Emotional labor involves the expression of socially desirable emotions in interpersonal interactions (Ashforth & Humphrey, 1993). It is generally associated with service work involving frequent interactions with customers that require workers to display a variety of emotions of varying degrees of intensity (Morris & Feldman, 1996). It is a form of emotion regulation that is accomplished through either surface or deep acting (Brotheridge & Lee, 1998). In performing surface acting, workers express required emotions without actually feeling them by hiding their real feelings and displaying ‘fake’ emotions. In performing deep acting, workers also express the required emotions but do so by attempting to summon these emotions within themselves; thus, they focus on bringing their external emotional display and internal feelings into alignment.
In general, existing research has focused primarily on emotional labor’s situational antecedents such as the frequency, variety, and intensity of emotional display (Morris & Feldman, 1996), and its consequences for worker well-being and performance. Existing research supports the need to distinguish surface acting from deep acting given that these aspects of emotional labor do not have a uniform impact on outcome variables. For example, whereas surface acting has been associated with increased emotional exhaustion, deep acting tends to generate feelings of personal accomplishment (Brotheridge & Lee, 2002). Indeed, although Hochschild (1983) argued that the external regulation of the emotions expressed by workers in service interactions would have pernicious effects, research suggests that what matters most is their internal regulation; i.e., how service workers choose to regulate their own emotional expression. As such, it is necessary to consider individual difference factors that might influence emotional regulation levels in service workers. The role of emotional intelligence and other individual difference variables in the process of performing emotional labor has generally been ignored (Bono & Vey, 2003). This is consistent with the dearth of research examining the relationship between individual differences and service delivery (Hurley, 1998). This research contributes to the literature that addresses the issue of the relative effects of the situation versus individual dispositions on behavior in the workplace (Davis-Blake & Pfeffer, 1989; George, 1992).
Research in self-monitoring (Gangestad & Snyder, 2000) and emotional regulation (Gross, 1998) has found that there are considerable differences in the extent to which individuals adjust their emotions in response to the social environment (Friedman & Miller-Herringer, 1991). Indeed, researchers considered dispositional factors to be more effective predictors of affective responses to work than situational factors (Newton & Keenan, 1991). As argued by Schneider (1987, p. 437), «the attributes of people, not the nature of the external environmental or organizational technology, or organizational structure, are the fundamental determinants of organizational behavior.» This line of reasoning was supported by research evidence that attitudes tend to be consistent despite changing situational contingencies (Staw & Ross, 1985).
The present research examines the extent to which emotional intelligence contributes to the explanation of surface and deep acting beyond situational factors (the frequency, duration, and intensity of emotional display). This study considers whether the nature of and the extent of emotional labor that is performed is due to individual differences, situational demands, or the interaction of dispositional and situational contingencies.
Individual differences serve as interactional competencies that facilitate or, in the case of negative affectivity, hinder role performance (Brotheridge & Lee, 2002). Emotional intelligence is included in the study rather than more general traits covered by five-factor model of personality because, as argued by Paunonen and Nicol (2001), unidimensional measures of personality traits are more likely to result in a more accurate understanding of the specific antecedents of behavior. Although multiple conceptualizations of emotional intelligence exist, the current study employs the definition developed by Mayer and Salovey (1997, p. 10):
Emotional intelligence involves the ability to perceive accurately, appraise, and express emotion; the ability to access and/or generate feelings when they facilitate thought; the ability to understand emotion and emotional knowledge; and the ability to regulate emotions to promote emotional and intellectual growth.
As measured in this study, emotional intelligence consists of four interrelated branches (Mayer & Salovey, 1997; Brackett & Salovey, 2006). The first and most basic capacity that is thought to serve as the foundation for the other branches is the ability to accurately perceive emotions generated internally or externally by other individuals or in other sources such as literature, art, or things (Branch 1 - Perception of Emotion). The second branch, the ability to use emotions to facilitate mental processes and communicate emotions, enhances one’s flexibility in these areas (Branch 2 - Emotional Facilitation of Thought). The third branch involves the ability to understand the nature and meaning of emotions and to use this information in reasoning processes (Branch 3 - Understanding Emotion). The final branch consists of managing or regulating emotions in others and oneself in a manner that facilitates understanding and well-being (Branch 4 - Managing Emotion).
Mayer and Salovey (1995) posited that emotionally intelligent people were more adaptive in regulating their emotions and, as a result, were more likely to: (a) attempt to present emotions that are pro-social and pro-individual; and (b) be flexible in regulating their emotions so that they are appropriate given the situation. They also argued that people who possess low levels of emotional intelligence were more likely to engage in emotional responses that are defensive in nature and that do not involve reframing. In contrast, highly emotionally intelligent individuals were more likely to engage in emotions that are adaptive to situational demands and to use reframing and other methods that «are consistent with one’s outlook on emotional responding» (Mayer & Salovey, 1995, p. 205). Thus, individuals with high levels of emotional intelligence should be able to regulate their emotions more flexibly than their low emotional intelligence counterparts (Lopes, Grewal, Kadis, Gall, & Salovey, 2006). The only study that has examined the relationship between emotional intelligence and aspects of emotional labor yielded mixed results. Totterdell and Holman (2003) found that emotional intelligence was associated with individuals’ ability to engage in positive refocusing but was not associated with differences in negative affect regulation, meeting display obligations, perspective taking, and faking emotions. These results may have been influenced by the self-report nature of the measure of emotional intelligence that was used in the study.
Mechanisms of influence for emotional intelligence
There are several potential mechanisms through which emotional intelligence may influence the performance of emotional labor (see figure 1; based on Moyle, 1995). Although the additive and moderating effects of personal characteristics on outcomes are familiar and popular (e.g., Houkes, Janssen, de Jonge, & Bakker, 2003), two additional relationships, confound and indirect effects, are possible. In a confound situation, emotional intelligence influences both the extent to which situational factors are perceived and the nature of the response to the situational demands. An indirect effect occurs when workers’ dispositions have an effect on how they perceive and experience their jobs, which in turn, affects the emotional labor that they perform (e.g., Schaubroek & Jones, 2000; Spector, Jex, & Chen, 1995). These four mechanisms of influence represent competing models and are hypothesized as follows:
H1: Model A - Direct or additive model: Both emotional intelligence and
situational demands directly influence how emotional labor is performed and
contribute independently to variance in emotional labor.
H2: Model B - Moderator model: Levels of emotional intelligence affect the
relationship between situational demands and emotional labor such that workers
with low levels of emotional intelligence are more attentive and reactive
to situational demands.
H3: Model C - Confound model: Emotional intelligence contributes to the
existence of spurious relationships between situational and emotional labor
variables, thus inflating their relationship and reducing their association
when emotional intelligence is controlled.
H4: Model D - Indirect model: Emotional intelligence conditions or influences
the perception of situational demands which, in turn, influence the extent
of emotional labor performed.
These hypotheses suggest the potential for several alternative relationships. Emotional intelligence will be associated with both situational demands and emotional labor (direct effects; figure 1a). At the same time, situational demands will predict emotional labor. Alternatively, controlling for emotional intelligence will reduce the strength of the relationship between situational demands and emotional labor (confound; figure 1c). Emotional intelligence will moderate the relationship between situational demands and emotional labor (moderator; figure 1b). Emotional intelligence will predict the perception of situational demands, which, in turn, will predict the nature of emotional labor performed (indirect effects; figure 1d).
Study participants included 188 undergraduate students at a Canadian university who were employed on a part-time basis in service occupations including: retail sales, restaurant servers, and hotel catering staff. The participants were an average of 23.17 years old (SD 6.46), and approximately 60 percent of the sample was male. Participants completed the measures in two stages over a two month period: Time 1: emotional intelligence; Time 2: emotional labor and its situational predictors.
Emotional Intelligence. This study employed the online version of the MSCEIT (Mayer, Salovey, & Caruso, 2000; 141 items, v 2.0) as its measure of emotional intelligence. The MSCEIT was employed because of its strengths in several areas: (a) since the MSCEIT measures ability, there are fewer concerns regarding response set and other issues that are common in self-report measures; (b) its overall and branch scores have strong reliability; and (c) its low correlations with personality measures indicates that it measures unique variance in criterion variables (Mayer et al., 2000). An overall emotional intelligence score was computed (MSCEIT total; α= .80) in addition to two area level scores including: (a) Emotional Experiencing (α= .86), which measures «how accurately a person can ‘read’ and express emotion, and how well a person can compare that emotional stimulation to other experiences» (Mayer et al., 2000, p. 47); and (b) Emotional Reasoning (α= .79), which measures «how accurately a person understands what emotions signify… and how emotions in him/herself and others can be managed» (Mayer et al., 2000, p. 47). The MSCEIT also provides scores for four branches as follows:
1. Branch 1- Perception of Emotion: the ability to accurately perceive
emotions in faces (Section A) and pictures (Section E);
2. Branch 2 - Emotional Facilitation of Thought: the ability to use emotions
to as an aid in reasoning (Section B: Facilitation; Section F: Synesthesia);
3. Branch 3 - Understanding Emotion: the ability to understand changes
(Section C) and blends (Section G) in emotions.
4. Branch 4 - Managing Emotion: the ability to manage emotions in relationships
(Section H) and oneself (Section D).
These scores were compared with those of a standardization sample whose mean is a score of 100 (SD 15).
Emotional Labor and its Situational Predictors. The Emotional Labour Scale (Brotheridge & Lee, 2003) was used to measure the frequency (3 items, a= .52), intensity (3 items, a= .81), and variety (3 items, a= .79) of emotional display as well as surface acting (3 items, a= .80) and deep acting (3 items, a= .83). Respondents used a five-point scale ranging from 1 (never) to 5 (always) to rate how frequently they performed specific behaviors on an average day at work. Sample items include: [How frequently do you] «display specific emotions required by your job» (frequency), «display many different emotions when interacting with others» (variety), «pretend to have emotions that I don’t really feel» (surface acting), and «try to actually experience the emotions that I must show» (deep acting).
Means, standard deviations and zero-order correlations
As indicated in table 1, whereas deep acting is moderately and significantly correlated with all three situational demands, it is only weakly (but significantly) correlated with MSCEIT (total score), the reasoning form of emotional intelligence, and emotional management. Deep acting is approaching significance in its relationship with experiencing and facilitating emotions. Surface acting is weakly but significantly correlated with frequency and variety of emotional display, and approaching significance in its relationship with identifying emotions.
Given that previous research has found gender differences in emotional intelligence when measured with the MSCEIT, gender was entered as a control variable in all the regression analyses that were performed. In Model A, emotional intelligence makes an independent contribution to the prediction of deep acting and surface acting beyond that of situational demands (figure 1a). To test this model, a hierarchical regression was performed in which gender was entered as the first step, situational demands were entered at the second step, and emotional intelligence was entered at the third step. These analyses found that emotional intelligence, whether measured as a total score, area scores, or branch scores, did not contribute to the explanation of emotional labor (either deep or surface acting) beyond the situational demands. In all cases, the value of the adjusted r-square remained the same as a result of the addition of the MSCEIT variable(s).
In Model B, emotional intelligence moderates the relationship between situational demands and emotional labor (figure 1b). In other words, a respondent’s level of emotional intelligence affects the impact of situational demands on levels of emotional labor. To test this model, hierarchical regression was performed in which gender was entered as the first step, situational demands were entered at the second step, emotional intelligence at the third step, and the interaction term (emotional intelligence x situational demands) was entered in the final step. Neither the addition of emotional intelligence nor the interaction term contributed to the explanation of emotional labor (either deep or surface acting) beyond situational demands. As with Model A, the value of the adjusted r-square did not change significantly as a result of the addition of the MSCEIT variable(s).
In Model C, emotional intelligence contributes to a spurious relationship between situational demands and emotional labor. If this spurious relationship exists, then controlling for emotional intelligence should reduce the extent of the relationship between situational demands and emotional labor (figure 1c). Baron and Kenney’s (1986) mediated regression test was used to examine the presence of confounding effects. The first requirement of this test is that emotional intelligence be significantly correlated with both situational demands and emotional labor. A review of the correlations provided in table 1 indicates that a significant positive zero-order correlation exists only between MSCEIT (total score) and deep acting and MSCEIT (total score) and frequency of emotional display. As such, only the deep acting, frequency, and MSCEIT (total score) variables were considered for further analyses. The second requirement, that frequency of emotional display be correlated with deep acting, was also met. The third requirement involves performing a regression analysis in which deep acting is simultaneously regressed onto MSCEIT (total score) and frequency of emotional display in Step 2 (with gender entered in Step 1). A confounding effect would exist if emotional intelligence, and not frequency, predicted deep acting. This was not found in the regression analysis.
In Model D, emotional intelligence predicts the perception of situational demands, which, in turn, predict the nature of emotional labor performed (figure 1d). These indirect effects were tested through the use of Baron and Kenney’s (1986) mediated regression test. As discussed in Model C, given that a significant correlation existed only between MSCEIT (total score), deep acting, and frequency of emotional display, only these three variables were retained for the regression analysis. A regression analysis was then performed in which deep acting was simultaneously regressed onto MSCEIT (total score) and frequency of emotional display. The frequency of emotional display became the only significant predictor of deep acting, thus, suggesting that it serves as a mediator between emotional intelligence and deep acting (Baron & Kenney, 1986).
The primary purpose of this research was to examine the role of emotional intelligence in the process of performing emotional labor. The current study illustrates the multifaceted relationship between emotional intelligence and emotional labor. The zero-order correlations indicated that: (a) overall emotional intelligence levels; and (b) an ability to read, understand, and manage emotions were positively associated with deep acting (i.e., making an effort to feel the emotions that need to be displayed) but not surface acting (faking displayed emotions). In contrast, surface acting was positively associated with the more basic capacity of accurately perceiving emotions. These results suggest that the nature and extent of one’s emotional intelligence may be implicated in how emotional labor is performed. Emotionally intelligent individuals appear to choose deep acting as the means of expressing expected emotions in interpersonal interactions. Deep acting is the most adaptive form of emotion labor given that it is associated with a sense of authenticity and personal accomplishment rather than emotional exhaustion and depersonalization as is surface acting (Brotheridge & Lee, 2002).
The connection between emotional intelligence and emotional labor, however, appears to be indirect in nature. Contrary to Model A, this study did not find that emotional intelligence added to the prediction of variance in emotional labor beyond situational demands. However, a potential explanation for this result may be that the nature of the situation affected the influence of personality variables on service provision (Hurley, 1998). The participants in this study were employed in occupations characterized by relatively strong normative expectations regarding the nature, frequency, and intensity of emotions to be displayed. This strong situation may overpower the influence of individual differences such as emotional intelligence and, thereby, weaken the strength of the relationship between emotional intelligence and emotional labor (Snyder & Ickes, 1985). In his review of the literature, Hurley (1998) found that personality had stronger effects on service delivery in jobs involving intimacy or in critical jobs. Given the short duration of the service provision, the standardization of the service delivery process (hence, the extensive external control mechanisms that exist), and the social versus intimate nature of the service interaction (Hurley, 1998), the service providers in the current sample were more likely to engage in brief encounters than in durable service relationships (Gutek, Bhappu, Liao-Troth, & Cherry, 1999).
Emotional intelligence did not moderate the relationship between situational demands and emotional labor in this study. In other words, levels of emotional intelligence did not affect the nature of the emotional labor that was performed given situational demands. Rather, the key role of emotional intelligence seems to be as a predictor of the perception of the situational demands themselves, which, in turn, predict the nature of emotional labor performed. Workers with higher levels of emotional intelligence were more likely to perceive the need to frequently display emotions as part of their work role and to perform deep acting in response to these situational demands. More broadly, this suggests that role requirements and demands will not necessarily be perceived in a uniform manner.
Limitations and future research
The data for this study were collected at two points in time, thus somewhat reducing the concerns associated with common method bias. However, research should be undertaken that examines the relationships among the study variables over time so that causal relationships can be established. A longitudinal design would also permit us to identify the sequential unfolding of deep acting and surface acting. For example, it is possible that, as workers begin to identify with their work over time, they progress from performing primarily surface acting to deep acting. Furthermore, research is needed that examines the role of emotional intelligence in the relationship between emotional labor and its outcomes given that emotional intelligence may intervene at several points in this process. For example, beyond influencing the perception of situational demands, emotional intelligence may act to bolster the positive effects of deep acting and/or to buffer the harmful effects of surface acting.
A cross-validation study should be undertaken that examines the relationships among study variables for workers who engage in durable service relationships (Gutek, Bhappu, Liao-Troth, & Cherry, 1999). As indicated earlier, it is possible that such occupations offer workers more room for autonomy and discretion in how they respond to the emotional demands of work than do occupations in which interactions are largely scripted in nature. It is possible that emotional intelligence will have more predictive power in the high-relationship, high-autonomy occupations given that they are more ambiguously structured (Weiss & Adler, 1984).
A potential implication of this study and an area for future study is the use of emotional intelligence measures in selecting service workers. This study tentatively indicates that doing so may yield a better person-job fit especially since emotional intelligence is associated with deep acting, a healthier means of performing emotional labor than surface acting. However, managers should be cautioned against using a single assessment instrument as the basis for selection screening, or attributing performance exclusively to individual differences (Hurley, 1998). Options such as employee socialization and emotion management training may be useful in supporting the productive displays of emotion in the workplace. In requiring that employees present a certain ‘face’ in the workplace, managers are acknowledging that work is, in essence, an interpersonal arena, and that emotions play a fundamental role in the day-to-day functioning of organizational life. If this is indeed the case, then understanding the role of emotional intelligence in building a healthy workplace is vital.
An earlier version of this manuscript was presented at the 18th annual meeting for the Society of Industrial and Organizational Psychologists, Orlando, Florida, 2003.
Ashforth, B.E. & Humphrey, R.H. (1993). Emotional labor in service roles: the influence of identity. Academy of Management Review, 18, 88-115.
Baron, R.M. & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: conceptual, strategic and statistical considerations. Journal of Personality and Applied Psychology, 51, 1173-1182.
Bono, J.E. & Vey, M.A. (2003). Toward understanding emotional management at work: a quantitative review of emotional labor research. Working Paper.
Brackett, M.A. & Salovey, P. (2006). Measuring emotional intelligence with the Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT). Psicothema, 18, supl., 34-41.
Brotheridge, C.M. & Lee, R.T. (1998). On the dimensionality of emotional labor: development and validation of an emotional labor scale. Paper presented at the First Conference on Emotions in Organizational Life, San Diego State University, San Diego, CA, August 7-8, 1998.
Brotheridge, C.M. & Lee, R.T. (2002). Testing a conservation of resources model of the dynamics of emotional labor. Journal of Occupational Health Psychology, 7, 57-67.
Brotheridge, C.M. & Lee, R.T. (2003). Development and validation of the emotional labour scale. Journal of Occupational and Organizational Psychology, 76, 365-379.
Davis-Blake, A. & Pfeffer, J. (1989). Just a mirage: the search for dispositional effects in organizational research. Academy of Management Review, 14, 385-400.
Friedman, H.S. & Miller-Herringer, T. (1991). Nonverbal display of emotion in public and in private: self-monitoring, personality and expressive cues. Journal of Personality and Social Psychology 61, 766-775.
Gangestad, S.W. & Snyder, M. (2000). Self-monitoring: appraisal and reappraisal. Psychological Bulletin, 126, 530-555.
George, J.M. (1992). The role of personality in organizational life: issues and evidence. Journal of Management, 18, 185-213.
Gross, J.J. (1998). Antecedent- and response-focused emotion regulation: divergent consequences for experience, 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 relationships and encounters. Journal of Applied Psychology, 84, 218-233.
Hochschild, A.R. (1983). The managed heart: the commercializationof human feeling. Berkeley, CA: University of California Press.
Houkes, I., Janssen, P.P.M., de Jonge, J., & Bakker, A.B. (2003). Personality, work characteristics and employee well-being: a longitudinal analysis of additive and moderating effects. Journal of Occupational Health Psychology, 8, 20-38.
Hurley, R.F. (1998). Service disposition and personality: a review and a classification scheme for understanding where service disposition has an effect on customers. Advances in Services Marketing and Management, 7, 159-191.
Lippa, R. (1983). Expressive behavior. In L. Wheeler & P. Shaver (eds.): Review of personality and social psychology (vol. 4, pp. 181-205). Beverly Hills, CA: Sage Publications.
Lopes, P.N., Grewal, D., Kadis, J., Gall, M., & Salovey, P. (2006). Evidence that emotional intelligence is related to job performance and affect and attitudes at work. Psicothema, 18, supl., 132-138.
Mayer, J.D. & Salovey, P. (1995). Emotional intelligence and the construction and regulation of feelings. Applied and Preventive Psychology, 4, 197-208.
Mayer, J.D. & Salovey, P. (1997). What is emotional intelligence. In P. Salovey & D. Sluyter (eds.): Emotional development, emotional literacy and emotional intelligence: implications for educators (pp. 3-31). New York: Basic Books.
Mayer, J.D., Salovey, P., & Caruso, D.R. (2000). MSCEIT (Version 2). Toronto, OH: Multi-Health Systems.
Morris, J.A. & Feldman, D.C. (1996). The dimensions, antecedents and consequences of emotional labor. Academy of Management Review, 21, 986-1010.
Moyle, P. (1995). The role of negative affectivity in the stress process: tests of alternative models. Journal of Organizational Behavior, 16, 647-668.
Newton, T. & Keenan, T. (1991). Further analyses of the dispositional argument in organizational behavior. Journal of Applied Psychology, 76, 781-787.
Paunonen, S.V. & Nicol, A.A.A.M. (2001). The personality hierarchy and the prediction of work behaviors. In B.W. Roberts & R. Hogan (eds.): Personality psychology (pp. 161-191). Washington D.C.: American Psychological Association.
Schaubroek, J. & Jones, J.R. (2000). Antecedents of workplace emotional labor dimensions and moderators of their effects on physical symptoms. Journal of Organizational Behavior, 21, 163-183.
Schneider, B. (1987). The people make the place. Personnel Psychology, 40, 437-454.
Snyder, M. & Ickes, W. (1985). Personality and social behavior. In G. Lindsey and E. Aronson (eds.): Handbook of social psychology (3rd ed., pp. 883-948). New York: Random House.
Spector, P.E., Jex, S.M., & Chen, P.Y. (1995). Relations of incumbent affect-related personality traits with incumbent and objective measures of characteristics of jobs. Journal of Organizational Behavior, 16, 59-65.
Staw, B.M. & Ross, J. (1985). Stability in the midst of change: a dispositional approach to job attitudes. Journal of Applied Psychology, 70, 469-480.
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.
Weiss, H.M. & Adler, S. (1984). Personality and organizational behavior. Research in Organizational Behavior, 6, 1-50.