Tải bản đầy đủ
A Reconceptualization of the Emotional Labor Construct: On the Development of an Integrated Theory of Perceived Emotional Dissonance and Emotional Labor

A Reconceptualization of the Emotional Labor Construct: On the Development of an Integrated Theory of Perceived Emotional Dissonance and Emotional Labor

Tải bản đầy đủ

RU B I N ET A L .

190

vated behavior undertaken to manage o r regulate the subjective state. We pro­
pose, as have others (Mann, 1999; Wakefield, 1989), that it is conceptually
necessary to separate emotional states from behavior. Thus, we contend that
separating the state (as experienced) from behavior (broadly defined) will in­
crease the theoretical clarity and ultimate utility of the emotional labor con­
struct.

THEORETICAL COMPONENTS
OF THE EMOTIONAL LABOR CONSTRUCT
Although diverse in nature, the previous conceptualizations of emotional labor
can be roughly collapsed into three categories: (a) those that conceive of emo­
tional labor as an emotional state arising from organizational or social norms
and requirements; (b) those that suggest emotional labor consists of the be­
haviors undertaken to manage an implied or explicit emotional state (Note:
"behavior" implies action and/or reaction, some of which may be unobserv­
able, such as self-talk or cognitive reappraisals) ; and (c) those that consist of
states, behavior, and/or situational factors. Most common, however, are con­
ceptualizations that combine two or more of these categories within a single
construct. For example, Morris and Feldman ( 1 996) defined emotional labor
as consisting of five dimensions, four of which are situational factors (fre­
quency, duration, variety, and intensity of emotional display) and one of which
is an individual state (emotional dissonance) . Others have proposed similar
definitions (Morris & Feldman, 1997; Rafaeli & Sutton, 1 989; Sutton, 199 1 ) .
Recently, Schaubroeck and Jones (2000) commented that the emotional state
of dissonance (similar to Morris & Feldman, 1997) was the most similar in
concept to their construct; however, their definition seems to also include sit­
uational factors ("in order to meet the demands of the job"), emotional states
("the experience of workers"), and behavior ("to suppress negative and ex­
press positive"). Mann ( 1 999) defined the construct as having three dimen­
sions: (a) expectations or rules about emotional display, (b) emotional sup­
pression, and (c) emotional faking. As she implied, the first dimension
represents a situational component and the second and third dimensions rep­
resent behavioral components. She remarked, "The last two dimensions in­
corporate the conflict idea, since emotions can only be suppressed or faked be­
cause of a discrepancy between expected and felt emotion" (p. 354) . Thus,
Mann's (1 999) conception involves a behavioral reaction to a state of conflict.
One of the earliest conceptualizations of emotional labor (in the organiza­
tional literature) by Ashforth and Humphrey ( 1 993) narrowed the scope con­
siderably by defining emotional labor as "the act of displaying appropriate
emotion" (p. 90) They added, ''We prefer to focus on behavior rather than on
the presumed emotions underlying behavior." Further, they explained that
.

11.

P E R C E IVED D I S S O NANCE AND E M O T I O NAL LAB O R

191

their definition "decouples the experience of emotion from the expression of
emotion" (p. 90). Thus, they made a clear distinction of identifying emotional
labor as a behavior versus an emotional state.
Although these formulations of emotional labor are certainly relevant and
highly related to the construct, the collective consequence of these conceptu­
alizations has been theoretical and practical confusion. Alicia Grandey (2000,
2003) developed a comprehensive model of emotional labor, defining emo­
tional labor as an act rather than an emotional state and specifying situational
factors and outcome variables. In particular, Grandey conceptualized emo­
tional labor as the emotion regulation process enacted in response to situa­
tional cues such as service interaction expectations (frequency, duration, vari­
ety, display rules) and emotional events. Emotional regulation is undertaken to
meet organizational demands felt by individuals, and can be achieved through
"surface acting" (regulating expression) and "deep acting" (regulating feel­
ings) (Hochschild, 1983). Thus, Grandey has separated the situational factors,
such as display rules, from other parts of the emotion regulation process. We
modify and extend her conceptualization by specifying the nature of an emo­
tion state-perceived emotional dissonance-and expand the notion of emo­
tional labor as one possible behavioral response to the state of perceived emo­
tional dissonance.

TOWARD A RECONCEPTUALIZATION
OF EMOTIONAL LABOR
We employed Wakefield's ( 1 989) framework of levels of explanation to disen­
tangle the emotional labor construct. His model makes distinctions between
distal explanations of behavior (e.g., primary traits such as the extraversion and
intelligence) and more proximal levels of explanation (e.g., specific psycholog­
ical states such as emotion and cognition) . Pertinent to our discussion is the
distinction between what he termed the intentional state, or what is going on
within the person, and the actual behavior being performed, called the moti­
vated act. Again, motivated acts are broadly defined such that the behaviors
employed are not always visible to others, such as cognitive reappraisal.
Wakefield used the term motivated act to make the distinction between unmo­
tivated behaviors/actions such as automatic responses, and motivated or voli­
tional behaviors/actions. He explained that because of its proximity to behav­
ior, "intentional states form the reasons for actions, and therefore, explain
behavior" (p. 335) . Thus, intentional states are the best predictors of moti­
vated acts within a given context. Current models of emotional labor have con­
fused and/or merged the intentional state experienced by an individual with
the motivated act or behavioral response to that intentional state. Wakefield's

192

RUBIN ET AL.

framework suggests that a given emotional state should influence behavior,
but is necessarily distinct from the behavior.
Certainly we are not the first to argue that confusing the state and behavior
may be theoretically unfavorable. Although we used Wakefield's model to
guide our thinking delineating the emotional state from the motivated act, oth­
ers have made similar distinctions in studying emotion constructs (d. Weiss &
Cropanzano, 1 996). For example, Mann addressed the challenges presented
by current emotional labor theory, stating, "The emphasis [of previous theory]
thus ranges from the internal effort on the one hand, to the external behavioral
display on the other. But which part is 'emotional labor' ? " (p. 348) . To date,
the distinction between states, behaviors, and situational factors remains un­
specified in the emotional labor literature.
Proposed Model
We believe that an additional construct is necessary to conceptually clarify the
literature and provide the critical link between the situational demands and
state conditions, and the outcome, behavior or act. This alternate and inter­
vening construct is needed to disentangle the response to the organizational
requirements before emotional labor is undertaken. One such construct that
has been examined in the literature is that of dissonance. We conceive of this
dissonance as a perceived emotional state representing the dissonance between
felt emotion, and emotion that is perceived to be required and contend that it
arises from situational demands combined with individual differences. Emo­
tional labor in the form of surface acting or deep acting, then, is one possible
motivated response to a psychological condition of high perceived dissonance.
In Fig. 1 1. 1 , we offer our proposed model, and we next discuss the theoretical
conceptualizations of each section. We begin with a discussion of dissonance,
emotional labor, and their relationship, and then move to a discussion of situa­
tional factors, individual differences, and outcomes.
The Affective State of Perceived Emotional Dissonance
As

mentioned earlier, we conceive of perceived dissonance as a disconnect be­
tween emotions that are genuinely felt and those that are required for the situ­
ation (Middleton, 1989) . We place a high importance on this state of disso­
nance because we believe perceived dissonance to be a necessary precursor to
emotional labor. A relatively recent but well-supported theory of job satisfac­
tion, affective events theory (AET; Weiss & Cropanzano, 1996), underscores
the critical nature of affective states as precursors to outcomes such as job atti­
tudes and behaviors. Similarly, we consider perceived dissonance to be the af­
fective state necessary for the behavioral response of emotional labor. There
have, however, been a number of dissonance constructs (e.g., emotive disso-

Possible Individual Differences: Direct and Moderate Effects





Individual Differences
Role Internalization
Emotional Intelligence
Positive and Negative Affectivity

Job Attitudes




Emotional Labor &

/

Job Specific


Job Characteristics

Emotion Specific


·



Perceived Dissonance

'\

r-+



vs.
Felt Emotion

• .. ...........

Frequency
Duration
Variety
Customer Expectations

Situational Demands

'\..






Requirement

Display Rules

• Turnover Intention


Alternative Behaviors

---..

Perceived Emotional





Surface Acting
Deep Acting
Violence
Passive-Aggressiveness

Job Satisfaction
Cynicism

Organizational Commitment

Health & Ps):ch. Well-Being




Stress

• Somatic Symptoms


Burnout

Task Avoidance
Genuine Behavior

Job Related Behaviors


./

Turnover

• Absenteeism


Individual State

Organizational Citizenship

• Performance

Possible Motivated Acts
Possible Outcomes

-

e

FIG. 1 1.1. Proposed research model of perceived emotional dissonance and emotional labor.

194

RUBIN ET AL.

nance, emotional deviance, emotional dissonance) utilized in emotional labor
research (Kruml & Geddes, 2000; Morris & Feldman, 1 997; Zerbe, 2000). Our
review of the literature shows generally two types of dissonance: (a) disso­
nance experienced prior to the behavior of emotional labor, and (b) dissonance
experienced after this behavior has occurred. Zerbe (2000) attempted to pro­
vide some clarity surrounding the dissonance construct by distinguishing be­
tween these two types. Specifically, he described one type of dissonance as,
"the degree of mismatch between felt emotions and displayed emotions, or
faking" (p. 202). He described the second type of dissonance as "the mismatch
between expressed emotions and local norms" (p. 202) . Although both forms
of dissonance are valid and intuitively appealing, only a mismatch betweenfelt
emotion and normative requirementsfor emotional display captures a purely af­
fective state. Dissonance between felt and displayed emotion (or faking) re­
quires a behavioral response to an affective state. Note that this type of disso­
nance (faking) is the "labor" of emotional labor and reflects a process whereby
one experiences an affective state and is motivated to deal with it.
Thus, in order to separate clearly emotional states from motivated behavior,
we define dissonance as a purely affective state occurring prior to emotional la­
bor. Morris and Feldman ( 1 996) remarked that "It should require little emo­
tional labor to sell products one genuinely believes in" (p. 992) . Thus all things
being equal, the larger the gap between genuine felt emotion and those re­
quired for the situation, the greater perceived dissonance. Again, note that dis­
sonance is not a product of conflict between felt emotions and objective orga­
nizational requirements per se (i.e., written rules or supervisor instructions) ;
rather, it is a discrepancy between felt emotions and what the individual per­
ceives to be the required display. Much like the well-known literature on cog­
nitive dissonance, emotional dissonance is likely to create psychological dis­
comfort that individuals are motivated to reduce. We believe that emotional
labor is the motivated behavior that individuals employ in an attempt to recon­
cile their feeling of emotional dissonance.
Emotional Labor: A Motivated Behavioral Response
to Perceived Dissonance
We argue that individuals typically engage in one of two strategies for dealing
with a situation of perceived dissonance. These strategies are similar to Hoch­
schild' s ( 1983) and Grandey's (2000) mechanisms for managing emotional
demands, but do not include the appraisal of the affective state (perceived dis­
sonance). The first behavioral strategy for managing perceived dissonance,
"surface acting," is accomplished when an individual simulates emotion that is
appropriate for the situation. An example of surface acting would be when a
customer service representative smiles and acts warmly in the face of an abra­
sive and disrespectful customer who churns up feelings of anger and resent-

1 1.

P E R C E IVED DISSO NANCE AND E M OT IO NA L LAB O R

195

ment. The expression of positive emotion is fitting for the customer-service
context, but negative emotion is being experienced.
A second strategy for managing perceived dissonance is through the proc­
ess of "deep acting," in which the individual endeavors to experience the emo­
tion that is appropriate for the situation. In this case, the individual is actually
changing the feeling experienced prior to the behavior by changing the under­
lying causes, or the mental construction of the feeling state (e.g., cognitive re­
appraisal) . Changing the feeling can occur by evoking or suppressing an emo­
tion, or through trained imagination (Hochschild, 1983) . For example, a ticket
counter agent may imagine himself successfully completing transactions and
noticing the smiles on customers' faces. As Grandey (2003) implied, deep act­
ing may require advanced emotional regulatory skills.
In sum, Hochschild claimed that surface acting can be likened to compliance,
while deep acting strategy can be likened to conversion. Both strategies can be
adaptive (i.e., allow for productive performance) in terms of dealing with per­
ceived dissonance. Over the long term, deep acting should be more adaptive in
that it allows the individual to internalize a role, and thus reduce the level ofper­
ceived disscmance (as shown by the feedback loop in Fig. 1 1. 1 ) . Regardless of
adaptability, both strategies are based on the notion of acting-that is, manufac­
turing a substitute for the emotion one genuinely wants to display. Thus, surface
and deep acting may temper perceived dissonance, but are unlikely to eliminate
completely the source of the motivation to act, or perceived dissonance.
Extending Hochschild's typology, Ashforth and Humphrey ( 1993) noted
that there are instances in which the sanctioned emotion corresponds to the
felt emotion, as is ordinarily the case in medical emergencies. Ashforth and
Humphrey (1 993) and Tews and Glomb (2000) classified this match between
genuine emotion and the situational requirements as a third and important
way of accomplishing emotional labor. Yet, note that we propose that if one
does not perceive dissonance, one will not be motivated to engage in behaviors
that might be described as emotional labor. Regardless of the organization's
normative requirements, i f one does not experience dissonance, there will be
no motivation to act, to "put on a face", and thus, no emotional labor-the
emotional display will be genuine. If an employee is required by the organiza­
tion to act pleasantly, and he or she does not perceive this to be emotionally
taxing, there will be no necessity for him or her to "act" pleasantly. In fact, a
match between organizational requirements and felt emotion may even facili­
tate or emotionally energize an individual to display particular emotions and
would most likely appear to be effortless. Taken together then, perceived dis­
sonance and emotional labor represent the entire emotion regulation process
as Grandey (2000) discussed; however, our framework separates situational
factors, emotional states, and behaviors: Perceived dissonance is the percep­
tual state of appraisal of the situational factors, and emotional labor is the be­
havioral strategy chosen to manage the dissonant state.

196

RUBIN E T A L .

As noted, we view emotional labor as one potential and typical response to
perceived dissonance. There are then, multiple other strategies for reconciling
the state of perceived emotional dissonance, including withdrawing from the
task or interaction completely (e.g., hanging up on a customer) ; becoming hos­
tile or aggressive (e.g., raising one's voice) ; becoming violent (e.g., physically at­
tacking a customer); or using passive-aggressive behavior (e.g., acting polite
during the interaction and then sabotaging a customer's order afterward) .
These behaviors, however, are unlikely to be viewed by the organization as con­
structive and representative of appropriate protocol or policy and therefore, are
likely to be punished or at least heavily discouraged. Because both emotional la­
bor strategies (surface acting and deep acting) are attempts to produce organi­
zationally desired and rewarded behavior, they are likely to be the strategy of
choice utilized by individuals attempting to reduce perceived emotional disso­
nance, while reaping organizational benefits and rewards (e.g., promotions) .
Thus, depending on the situational requirements, the actual behavioral display
could be positive or negative (e.g., A debt collector must be intimidating and
downright hostile at times to be successful) . Although this chapter does not di­
rectly address at length the numerous possible responses beyond labor, we do
present some of them in the model to emphasize their potential importance in
future research. Thus, Fig. 1 1 . 1 is illustrative, not exhaustive.
Situational Demands of Perceived Dissonance
As stated earlier, our premise is that much of the early work on emotional labor
is descriptive of an emotional state influenced by a variety of factors. Given the
proposed model, our attempt here is to identify situational demands that im­
pact perceived dissonance and thus indirectly impact emotional labor. We dis­
cuss two categories of situational demands we believe largely contribute to the
perception of dissonance and subsequent emotional labor. As the model de­
picts, these situational demands are comprised of (a)job-speciJic demands, in­
cluding characteristics of the job, and (b) emotion-speciJic demands, including
emotional characteristics of the job, display rule norms, and a special case of
display rules, customer expectations. Interestingly, these variables have all ap­
peared in prior research on emotional labor; however, many have actually been
labeled as dimensions of the emotional labor construct (e.g., display rules and
frequency of emotional display) , and not as situational factors. Our view is that
these important variables work together to create the critical environmental
stimuli that may lead to an appraisal of perceived dissonance, and subsequent
emotional labor acts.
Job Characteristics. Humphrey (2000) suggested that the critical psycho­
logical states resulting from job characteristics (Hackman & Oldham, 1975)
clearly impact affective conditions like job satisfaction and thus they should

1l.

P E R C E IVED DIS SONAN C E AND E MOTIONAL LAB O R

197

strongly impact secondary work behavior ("behavior that is influenced by the
work task, but not actually part of the task," p. 239) including emotional dis­
plays. He further argued that job characteristics influence the social context of
the work environment. Thus, job characteristics are powerful contextual fac­
tors that contribute to an individual's affective state. In our terms, the affective
state of perceived emotional dissonance should similarly be impacted by job
characteristics.
To date, research on emotional labor and job characteristics has focused
primarily on job autonomy. For example, Morris and Feldman ( 1997) reported
a negative relationship between autonomy and emotional dissonance. Pugliesi
(1 999) explored the relationship between emotional labor and certain charac­
teristics of the job among university employees. She found negative correla­
tions between emotional labor and job characteristics such as control (auton­
omy) and complexity (skill required, level of challenge, variety and intrinsic
interest) of the job. Other research yielded similar results (Rafaeli & Sutton,
1989; Wharton, 1993) . Yet beyond skill variety and autonomy, very little re­
search exists linking other job characteristics to emotions. However, given the
effects found for autonomy, we argue that other job characteristics might im­
pact perceived dissonance similarly. For example, individuals who experience
a high degree of meaningfulness via skill variety, task identity, and task signifi­
cance are likely to experience a more positive psychological state (Humphrey,
2000) . This state should positively influence individuals' emotional appraisal
of their work situation and lead to lower perceived emotional dissonance, or at
least make it easier to maintain a positive emotional state. Similarly, high re­
sponsibility (autonomy) and knowledge of results (feedback) from the job
should lead to a lesser degree of perceived emotional dissonance.
Display Rule Norms. Display rules can be conceptualized as societal, oc­
cupational, and organizational norms that provide structure for service trans­
actions (Ashforth & Humphrey, 1993; Ekman, 1973a; Rafaeli & Sutton, 1989) .
Display rules serve to regulate the type of sanctioned expression in a given sit­
uation, as well as the degree to which it is expressed. In an organizational ser­
vice setting, display rules are often highly explicit to ensure consistency of ser­
vice. Additionally, display rules are found in all types of organizational
communication, including handbooks, performance appraisals, and new em­
ployee orientation classes. For example, at the Ritz-Carlton hotels, employees
are required to memorize service credo cards, which explicitly state what is
meant by service, and which all employees carry on their person at all times.
One service "basic" reads, "Smile-We are on stage. Always maintain positive
eye contact. Use proper vocabulary with our guests (e.g., "Good morning",
"Certainly" and "My pleasure")." Service employees are "on stage" in the
sense that these display rules require employees to become actors (Grandey,
200 1 ), acting out organizationally required emotions that may not be congru­
ent with their own personal emotions.

198

RUBIN ET AL.

Empirical research demonstrating a link between display rules and emo­
tional labor has provided mixed results. For instance, Hochschild (1983)
found that explicit display rules for flight attendants heavily impacted their be­
havior on the job. In contrast, Morris and Feldman ( 1997) found that the ex­
plicitness of display rules was negatively related to the frequency of emotional
labor. These differences are most likely due to the discrepancies in conceptu­
alizations of the construct. As we mentioned, it is important to disentangle the
stimulus (e.g., the organizational context) from an individual's internal state
(e.g., the felt experience) . Our model suggests that an incongruence between
the desired emotional expression required by the organization and the felt
emotion of the individual creates an experience of perceived dissonance.
Therefore, highly explicit display rules would exacerbate the importance of
sanctioned emotional expression and create a larger gap between required and
felt emotion. Thus, the less latitude employees have (i.e., the more explicit the
display rules) in expressing their own felt emotions in lieu of organizationally
sanctioned emotion, the greater is the possible perceived dissonance.
Customer Expectations. Schneider and Bowen ( 1 999) asserted, "By con­
centrating on [customer] needs, we have an opportunity to delve more deeply
into the customer's internal states to offer managers insights about how to
create customer delight, as well as how to avoid customer outrage" (p. 38) . In
addition, they highlighted multiple needs customers expect to receive from
service transactions that contribute to positive organizational outcomes, in­
cluding keeping promises and commitments, help when needed, and friendli­
ness, honesty, and politeness. Similarly, Zeithamal, Parasuraman, and Berry
(1990) found that customer evaluations are a function of 10 dimensions, in­
cluding responsiveness, courtesy, credibility, access, communication, and un­
derstanding (in Ashforth & Humphrey, 1993, p. 9 1 ) . Clearly, customer needs
and criteria for good service are laden with emotional content (e.g., "Greet ev­
ery customer with a smile") . Thus, customer expectations are a special and
highly salient type of display rule.
For example, a customer can easily interpret a simple sigh exuded by a cus­
tomer service agent to mean "I don't care about you-I want to go home,"
when in fact the agent may simply be tired. Recently, Pugh (2001 ) found a pos­
itive relationship between the display of positive emotion by bank employees
and customer positive affect. Similarly, Daus (2001 ) found converging evi­
dence from a laboratory experiment where expression of positive emotion con­
tributed to higher evaluations of performance for waiters/waitresses. This re­
search confirms what organizations have intuitively known for years-positive
employee displays satisfy customers. Organizations are increasingly advertis­
ing and educating their customers about this relationship, and customers now
approach service exchanges with clear expectations for what good service
should look like (e.g., Delta Airline's slogan, "We love to fly and it shows").

11.

P ER C EIVED D I S S O NANCE AND E M O T IO NAL LAB O R

199

Thus, for some organizations, customers will have incredibly high service ex­
pectations, including expectations for how employees will display emotions.
These expectations often are built into the fabric of the organization and serve
as yet another source of information about how employees should behave in
customer service interactions. Yet the paradox is that this may be taxing to em­
ployees because organizations that promise customers consistent top-notch
service thereby increase the emotional requirements of their employees. We
therefore propose that customers' expectations will be positively associated
with the perceived dissonance of employees.
Emotional Job Characteristics. Service-role jobs vary considerably across
industries and job categories. Logically, these jobs also differ in their emo­
tional characteristics, or the specific type and level of emotional expression in­
herent in the job itself. These specific emotional characteristics of the job may
lead to variance in how an individual actually feels on the job. Therefore, we in­
clude them in our model as predictors of perceived dissonance. Three emo­
tional job characteristics that have appeared in prior research are the fre­
quency of emotional display or interaction, duration of emotion display or
interaction, and variety ofemotions required. Again, these have previously been
presented as dimensions of the emotional labor construct (Hochschild, 1983;
Morris & Feldman, 1996, 1 997) . We separate them from the behavior of emo­
tional labor as Grandey (2000) did, and extend her work to conceptualize them
as precursors to perceived dissonance.
Frequency of emotional display or interaction can be defined as the rate of
emotional displays or interactions between service providers and clients
(Grandey, 2000; Hochschild, 1983; Morris & Feldman, 1997) . Some jobs in­
herently require higher frequencies of emotional displays than others. For ex­
ample, a cashier at a busy fast-food restaurant may be required to serve hun­
dreds of people in one shift, whereas a concierge at a small hotel may only
service a handful of people in a day. In their study of supermarket cashiers,
Rafaeli and Sutton (1 990) found that during particularly busy hours, cashiers
displayed fewer positive emotions. Other research has shown both positive
(Brotheridge & Lee, 1998) and negative (Grandey, 1999) relationships be­
tween frequency and emotional labor. Note that these studies did not measure
the impact of frequency on the emotional state, but rather the impact of fre­
'
quency on behavior, which may explain the inconsistent findings. The state
should impact behavior more strongly than contextual circumstances (Wake­
field, 1989). Therefore, for any particular job, the greater the frequency of
emotional display required, the more likely it is that an individual will experi­
ence a higher degree of perceived emotional dissonance.
Duration ofemotional display or interaction is generally referred to as dura­
tion. Service interactions that are relatively short tend to follow a scripted in­
teraction schema (Rafaeli, 1989) and require less emotional effort. As the du-

200

R U B I N ET A L .

ration increases, so does the level of emotional effort required to go beyond the
social script in order to comply with organizational demands. Back to our pre­
vious example: The cashier at the fast-food restaurant may on average spend
about 3 or 4 minutes with a customer, whereas the concierge at the hotel may
spend upward of 10 to 1 5 minutes working with a customer planning entire
outings. Thus, the concierge's job requires him or her to engage in positive
emotional displays for more extended periods. Long durations could lead to a
feeling of heightened perceived dissonance if in fact the discrepancy between
what the job requires from the concierge and how he or she feels is too great.
Thus, we propose that the longer the duration of interaction, potentially the
greater is the amount of perceived dissonance.
Certain jobs require a greater variety of emotions to comply with organiza­
tional requirements. For example, some occupations "often require frequent
changes of emotions that are displayed: positive emotions to build enthusiasm,
negative emotions to support discipline and neutrality of emotions to demon­
strate fairness and professionalism" (Morris & Feldman, 1996, p. 992) . In the
theatrical arts, actors often talk about "range" in reference to the ability to play
a wide variety of characters and across multiple genres (e.g., slapstick comedy
vs. Shakespearean theater) . Similarly, service employees are often called on to
express a wide range of emotions in order to successfully perform on the job.
Some of these emotional expressions may even contradict one another. For ex­
ample, an employee might force a smile while communicating a tough policy
issue to a customer in a firm voice. Few researchers have explored this idea
empirically, but there are similar theoretical conceptions (Grandey, 2000;
Wharton & Erickson, 1993) . Our proposition is that jobs requiring a wide
range of emotions and/or multiple emotions to express simultaneously have a
greater impact on one's emotional state than jobs requiring only a few emo­
tions to be expressed. Much like cognitive load, perceived dissonance could in­
crease under such circumstances.
The situational demands just outlined seem to contribute to an individual's
overall affective state. Clearly, a single property may not produce a significant
change in the emotional state; however, the cumulative effect contributes to
the experienced state. Thus, it should be acknowledged that although the
main effect of any single antecedent may impact perceived dissonance (posi­
tively or negatively), the potential interactions are likely to yield more complex
types of relationships and may strengthen, weaken or nullify the main effects. l
The Direct and Moderating Effect of Individual
Differences
In this model, perceived dissonance mediates the relationship between situa­
tional demands and emotional labor. All things being equal, a number of indi'We thank Blake Ashforth for bringing this important point to our attention.