166; cf. Bottomley &
Holden, 2001). Specifically, we calculated the differ-
ence between an employee’s deep (surface) acting
score and his/her customer’s perception of the em-
ployee’s deep (surface) acting. For both deep and
surface acting, we used the mean of the individual
items to calculate the difference. As we were primar-
ily interested in dichotomous categories (i.e., hits vs.
misses), we then applied a median split, assigning a
value of 1 to cases with a difference score higher than
the median (i.e., low accuracy) and a value of 2 to
cases with a difference score lower than the median
(i.e., high accuracy). Mersman and Donaldson (2000)
also recommended this approach to mitigate the reli-
ability concerns associated with difference scores.
Next, we estimated the cross-product residuals of a
regression of deep (surface) acting detection accuracy
and employee deep (surface) acting on the cross-
product of the two variables (i.e., deep (surface) act-
ing detection accuracy and employee deep (surface)
acting). We then used cross-product residuals of these
regressions as the deep (surface) acting accuracy by
employee deep (surface) acting interaction variables.
Service type. We assigned moderate-contact ser-
vice a value of 1 and high-contact service a value of
2. We again used residual centering to generate
interaction term variables, running regressions
with the cross-product of service type and em-
ployee deep (surface) acting as a dependent vari-
able and service type and employee deep (surface)
acting as regressors. We use the residual terms of
these regressions as service type by employee deep
(surface) acting interaction variables when estimat-
ing the structural model.
RESULTS
Reliability, Validity, and Common Method Bias
The means, standard deviations, reliability esti-
mates, and correlation coefficients of all variables
appear in Table 1. The reliability of all reflective
scales is satisfactory, with
scores ranging from .84 to
TABLE 1
Means, Standard Deviations, and Correlationsa
Variables Mean s.d. 1 2 3 4 5 6 7 8 9 10 11
1. Employee positive affectivity 3.76 0.60 (.85)
2. Employee negative affectivity 2.05 0.67 .16** (.87)
3. Employee deep acting 3.76 1.43 .16** .01 (.84)
4. Employee surface acting 2.88 1.58 .23** .24** .06 (.88)
5. Customer perceptions of
employee deep acting
4.21 1.41 .18** .00 .20** .02 (.86)
6. Customer perceptions of
employee surface acting
2.86 1.29 .05 .12* .04 .21** .19** (.89)
7. Customer deep acting detection
accuracy
.12* .02 .31** .03 .00 .06
8. Customer surface acting detection
accuracy
.05 .15* .02 .23** .01 .19** .03
9. Perceived customer orientation 5.35 1.08 .18** .12* .13* .05 .42** .22** .03 .04 (.85)
10. Perceived service quality 5.22 1.27 .19** .14* .17** .03 .41** .10 .03 .04 .54** (.88)
11. Customer loyalty intentions 5.00 1.29 .04 .10 .08 .01 .18** .24** .09 .08 .42** .63** (.91)
12. Service typeb .08 .03 .15* .03 .16* .08 .08 .08 .09 .14* .05
a n
285. Values in parentheses on the diagonal are internal consistency estimates. Absent means and standard deviations indicate the
variable in the row is binary.
b 1
“moderate-contact services,” 2
“high-contact services.”
* p  .05
** p  .01
Two-tailed tests.
2009 965Groth, Hennig-Thurau, and Walsh

.91. To assess convergent and discriminant validity of
all emotional labor strategy and customer outcome
measures, we subjected the measurement models of
all multi-item scales for both the employee and cus-
tomer variables to confirmatory factor analyses (CFA)
using AMOS 6.0. For the employee measures, we
estimated a two-factor model with deep acting and
surface acting as separate constructs. The overall fit
statistics for the two-factor model indicate a good fit
to the data (2[8, n
285]
12.40, p
.13; compar-
ative fit index [CFI]
.99; incremental fit index
[IFI]
1.00; Tucker-Lewis index [TLI]
.99; and
root-mean-square error of approximation [RMSEA]

.06). The fit of the two-factor structure was signifi-
cantly better than that of a one-factor structure
(2[1]
386.3, p .01). For the customer measures,
we estimated a five-factor model (with customer per-
ceptions of employee deep acting and surface acting
and the three customer outcomes: customer orienta-
tion, service quality, and customer loyalty intentions)
that also provided a good fit to the data (2[209, n

285]
304.22, p .01; CFI
.94; IFI
.94; TLI
.92;
RMSEA
.08). The five-factor model provided a
better fit than either a four-factor model in which we
combined deep and surface acting into one factor)
(2[4]
473.04, p  .01) or a one-factor model
(2[10]
1,455.83, p  .01). We report the factor
loadings for all items in the Appendix.
To further assess the discriminant validity of our
measures, we followed the procedures outlined by
Fornell and Larcker (1981), which require that the
average variance extracted for two constructs ex-
ceed the squared correlation between the con-
structs to demonstrate discriminant validity. Re-
sults confirmed that all our study constructs have
sufficient discriminant validity.
Finally, although our data come from three differ-
ent sources (employees, customers, and coders, in the
case of service type), common method bias might still
influence some postulated relations in our model,
such as the links among the customer variables and
those between customer perceptions of emotional la-
bor strategies and customer outcomes. To rule out the
existence of such a bias, we used methods recom-
mended by Podsakoff, MacKenzie, Lee, and Podsa-
koff (2003). Specifically, we used structural equation
modeling to estimate a variation of the model that
includes only those variables collected from custom-
ers, as well as a latent common method variance
factor on which every item in the model was allowed
to load (in addition to its loading on its respective
construct). We compared the significance of all theo-
rized model paths between the models with and
without the additional factor and found no differ-
ences, which indicates the absence of common
method bias (Podsakoff et al., 2003).
Hypothesis Testing
We tested our theoretical model with PLS struc-
tural equation modeling, a distribution-free method
with fewer constraints and statistical specifications
than covariance-based techniques such as LISREL
(Fornell & Bookstein, 1982). We used SmartPLS (Ver-
sion 2.3, Ringle, Wende, & Will, 2005) and estimated
the inner weightings with the path method (Chin,
2001). The t-values were generated through a boot-
strapping procedure with 500 resamples with 285
cases each (Fornell & Bookstein, 1982). The structural
model contains the two employee emotional labor
strategies, the two employee deep (surface) acting by
customer detection accuracy variables, the two em-
ployee deep (surface) acting by service type variables,
and the three customer outcome variables of per-
ceived customer orientation, perceived service qual-
ity, and customer loyalty intentions. In addition to
the theoretically proposed model paths, we also in-
cluded the main effects from the emotional labor de-
tection accuracy variables, service type, and paths
from negative and positive affectivity to the customer
outcomes variables of perceived customer orientation
and service quality as controls. We also tested a base-
line model that includes only the control variables of
positive and negative affectivity; it allowed us to iso-
late the incremental variance explanation for the cus-
tomer outcomes that can be attributed to the model
constructs. In the baseline model, positive and nega-
tive affectivity were linked to each of the three cus-
tomer outcomes. The PLS results for both the theoret-
ical and the baseline model appear in Table 2; Figure
2 highlights the significant model paths.
The composite reliability of the theoretical
model is .88 or greater for all model constructs, and
the average variance extracted (AVE) is greater than
.75 except for perceived customer orientation
(AVE
.64). The results support several, but not
all, of our hypotheses regarding the inner model
relationships. Specifically, we find a significant
link between employee deep acting and perceived
customer orientation (
.11, p  .05) in support
of Hypothesis 1a, but the path between employee
surface acting and perceived customer orientation
is not significant, providing no support for Hypoth-
esis 1b. The direct effect of employee deep acting
on service quality is nearly as strong as the one
between deep acting and perceived customer ori-
entation (
.10), though only significant at the .10
level. However, when examining the total effect
that also accounts for the indirect effect through
perceived customer orientation, results are in line
with the proposed effect of employee deep acting
on service quality (
.15, p  .01). The limited
direct effect provides support for Hypothesis 2a,
966 OctoberAcademy of Management Journal

again only at the .10 level. Both the direct and the
total effect of employee surface acting on service
quality fail to reach significance, so Hypothesis 2b
does not receive support. We also found that em-
ployee deep acting exerts a significant total effect
on customer loyalty intentions (
.11, p  .01), a
finding not formally hypothesized.
Regarding the effects of customers’ emotional labor
detection accuracy on service outcomes, we find that
the interaction of employee deep acting by deep act-
ing detection accuracy has the proposed positive ef-
fect on perceived customer orientation (
.11, p 
.05) in support of Hypothesis 3a. In other words, if
customers accurately detect the extent of employee
deep acting, the deep acting strategy has a stronger
positive effect on perceived customer orientation. Al-
though the coefficient on the path from this interac-
tion to service quality is somewhat weaker (
.08)
and marginally significant (p .10), the total effect of
this relationship is significant (
.13, p  .05).
Again, as a result of the limited direct effect, Hypoth-
esis 3b is supported at only the .10 level. We also find
a significant and substantial negative direct path from
the interaction of employee surface acting and surface
TABLE 2
Path Coefficients from Partial Least Squares Analysesa
Hypothesis Path from To
Theoretical
Model
Baseline
Model
Path
Coefficient
(t)
Path
Coefficient
(t)
1a Employee deep acting Perceived customer orientation .11* (2.06)
1b Employee surface acting Perceived customer orientation .02 (0.35)
2a Employee deep acting Perceived service quality .10 (1.78)
2b Employee surface acting Perceived service quality .03 (0.65)
3a Employee deep acting  customer deep
acting detection accuracy
Perceived customer orientation .11* (1.99)
3b Employee deep acting  customer deep
acting detection accuracy
Perceived service quality .08 (1.68)
4a Employee surface acting  customer
surface acting detection accuracy
Perceived customer orientation .16* (2.85)
4b Employee surface acting  customer
surface acting detection accuracy
Perceived service quality .08 (1.55)
5 Service type  employee deep acting Perceived customer orientation .01 (0.16)
5 Service type  employee surface acting Perceived customer orientation .01 (0.27)
6 Service type  employee deep acting Perceived service quality .08 (1.40)
6 Service type  employee surface acting Perceived service quality .00 (0.11)
7 Perceived customer orientation Perceived service quality .50* (10.56)
8 Perceived customer orientation Customer loyalty intentions .09 (1.70)
9 Perceived service quality Customer loyalty intentions .63* (14.08)
Negative affectivity Perceived customer orientation .15* (2.93) .15* (2.88)
Positive affectivity Perceived customer orientation .13* (2.17) .16* (2.71)
Negative affectivity Perceived service quality .04 (0.98) .05 (1.13)
Positive affectivity Perceived service quality .07 (1.41) .10 (1.94)
Customer deep acting detection accuracy Perceived customer orientation .10 (1.71)
Customer surface acting detection accuracy Perceived customer orientation .01 (0.40)
Service type Perceived customer orientation .07 (1.43)
Customer deep acting detection accuracy Perceived service quality .01 (0.33)
Customer surface acting detection accuracy Perceived service quality .02 (0.57)
Service type Perceived service quality .06 (1.41)
R2, perceived customer
orientation
.12 .05
R2, perceived service
quality
.34 .30
R2, customer loyalty
intentions
.47 .47
a Values of t were calculated through bootstrapping with 500 resamples with 285 cases per sample.
* p  .05
2009 967Groth, Hennig-Thurau, and Walsh

acting detection accuracy interaction to perceived
customer orientation (
.16, p .01), which sup-
ports Hypothesis 4a. The negative direction of this
path implies that though employee surface acting
does not have a direct main effect on perceived cus-
tomer orientation, the effect becomes significant in
cases in which a customer accurately detects an em-
ployee’s surface acting strategy, so that surface acting
leads to reduced levels of perceived customer orien-
tation in those circumstances. No such effect (either
direct or total) emerges for service quality, providing
no support for Hypothesis 4b.
Given the relevance of detection accuracy for per-
ceived customer orientation (for both deep and sur-
face acting), it is interesting to note that employee
deep acting and customers’ perceptions of employee
deep acting correlate significantly (r
.20, p  .01),
as is the case for the correlation between employee
surface acting and customers’ perceptions of em-
ployee surface acting (r
.21, p  .01). Thus, cus-
tomers are able to detect employees’ emotional labor
strategies, although the relatively low correlation co-
efficients suggest that such ability is far from perfect.
This inference is supported by the means of the un-
adjusted difference between employee and customer-
perceived deep acting and the one between employee
and customer-perceived surface acting, which are
1.45 and 1.42, respectively, with both measures being
significantly different from 0 (p  .01). Also, the
standard deviations of both difference measures sug-
gest that the accuracy of detecting employee emotions
differs substantially among customers (deep acting
1.17;
surface acting
1.14).
Counter to our expectations, no support emerged
for the proposed moderating effect of service type.
Specifically, we find that the service interaction
term variables have no significant impact on cus-
tomer orientation and service quality; the main ef-
fects of service type are also nonsignificant. Thus,
Hypotheses 5 and 6 do not receive support.
Finally, the relations among the service outcomes
are largely as proposed, showing that perceived cus-
tomer orientation has a strong direct effect on service
quality (
.50, p  .01) and that service quality is
strongly linked to customer loyalty intentions (

.63, p  .01), in support of Hypotheses 7 and 9.
FIGURE 2
Significant Model Paths
Customer
Loyalty
Intentions
Employee
Surface
Acting
Employee
Deep Acting
Perceived
Customer
Orientation
Perceived
Service Quality
H1a(+)
H3a(+)
H7(+)
H8(+)
H9(+)
H2a(+)
H3b(+)
H4a(–)
Path significant at p < .05
Path significant at p < .10
Customer
Deep Acting
Detection Accuracy
× Employee Deep
Acting
Customer
Surface Acting
Detection Accuracy
× Employee
Surface Acting
Service Type
968 OctoberAcademy of Management Journal

Counter to our expectations, the direct effect of per-
ceived customer orientation on customer loyalty in-
tentions (Hypothesis 8) fails to reach significance at
the .05 level, but it is marginally significant (p 

.10). The total effect of perceived customer orienta-
tion on loyalty intentions is quite strong (
.41, p
.01), which suggests that perceived customer orienta-
tion influences customer loyalty intentions mostly
through service quality. This finding is in line with
that of Brady and Cronin (2001), who demonstrated
an indirect impact of customer orientation on loyalty,
and Hennig-Thurau’s (2004) finding of a direct im-
pact in only one of two service industries.
Variance Explanation
Regarding variance explanation, we find that the
theoretical model that contains employees’ emotional
labor strategies and detection accuracy variables, as
well as the service type moderator, explains 11.5 per-
cent of perceived customer orientation. The baseline
model (which includes only the control variables of
positive and negative affectivity) explains 5.4 percent
of the variance in the same outcome variable, so we
conclude that the difference of 6.1 percent in variance
explanation can be attributed to employees’ emo-
tional labor, customers’ emotional labor detection ac-
curacy, and service type.
As we were mainly interested in employees’ emo-
tional labor and customers’ emotional labor detection
accuracy, we computed an additional structural
model that includes these variables but not service
type. This model explains 5.6 percent more variance
than the baseline model, an increase that can be at-
tributed solely to emotional labor and detection ac-
curacy. Because PLS does not offer formal signifi-
cance tests between different models, we conducted a
blockwise OLS regression analysis with the variables
included in the baseline model and the emotional
labor and detection accuracy variables and found that
this increase in explained variance was significant
(p .01). The effects of emotional labor and detection
accuracy on service quality could not be examined in
the same straightforward manner because the theoret-
ical model implies a path from perceived customer
orientation to service quality in all models. To isolate
the increase in explained variance in service quality
caused by emotional labor and detection accuracy,
we ran the structural model (without the service type
variables) with PLS but deleted the path from cus-
tomer orientation to service quality. In doing so, we
found that explained variance (measured as the R2 of
service quality) rose from 5.5 percent (baseline
model) to 10.2 percent, an increase of 4.7 percent that
can be attributed to emotional labor and detection
accuracy. Again, using blockwise OLS regression, we
found the increase was significant (p  .01).
DISCUSSION
Summary of Results and Theoretical
Implications
With this research, we attempt to examine the re-
lationship between service employees’ emotional la-
bor and customers’ resulting service experiences, an
important yet underresearched facet of service man-
agement. We developed and empirically tested a the-
oretical model of the differential effects of the emo-
tional labor strategies of deep and surface acting and
customers’ ability to detect these strategies accurately
on customer outcomes in real-world service interac-
tions. We did so by collecting survey data from 285
dyads of employees and customers immediately fol-
lowing a service encounter. By focusing on the ser-
vice encounter as the unit of analysis and collecting
the data immediately after the service transaction, we
contribute to emotional labor research in that we ex-
amine immediate emotional labor behaviors as they
are experienced by employees and customers rather
than examine them on average, as is predominantly
done in research.
Our results demonstrate that service employees’
internal regulatory emotional labor strategies differ-
entially influence customer outcomes and that cus-
tomers’ ability to judge the employees’ strategies ac-
curately moderates these impacts. We find that deep
acting provides positive benefits for customers, a re-
sult that is in line with research that shows the pos-
itive benefits of deep acting for workers (Grandey,
2003). Deep acting therefore emerges as an important
driver of service delivery outcomes such as perceived
customer orientation and service quality. Surface act-
ing does not exert the same positive effect, but we do
not find a negative main effect on customers either.
As another key contribution, this research sheds light
on the crucial role of customers’ accuracy in detecting
employees’ strategies; results show that surface acting
exerts negative effects when customers perceive it as
such. Put differently, surface acting is not a problem
as long as customers do not recognize it. The crucial
role of emotion detection in turn becomes even more
obvious through our finding that detection accuracy
also increases the positive impact of deep acting on
customer outcomes. As we do not find support for a
hypothesized moderating role of service type, it
seems that the impact of emotional labor strategies on
the customer experience that we do find is not spe-
cific to a certain type of service, but tends to exist
regardless of the kind of service offered.
To what extent are customers able to detect emo-
2009 969Groth, Hennig-Thurau, and Walsh

tional labor accurately in the first place? Preliminary
insights can be taken from our emotional labor detec-
tion accuracy measures and the relationships be-
tween employee and customer-perceived emotional
labor strategies. The significant correlations between
employee emotional labor and customer perceptions
of that labor suggest that customers can indeed “de-
code” employees’ emotional labor strategies, though
the limited strength of the correlation coefficients and
the relatively high average differences between the
employees’ behaviors and the customers’ perceptions
of those behaviors also indicate that this decoding
ability is far from perfect, as has been theoretically
argued (Ekman, 2001; Ekman et al., 1999). It appears
that no matter which emotional labor strategy em-
ployees choose, their true emotions often leak out
(Ekman, 2001, 2003) to be detected by customers. But
this detection process is error-prone, in keeping with
social psychological research that indicates people
can generally detect deceptive emotions, although
their accuracy is only slightly better than chance
(Ekman & O’Sullivan, 1991; Ekman et al., 1999). It
should be noted that this finding contradicts emo-
tional labor theories that predict deep and surface
acting differ only in terms of internal cognitive regu-
latory processes and not in terms of emotional dis-
plays discernible to other people.
Implications for Service Managers
The provision of high-quality customer service
has long been considered a competitive advantage
in service industries, and the delivery of “service
with a smile” has received increasing attention in
an effort to satisfy customers and increase their
loyalty to service firms (Schneider, 1994; Schnei-
der, Ashworth, Higgs, & Carr, 1996). Service em-
ployees, as the face of a service firm, must create
this image through their own emotional displays
(Rupp, Holub, & Grandey, 2007). Service managers
should note that the emotional displays of frontline
service workers play an important role in driving
customer experiences and thus in customers’ stay-
ing loyal to a service provider. Specifically, pursu-
ing an “always smile” customer service strategy
may not be the most effective means of improving
customers’ experiences. Such a strategy may have
limited benefit or even a detrimental effect on cus-
tomers’ service experiences and eventually their
loyalty. Prior research suggests that faking emo-
tions may increase staff turnover (Coˆte´ & Morgan,
2002); our research suggests that surface acting may
be less effective than deep acting in eliciting de-
sired customer responses. Although organizational
efforts to increase employee smiling are often well-
intentioned, their effectiveness for both employee
and customer outcomes demands careful scrutiny.
Service managers should encourage deep acting
strategies by employees. For example, during hir-
ing, managers might focus on individual differ-
ences that indicate that some people are more ef-
fective at, and more likely to engage in, deep acting
(Gosserand & Diefendorff, 2006). Training might
also address this issue by suggesting ways to engage
effectively in deep acting and thus maximize the
chances that customers detect genuine emotional
displays. One such approach, perspective taking
(Parker & Axtell, 2001), uses empathy training and
asks employees to put themselves in the shoes of
their customers and thereby view the world
through their eyes (Wharton, 1993).
Our finding that customers’ often inaccurate per-
ceptions of emotional labor influence customer
outcomes has important implications for managers.
Service managers would be well advised to manage
both employees’ behavior and customers’ subjec-
tive experience of service. For example, the latter
could be achieved by more actively managing as-
pects of the “servicescape” (Bittner, 1992)—the
physical aspects of the service environment—or
the appearance and aesthetic quality of service em-
ployees (Nickson, Warhurst, Witz, & Cullen, 2001)
to influence customers’ subjective perceptions of
employees’ emotional display.
Limitations and Future Research Directions
Although this research is the first study to use
field data from a variety of services to assess the
relationships between emotional labor strategies
and customer outcomes, our empirical design did
not allow us to control the dyads for situational
influences, such as the dynamics of an employee-
customer interaction, service environment (e.g.,
aesthetics, music), and distractions by colleagues
or other customers. Future studies on the topic
might control for such factors to explain why some
customers are better able to read employees’ emo-
tional labor strategies.
Even though our sample covers a variety of gender
and age groups, it is not random, which limits the
generalizability of the results. An interesting question
that arises from our findings is whether other vari-
ables moderate the relationship between employee
deep and surface acting and customer outcomes. For
example, Parker and Axtell (2001) reported that in-
teraction frequency correlated with employee and
customer emotions. It would be informative to inves-
tigate how variables such as frequency of service use
or the longevity of a relationship moderate the emo-
tional labor–customer outcomes link.
970 OctoberAcademy of Management Journal

A related question is to whether customers’ expec-
tations of emotional labor vary depending on the type
of service relationship in which they are involved.
Gutek and colleagues (Gutek, Bhappu, Liao-Troth, &
Cherry, 1999; Gutek, Groth, & Cherry, 2002) concep-
tualized service relationships by distinguishing be-
tween interactions of two people who have a shared
history of interactions and expect to interact again in
the future (i.e., service relationships) and interactions
of people who do not know or expect to see each
other again (i.e., service encounters). It would be in-
teresting to investigate whether customer expecta-
tions of, and reactions to, emotional labor differ sub-
stantially between these two service delivery types.
Additional research might benefit from exploring
the potential moderating effects of such service
characteristics.
Finally, though the survey items we used in this
research to measure emotional labor have been val-
idated, they do not separate out the effects of the
different facets of emotional display (e.g., smiling,
eye contact, body language) on customers. Further
research could employ new and multifaceted mea-
surement approaches.
REFERENCES
Albrecht, K., & Zemke, R. 1985. Service America! Doing
business in the new economy. Homewood, IL:
Dow Jones-Irwin.
Barsade, S. G. 2002. The ripple effect: Emotional conta-
gion and its influence on group behavior. Adminis-
trative Science Quarterly, 47: 644–675.
Bitner, M. J. 1990. Evaluating service encounters: The
effects of physical surroundings and employee re-
sponses. Journal of Marketing, 54(2): 69–82.
Bitner, M. J. 1992. Servicescapes: The impact of physical
surroundings on customers and employees. Journal
of Marketing, 56(2): 57–71.
Bitner, M. J., & Hubbert, A. R. 1994. Encounter satisfac-
tion versus overall satisfaction versus quality. In
R. T. Rust & R. L. Oliver (Eds.), Service quality: New
directions in theory and practice: 72–94. Thousand
Oaks, CA: Sage.
Bottomley, P. A., & Holden, S. J. S. 2001. Do we really
know how consumers evaluate brand extensions?
Empirical generalizations based on secondary anal-
ysis of eight studies. Journal of Marketing Re-
search, 38: 494–500.
Bowen, J. 1990. Development of a taxonomy of services
to gain strategic marketing insights. Journal of the
Academy of Marketing Science, 18: 43–49.
Brady, M. K., & Cronin, J., Jr. 2001. Some new thoughts
on conceptualizing perceived service quality: A
hierarchical approach. Journal of Marketing, 65:
34–49.
Brotheridge, C. M., & Grandey, A. A. 2002. Emotional
labor and burnout: Comparing two perspectives of
“people work.” Journal of Vocational Behavior, 60:
17–39.
Brotheridge, C. M., & Lee, R. T. 2003. Development and
validation of the emotional labour scale. Journal of
Occupational and Organizational Psychology, 76:
365–379.
Brown, T. J., Mowen, J. C., Donavan, D. T., & Licata, J. W.
2002. The customer orientation of service workers:
Personality trait effects on self- and supervisor per-
formance ratings. Journal of Marketing Research,
19: 110–119.
Chin, W. W. 2001. PLS-Graph user’s guide: Version 3.0.
Soft Modeling Inc.
Cook. D. P., Goh, C.-H., & Chung, C. H. 1999. Service
typologies: A state of the art survey. Production and
Operations Management, 8: 318–338.
Coˆte´, S. 2005. A social interaction model of the effects of
emotion regulation on work strain. Academy of
Management Review, 30: 509–530.
Coˆte´, S., & Morgan, L. M. 2002. A longitudinal analysis of
the association between emotion regulation, job sat-
isfaction, and intentions to quit. Journal of Organi-
zational Behavior, 23: 947–962.
Crosby, L. A., Evans, K. R., & Cowles, D. 1990. Relation-
ship quality in services selling: An interpersonal
influence perspective. Journal of Marketing, 54(3):
63–81.
DeWitt, T., & Liu, Y. 2002. The customer orientation-
loyalty model: The role of emotional contagion and
rapport in the service encounter. In W. J. Kehoe &
J. H. Lindgren, Jr. (Eds.), AMA summer educators’
conference: Enhancing knowledge development in
marketing: 321. Chicago: AMA.
Ekman, P. 2001. Telling lies: Clues to deceit in the
marketplace, politics, and marriage. New York:
Norton.
Ekman, P. 2003. Emotions revealed: Recognizing faces
and feelings to improve communication and emo-
tional life. New York: Times Books.
Ekman, P., & Friesen, W. V. 1969. The repertoire of non-
verbal behavior: Categories, origins, usage, and cod-
ing. Semiotica, 1: 49–98.
Ekman, P., & Friesen, W. V. 1978. The facial action
coding system. Palo Alto, CA: Consulting Psycholo-
gists Press.
Ekman, P., & O’Sullivan, M. 1991. Who can catch a liar?
American Psychologist, 46: 913–920.
Ekman, P., O’Sullivan, M., & Frank, M. G. 1999. A few
can catch a liar. Psychological Science, 10:
263–266.
Elfenbein, H. A., Foo, M., White, J. B., Tan, H. H., & Aik,
V. C. 2007. Reading your counterpart: The benefit of
emotion recognition accuracy for effectiveness in
2009 971Groth, Hennig-Thurau, and Walsh

negotiation. Journal of Nonverbal Behavior, 31:
205–223.
Elfenbein, H. A., Marsh, A. A., & Ambady, N. 2002.
Emotional intelligence and the recognition of emo-
tion from facial expressions. In L. Feldman Barrett &
P. Salovey (Eds.), The wisdom in feeling: Psycho-
logical processes in emotional intelligence: 37–59.
New York: Guilford Press.
Fornell, C., & Bookstein, F. L. 1982. Two structural equa-
tion models: LISREL and PLS applied to consumer
exit-voice theory. Journal of Marketing Research,
19: 440–452.
Fornell, C., & Larcker, D. F. 1981. Evaluating structural
equation models with unobservable variables and
measurement error. Journal of Marketing Research,
18: 39–50.
Gosserand, R. H., & Diefendorff, J. M. 2006. Emotional
display rules and emotional labor: The moderating
role of commitment. Journal of Applied Psychol-
ogy, 90: 1256–1264.
Grandey, A. A. 2000. Emotion regulation in the work-
place: A new way to conceptualize emotional labor.
Journal of Occupational Health Psychology, 1:
95–110.
Grandey, A. A. 2003. When “the show must go on”:
Surface acting and deep acting as determinants of
emotional exhaustion and peer-rated service deliv-
ery. Academy of Management Journal, 46: 86–96.
Grandey, A. A. 2008. Emotions at work: A review and
research agenda. In C. Cooper & J. Barling (Eds.),
Handbook of organizational behavior: 235–261.
London: Sage.
Grayson, K. 1998. Customer responses to emotional labor
in discrete and relational service exchanges. Inter-
national Journal of Service Industry Management,
9: 126–154.
Gross, J. J. 1998a. The emerging field of emotion regula-
tion: An integrative review. Review of General Psy-
chology, 2: 271–299.
Gross, J. J. 1998b. Antecedent- and response-focused
emotion regulation: Divergent consequences for ex-
perience, expression, and physiology. Journal of
Personality and Social Psychology, 74: 224–237.
Gutek, B. A., Bhappu, A. D., Liao-Troth, M. A., & Cherry,
B. 1999. Distinguishing between service relation-
ships and encounters. Journal of Applied Psychol-
ogy, 84: 218–233.
Gutek, B. A., Groth, M., & Cherry, B. 2002. Achieving
service success through relationships and enhanced
encounters. Academy of Management Executive,
16(4): 132–144.
Hatfield, E., Cacioppo, J. T., & Rapson, R. L. 1992. Prim-
itive emotional contagion. In M. S. Clark (Ed.), Re-
view of personality and social psychology: Emo-
tion and social behavior: 151–177. Newbury Park,
CA: Sage.
Hatfield, E., Cacioppo, J. T., & Rapson, R. L. 1994. Emo-
tional contagion. Cambridge, U.K.: Cambridge Uni-
versity Press.
Hennig-Thurau, T. 2004. Customer orientation of service
employees: Its impact on customer satisfaction, com-
mitment, and retention. International Journal of
Service Industry Management, 15: 460–478.
Hennig-Thurau, T., Groth, M., Paul, M., & Gremler, D.
2006. Are all smiles created equal? How emotional
contagion and emotional labor affect service rela-
tionships. Journal of Marketing, 70(3): 58–73.
Heskett, J. L., Jones, T. O., Loveman, G. W., Sasser, Jr.,
W. E., & Schlesinger, L. 1994. Putting the service-
profit chain to work. Harvard Business Review,
72(2): 164–174.
Hochschild, A. R. 1979. Emotion in work, feeling rules
and social structure. American Journal of Sociol-
ogy, 85: 551–575.
Hochschild, A. R. 1983. The managed heart: Commer-
cialization of human feeling. Berkeley: University
of California Press.
Kruml, S. M., & Geddes, D. 2000. Exploring the dimen-
sions of emotional labor: The heart of Hochschild’s
work. Management Communication Quarterly, 14:
8–49.
Lance, C. E. 1988. Residual centering, exploratory and
confirmatory moderator analysis, and decomposi-
tion of effects in path models containing interac-
tions. Applied Psychological Measurement, 12:
163–175.
Mattila, A. S., & Enz, C. A. 2002. The role of emotions in
service encounters. Journal of Service Research, 4:
268–277.
Mersman, J. L., & Donaldson, S. I. 2000. Factors affecting
the convergence of self-peer ratings on contextual
and task performance. Human Performance, 13:
299–322.
Morris, J. A., & Feldman, D. C. 1997. Managing emotions
in the workplace. Journal of Managerial Issues, 9:
257–274.
Nickson, D., Warhurst, C., Witz, A., & Cullen, A.-M. 2001.
The importance of being aesthetic: Work, employ-
ment and service organisations. In A. Sturdy, I.
Grugulis, & H. Willmott (Eds.), Customer service:
Empowerment and entrapment: 170–190. New
York: Palgrave.
Oliver, R. L. 1997. Satisfaction: A behavioral perspec-
tive on the consumer. New York: McGraw-Hill.
Parasuraman, A., Berry, L. L., & Zeithaml, V. A. 1991.
Refinement and reassessment of the SERVQUAL
scale. Journal of Retailing, 67: 420–450.
Parasuraman, A., Zeithaml, V. A., & Berry, L. L. 1988.
SERVQUAL: A multiple-item scale for measuring
consumer perceptions of service quality. Journal of
Retailing, 64: 12–40.
972 OctoberAcademy of Management Journal

Parker, S. K., & Axtell, C. M. 2001. Seeing another view-
point: Antecedents and outcomes of employee per-
spective taking. Academy of Management Journal,
44: 1085–1100.
Podsakoff, P. M., MacKenzie, S. B., Lee, J.-Y., & Podsa-
koff, N. P. 2003. Common method biases in behav-
ioral research: A critical review of the literature and
recommended remedies. Journal of Applied Psy-
chology, 88: 879–903.
Rafaeli, A. 1989. When cashiers meet customers: An
analysis of the role of supermarket cashiers. Acad-
emy of Management Journal, 32: 245–273.
Rafaeli, A., & Sutton, R. I. 1987. Expression of emotion as
part of the work role. Academy of Management
Review, 12: 23–37.
Richards, J. M., & Gross, J. J. 2000. Emotion regulation
and memory: The cognitive costs of keeping one’s
cool. Journal of Personality and Social Psychology,
79: 410–424.
Ringle, C., Wende, S., & Will, A. 2005. SmartPLS version
2.0 M2. www.smartpls.de.
Rupp, D., Holub, S., & Grandey, A. A. 2007. A cognitive-
emotional theory of customer injustice and emo-
tional labor: Implications for customer service, fair-
ness theory, and the multifoci perspective. In D.
DeCremer (Ed.), Advances in the psychology of jus-
tice and effect: 199–226. Charlotte, NC: IAP.
Scherer, K. R., & Ceschi, G. 2000. Criteria for emotion
recognition from verbal and nonverbal expression:
Studying baggage loss in the airport. Personality
and Social Psychology Bulletin, 26: 327–339.
Schmit, M. J., & Allscheid, S. P. 1995. Employee attitudes
and customer satisfaction: Making theoretical and
empirical connections. Personnel Psychology, 48:
521–537.
Schneider, B. 1994. HRM—A service perspective: To-
wards a customer-focused HRM. International Jour-
nal of Service Industry Management, 5: 64–76.
Schneider, B., Ashworth, S. D., Higgs, A. C., & Carr, L.
1996. Design, validity, and use of strategically fo-
cused employee attitude surveys. Personnel Psy-
chology, 49: 695–705.
Schneider, B., & Bowen, D. E. 1985. Employee and cus-
tomer perceptions of service in banks: Replications
and extension. Journal of Applied Psychology, 70:
423–433.
Sutton, R. I., & Rafaeli, A. 1988. Untangling the relation-
ship between displayed emotions and organizational
sales: The case of convenience stores. Academy of
Management Journal, 31: 461–487.
Tan, H. H., Foo, M. D., & Kwek, M. H. 2004. The effects of
customer personality traits on the display of positive
emotions. Academy of Management Journal, 47:
287–296.
Taylor, S. A., & Baker, T. L. 1994. An assessment of the
relationship between service quality and customer
satisfaction in the formation of consumers’ purchase
intentions. Journal of Retailing, 70: 163–178.
Totterdell, P., & Holman, D. 2003. Emotion regulation in
customer service roles: Testing a model of emotional
labor. Journal of Occupational Health Psychology,
8: 55–73.
Tsai, W.-C. 2001. Determinants and consequences of em-
ployee displayed positive emotions. Journal of
Management, 27: 497–512.
Tsai, W.-C., & Huang, Y.-M. 2002. Mechanisms linking
employee affective delivery and customer behavioral
intentions. Journal of Applied Psychology, 87:
1001–1008.
Walsh, G., & Beatty, S. E. 2007. Customer-based corporate
reputation of a service firm: Scale development and
validation. Journal of the Academy of Marketing
Science, 35: 127–143.
Watson, D., Clark, L. A., & Tellegen, A. 1988. Develop-
ment and validation of brief measures of positive
and negative affect: The PANAS scales. Journal of
Personality and Social Psychology, 54: 1063–1070.
Wharton, A. S. 1993. The affective consequences of ser-
vice work. Work and Occupations, 20: 205–232.
Zapf, D. 2002. Emotion work and psychological well-
being: A review of the literature and some concep-
tual considerations. Human Resource Management
Review, 12: 237–268.
Zeithaml, V. A., Berry, L. L., & Parasuraman, A. 1996.
The behavioral consequences of service quality.
Journal of Marketing, 60(2): 31–46.
Zinkhan, G., Burton, S., & Wallendorf, M. 1983. Market-
ing applications for snowball sampling: Potential
benefits and problems. In W. R. Darden, K. B. Mon-
roe, & W. R. Dillon (Eds.), Research methods and
causal modeling in marketing, vol. 1: 5–8. Chicago:
American Marketing Association.
2009 973Groth, Hennig-Thurau, and Walsh

Markus Groth (markusg@unsw.edu.au) is a senior lecturer
in the School of Organisation and Management at the Uni-
versity of New South Wales. He earned his Ph.D. in man-
agement at the University of Arizona. His research interests
include service management and emotions in the workplace.
Thorsten Hennig-Thurau (tht@medien.uni-weimar.de) is
a professor of marketing and media research at Bauhaus–
University of Weimar and a research professor of market-
ing at Cass Business School, City University London. He
received his doctorate in marketing from the University
of Hannover. His main research interests include custom-
er-employee interfaces in services, customer-firm rela-
tionships, and media management.
Gianfranco Walsh (walsh@uni-koblenz.de) is a professor
of marketing and electronic retailing at the University of
Koblenz–Landau and a visiting professor in marketing at
the University of Strathclyde Business School. He earned
his Ph.D. and his habilitation (postdoctoral degree) from the
University of Hannover. His research focuses on manage-
ment issues and consumer behavior.
APPENDIX
Construct Measures
TABLE A1
Results of Confirmatory Factor Analysisa
Items
Standardized
Coefficient
Composite
Reliability
Employee deep acting .90
I tried to actually experience the emotions I had to show to the customer. .66
I worked hard to feel the emotions that I needed to show to this customer. .90
I made a strong effort to actually feel the emotions that I needed to display toward this
customer.
.85
Employee surface acting .92
I just pretended to have the emotions I needed to display to this customer. .97
I put on a ‘mask’ in order to display the emotions my manager wants me to display. .83
I put on a ‘show’ or ‘performance’ when interacting with this customer. .74
Perceived customer orientation .90
The employee tried to help me achieve my goals. .76
The employee seemed to achieve his/her own goals by satisfying me. .75
The employee got me to talk about my service needs with him/her. .63
The employee kept the best interests of the customer in mind. .78
The employee was able to answer my questions correctly. .76
Perceived service quality .94
I would say that this firm provides superior service. .83
I believe this firm offers excellent service. .94
Customer loyalty intentions .94
I will say positive things about this service provider to other people. .92
I will recommend this service provider to someone who seeks my advice. .92
I will consider this service provider my first choice. .73
I will encourage friends and relatives to do business with this service provider. .85
Customer perceptions of employee deep acting .92
The employee tried to actually experience the emotions s/he had to show to me. .70
The employee worked hard to feel the emotions that s/he needed to show to me. .86
The employee made a strong effort to actually feel the emotions that s/he needed to
display toward me.
.90
Customer perceptions of employee surface acting .93
The employee just pretended to have the emotions s/he displayed to me. .92
The employee put on a ‘mask’ in order to display the emotions his/her boss wants him/
her to display.
.83
The employee showed feelings to me that are different from what s/he actually felt. .79
a n
285. All factor loadings are significant at p  .01.
974 OctoberAcademy of Management Journal