How Does Emotional Labor Impact Employees’ Perceptions of Well-Being? Examining the Mediating Role of Emotional Disorder

Abstract

Since bank employees are prone to high psychological pressure, it is key to explore the influencing mechanism of their emotional labor so as to relieve their pressure, as well as improve organizational performance and service quality. This study aimed to investigate the effects of emotional labor on bank employees’ well-being and to determine the mediating role of emotional disorder in this relationship. Employees responded to a survey regarding their use of emotional labor as well as perceptions of their well-being and emotional disorder. The results showed that employees’ use of emotional labor was related to their perceptions of well-being and confirmed the mediating role of emotional disorder in this relationship. The results indicated that surface acting has a significant negative impact on employee well-being, while deep acting has a significant positive impact. Moreover, emotional disorder played a role in mediating emotional labor and employee well-being, and emotional disorder was positively correlated with surface acting and negatively correlated with deep acting. The results revealed that developing deep-acting skills is important for increasing front-line bank staff’s well-being in China, who are accustomed to repressing their emotions, and emotional disorder might occur more often than has been previously believed, which worsens their well-being.

1. Introduction

Under the conditions of high-intensity and long-term emotional labor, bank employees are prone to psychological pressure. According to the Chinese Medical Association, 72.8% of Chinese financial industry employees self-reported experiencing high pressure all the time [1]. With an increasing number of young people using online banking, most off-line consumers are old people who require more patience, which causes the front-line bank staff more emotional pressure than they previously experienced. In contrast to the work of service employees in other industries, every move that front-line bank staff makes is monitored, and their hard work consumes emotional resources. To create a good service image, the frequency of and attention paid to emotional expression are quite high; thus, compared with other types of workers, front-line bank staff must perform more physical and mental labor, they consume more psychological resources, and they are more prone to mental health problems. Employees experience heavy emotional labor while controlling their feelings to express the emotions expected by their organizations through facial and bodily displays [2]. Over the past decades, much attention has been devoted to the association of emotional labor with employee outcomes. For example, evidence suggests that the psychological states of employees can function as a mediator between emotional labor and employee outcomes [3,4,5,6].

Despite the increasing interest in emotional labor, most studies have focused on direct relationships between the predictors and outcomes of emotional labor. Unfortunately, there is little evidence regarding variables that mediate the relationship between employees’ emotional labor and work outcomes, especially well-being [3,4]. Emotion regulation is crucial for social adjustment and well-being [7], and it is necessary to explore the role employees’ emotional labor plays in their well-being. According to the Conservation of Resources (COR) theory [8,9], employee resources are consumed when they perform surface or deep acting. If their resources are not replenished in a timely manner, employees will experience an imbalance in resources that leads to negative outcomes, such as job burnout, emotional disorder, and, ultimately, lower well-being. On the other hand, when employee resources are replenished in a timely way, their emotional labor will play a positive role, leading to better well-being.

Therefore, the overall purpose of this study is to investigate the effects of emotional labor on bank employees’ well-being and to determine the mediating role of emotional disorder in this relationship. Specifically, the objectives are threefold. First, this study examines the relationship between the two acting modes of emotional labor (i.e., deep acting and surface acting) and well-being in a banking service setting. Second, this study develops a clearer understanding of the effects of emotional disorder on the relationship between emotional labor and well-being. Third, this study provides a sample from China that is distinct from Western cultures, where most of the emotional labor research has occurred, to discover the underlying mechanism in the demand-strain process of emotion regulation.

2. Literature Review

2.1. Defining Emotional Labor

Emotional labor refers to a process by which employees must manage their emotions to meet organizational rules for emotional displays or norms concerning the appropriate emotional reactions in specific situations [10]. Two emotional labor strategies have been identified: (1) surface acting refers to employees’ modification of only their observable expressions in order to adhere to expectations about emotional expression, and (2) deep acting involves the modification of both felt and underlying emotions to adhere to rules about emotional expression [6].

In surface acting, employees change their observable features, gestures, facial expressions, or vocal tones, whereas in deep acting, they change both these observable features and their unobservable inner states [11]. Of these forms of emotional labor, passive acting and deep acting are believed to provide better and more positive outcomes for both service providers and customers [9,12]. A large body of cross-sectional and longitudinal research [13,14] has documented that surface acting negatively affects employees’ job satisfaction and leads to emotional exhaustion. Surface acting aims to alter the outward emotional expression without changing the actual feeling. It involves the suppression of negative emotions and mimicry of positive emotional expression for jobs that require positive displays of emotion [4]. Surface acting may affect employee well-being in general, emotional exhaustion and job satisfaction in particular. This is because surface acting is effortful, depletes mental resources, undermines employees’ sense of authenticity, and hampers positive social interactions with customers [13,15,16,17]. Evidence has revealed that employees frequently engage in surface acting, especially when they interact with unpleasant clients and confront stress and anger [18,19,20].

Deep acting involves the regulation of feelings to seem authentic, whereas surface acting involves the regulation of emotional expression to fulfill job duties. During deep acting, people regulate the precursors of emotion (i.e., they use antecedent-focused regulation) by modifying their perception of the situation. Deep acting is believed to provide positive and better outcomes for both service providers and customers [9,12]. Kim [12], for example, revealed that emotional labor does not cause burnout among hotel service personnel, especially when they sincerely try to have appropriate emotions (deep acting) rather than project false ones (surface acting). Evidence also implies that deep acting is more efficacious than surface acting for the performance and well-being of customer service employees [21]. Both deep and surface acting involve attempts to display the required emotions: surface acting changes the focus of thoughts to things that induce the required emotions, while deep acting changes the cognitive basis for emotions. For example, regulating emotion using deep acting, such as recalling pleasant memories, can eliminate feelings of sadness [13,22]. Thus, emotional labor does not cause burnout among hotel service personnel, especially when they sincerely try to have the appropriate emotions (deep acting) rather than adopting false ones (surface acting).

2.2. Construct of Well-Being

Well-being is a good psychological state of employees in their own fields of work. It can reflect the level of individual psychological arousal and is an effective index for measuring mental health. The construct of well-being includes (1) individual well-being, which relates to job satisfaction and burnout, and (2) organizational well-being, which relates to performance and withdrawal behaviors [21]. Employees’ job well-being can increase their attention and cognitive ranges [23] and enhance their self-efficacy [24]. Individual motivation can affect well-being. Waterman [25] proposed that different tasks are associated with different types of well-being. Individuals’ pursuit of goals consistent with their internal motives generate higher satisfaction. Social support in interpersonal relationships positively impacts well-being [26]. Social support also provides employees with an opportunity to modify their emotions and maintain good emotions when coping with stress. Obtaining organizational support can encourage employees to devote more energy to their work. Of course, stress from being overworked can reduce employees’ well-being, which may occur because of the emotional disorder caused by work stress, leading to a decline in employee job satisfaction.

Because young people are choosing to bank online, front-line bank staff in China often complain of high stress and poor well-being because they are faced with an increasing number of older, “offline” consumers. The elderly are respected in the Chinese culture and require more patience, which increases the pressure on front-line bank staff. Our research aimed to explore the mechanism between emotional labor and well-being to improve the well-being of front-line bank staff by developing emotional skills. Much evidence has suggested that emotional labor influences individuals’ well-being, which is considered an indicator of psychological health [27]. Initial evidence has suggested that burnout is a response related to frequent social and emotional interaction with clients [22]. Emotion regulation using surface acting is negatively related to well-being [22,28], and emotion regulation with deep acting is positively related to well-being [13,21,22,28]. A large number of studies have found that emotional labor not only leads to a decline in employee well-being, but also causes job burnout, which greatly reduces job satisfaction. Grandey [13] showed that employees who adopt surface behavioral strategies can consequently experience negative emotions and an impaired sense of self-worth [29,30]. Lin & Huang [31] found that surface acting was positively associated with job burnout among workers in hospital emergency departments. Relevant studies have also shown that the employee use of deep-seated behavioral strategies can result in positive emotional experiences for employees and have a positive impact on employee well-being [31]. Tolich [32] showed that employees feel inner satisfaction and pleasure when they are in harmonious communication with customers. Through a control group experiment, the surface-acting strategy was found to reduce the well-being of employees to some extent, while the deep-acting strategy improved their well-being [33]. Cheung (2006) [10] found that well-being was negatively related to surface acting in China. Earlier studies emphasized the potential negative impact of emotional labor on the well-being of service workers [30,34], and some recent studies argued that emotional labor can produce positive employee outcomes [30,35]. The most significant factor determining the effects of emotional labor was found to be the emotional labor strategy, which is the emotion regulation (i.e., surface versus deep acting) utilized by an employee to adhere to the organization’s rules for emotional displays [28,36]. The effortful processes of surface acting and deep acting may be related to employee stress and health as well as organizational well-being [37]. Specifically, evidence indicates a positive association with employees’ well-being and deep acting but a negative association with surface acting [9,16]. In line with this perspective, Lv et al. [4] investigated hotel employees in China and noted that, while surface acting contributed to the employees’ emotional exhaustion, deep acting reduced their sense of emotional depletion. Similar results were derived from the studies of Kim [13] and Lam and Chen [38] among front-line employees in the hotel industries in India, the US, and China. The following hypotheses (Hs) are thus proposed:

H1. 

Surface acting has a negative effect on well-being.

H2. 

Deep acting has a positive effect on well-being.

2.3. Emotional Disorder and its Mediating Role

Emotional disorder is a common psychological state of employees in the workplace. When employees realize that their true feelings are inconsistent with the emotions required by the established rules of the organization, emotional disorder will occur [39]. Emotional disorder has three definitions. The first considers emotional disorder to be an antecedent of emotional labor, which is the difference between employees’ feelings and their organization’s requirements [40,41]. The second considers emotional disorder to be one of the components of emotional labor [42,43]. The third considers emotional disorder to be the outcome of emotional labor, or the inner feelings of employees if there was an inconformity between employees’ expressions and their true emotion feelings [44]. Recent studies consider emotional disorder as an emotional state of restlessness, anxiety, and nervousness that occurs when employees’ expressions and their true emotion feelings were nonconforming, so this paper also considered emotional disorder to be the outcome of emotional labor.

Emotional disorder leads to a decline in employee satisfaction and self-worth and reduces employee well-being. Deep-seated strategies create more true or positive emotions by adjusting internal psychological emotions to lessen emotional disorder [45]. However, surface acting consumes many resources to prevent inappropriate displays of emotion, so it creates emotional disorder [19]. Deep-seated strategies may have a positive impact on employees’ work, allowing them to receive praise from their leaders and enhancing their self-confidence and sense of achievement [30]. This outcome occurs because of a negative correlation between deep acting and emotional disorder such that employees seek more positive emotional experiences to reduce emotional disorder. On the other hand, surface acting consumes a great deal of psychological resources, causing employees to experience emotional exhaustion or role conflict [13]. Based on the COR theory, employees devote their resources to work and hope to receive a positive response. Employees consume many resources in terms of emotional labor, and an imbalance will result if they do not receive a timely resource supplement, which will lead to emotional disorder. Specifically, surface acting focuses on controlling the external emotional behaviors which will damage employees’ true feelings, and it consumes many resources without replenishing them in a timely manner [14]. Therefore, employees can easily experience emotional disorder when they use surface acting, because surface acting has a positive association with emotional disorder. However, deep acting focuses on people’s inner feelings and adjusts their emotions to the cognitive evaluation process, so it obtains a balance between inner and external emotions, which is a resource acquisition process, thus attenuating emotional disorder [14].

In summary, emotional disorder plays an important role in the relationship between emotional labor (surface- and deep-acting strategies) and well-being, thus affecting the mental health of employees. Emotional disorder occurs when employees’ real feelings are inconsistent with the emotions dictated by organizational rules, as employees strive to show the emotions that the organization requires them to express. If they focus too much on their external display of emotions, they are likely to experience emotional disorder, resulting in psychological fatigue [9]. In addition, if employees’ emotional experiences conflict with organizational requirements, they will try to compensate for their inappropriate emotional expressions, increasing their chances of having emotional disorder [17]. Using the COR theory and empirical research, Grandey [15] has proven that employees’ surface-acting strategies are positively correlated with emotional exhaustion. Deep-acting strategies focus on adjusting employees’ inner experiences, resulting in a consistency between their internal and external emotions [9]. The use of surface-acting strategies can cause employees to be mentally tired, while the use of deep-acting strategies has a positive effect on employee enthusiasm [21], resulting in more positive and real emotional experiences and reducing emotional disorder [20]. Therefore, we find that the use of different emotional labor strategies has different effects on experiences of emotional disorder. When employees’ inner emotional experiences are inconsistent with the emotions they want to show, they will experience many negative thoughts and emotions that have long-term effects on their mental health. In an experimental investigation of police, Chapman [9] found that long-term emotional disorder can lead to emotional exhaustion among police, which has a negative impact on their job satisfaction. From the perspective of the COR theory, emotional disorder consumes psychological energy. Chen [46] found that emotional disorder in children with attention deficit disorder (ADD) has its own particular context. When these children encounter negative situations, they experience a great deal of emotional disorder. Emotional disorder has a significant negative impact on well-being. A study of Indian medical equipment salesmen also showed that emotional disorder is negatively correlated with employee well-being.

Social service workers use different strategies to express their emotional experiences in their daily work. Research shows that surface behavior has a significant positive correlation with turnover intention and job burnout. When employees adopt emotional labor strategies to show appropriate emotions, if they do not recognize their emotions, they will experience a great deal of disordered thinking, which is closely related to emotional disorder, causing emotional exhaustion, lower job satisfaction, poor well-being, and other adverse effects. If employees feel that their emotional experiences reflect their true inner feelings, they will experience less emotional stress and lower levels of emotional disorder [47].

Based on the above analysis, this study concludes that surface acting is negatively correlated with employee well-being, while deep acting is positively correlated with employee well-being. Emotional disorder plays an intermediary role and is positively regulated by surface acting and negatively regulated by deep acting. The following was hypothesized:

H3. 

Surface acting has a positive effect on emotional disorder.

H4. 

Deep acting has a negative effect on emotional disorder.

H5. 

Emotional disorder has a negative effect on well-being.

This study thus proposed the development of a structural model of emotional labor strategies (i.e., surface acting versus deep acting), well-being, and emotional disorder (Figure 1). The following is our proposed conceptual framework:

3. Research Methods

3.1. Data Collection

We sought to recruit participants who represent the general emotional labor status in China, which thus led to recruiting front-line bank staff from two representative provinces of China. The first province, Zhejiang, is located in the eastern coastal region and represents the developed regions of China. The other, the central province of Hubei, represents the underdeveloped regions. A total of 160 participants were recruited from four large state-owned banks (i.e., Bank of China, Agricultural Bank of China, Industrial and Commercial Bank of China, and Construction Bank of China) and 175 other banks (i.e., City Commercial Bank, Development Bank, China CITIC Bank).

Based on the study aim, an electronic survey instrument was developed using information from participants. The survey consisted of five items pertaining to the general characteristics of the respondents, such as gender, age, years of work experience, marital status, and education level (

Table 1. Profile of the respondents.

	N	%
Gender (n = 329)
Male	141	42.9
Female	188	57.1
Age (n = 329)
18–25 years old	96	29.2
26–35 years old	112	34.0
36–45 years old	85	25.8
46 years old and above	36	10.9
Mean (in years)		32.4
Standard Deviation		9.8
Education Level (n = 329)
Bachelor’s college or below	222	67.5
Master’s degree	64	19.5
PhD or doctorate degree	43	13.1
Marital status (n = 329)
Single	188	57.1
Married	129	84.9
Years of professional work (n = 329)
3 years and under	188	57.1
4–6 years	111	33.7
7–12 years	12	3.6
13 years and above	18	5.5
Mean (in years)		4.0
Standard Deviation		3.3

); 11 items related to emotional labor as the independent variable; 12 items related to well-being as the dependent variable; and five items related to emotional disorder as the mediating variable. A 5-point Likert-type response scale was used for these items (1 = strongly disagree to 5 = strongly agree). To measure emotional labor, 11 items, including seven for surface acting and four for deep acting, were used for the measurement developed by Gandey’s (2003) and Dfendorff’s (2005) Emotional Labor Questionnaire [47]. A higher score indicated a higher degree of emotional labor (see

Table 2. Descriptive statistics about the respondents’ use of emotional labor strategies and perceived well-being and emotional disorder.

Items	M	SD
Surface Acting (α = 0.94)	3.83	0.83
1. In order to serve customers in the right way, I will show the right emotions and hide my true emotional feelings.	3.88	0.80
2. At work, when I communicate with others, I will show a good mood, even if it is not in my heart.	3.83	0.86
3. Exhibiting appropriate emotions at work is like acting to me.	3.78	0.83
4. I just simply pretend to be emotional about what I need to show at work.	3.80	0.83
5. I will show the emotions I need at work, but I will not change my true emotional feelings.	3.84	0.83
6. In my work, the emotions I express in the face of others are different from the real feelings I feel inside.	3.83	0.78
7. When communicating with others, my emotions are adjusted and modified.	3.88	0.85
Deep Acting (α = 0.88)	3.88	0.70
1. I try to experience the emotions that should be expressed at work, rather than just only pretending to feel them.	3.81	0.77
2. I try to express my feelings to customers from my heart rather than pretending.	3.91	0.67
3. I try to feel the emotions I need to express to my customers in my work.	3.92	0.70
4. I try to cultivate the emotions I need to express to my customers.	3.90	0.65
Perceived Well-Being (α = 0.90)	3.67	0.89
1. Work makes me feel terrible. Nothing makes me happy.	3.71	0.86
2. Work makes me feel nervous.	3.70	0.82
3. Work makes me feel fidgety.	3.67	0.90
4. Work makes me feel hopeless.	3.65	0.91
5. Work frustrates me.	3.61	0.89
6. Work makes me feel worthless.	3.70	0.93
7. Work makes me feel happy.	3.64	0.86
8. Work makes me feel energetic.	3.60	0.87
9. Work makes me feel very happy.	3.67	0.92
10. Work makes me feel calm.	3.65	0.91
11. I am satisfied with my work.	3.67	0.87
12. Work makes me feel full of life.	3.73	0.84
Emotional Disorder (α = 0.96)	3.74	1.18
1. The emotions I display at work are consistent with the true emotional feelings in my heart.	3.71	0.87
2. The emotions I express at work are not consistent with my true feelings.	3.71	0.88
3. The norms of service attitude at work require that the emotions I should express differ from my true feelings.	3.85	0.85
4. Usually, the emotions I express to customers are consistent with my true feelings.	3.70	0.91
5. At work, I often hide my true emotional feelings.	3.77	0.86

for details).

To measure well-being, 12 items were developed based on the work well-being scale developed by Kelloway et al. [43] and modified by Lee [26]. The scale includes 6 items each for positive emotions (questions 7, 8, 9, 10, 11, and 12) and negative emotions (questions 1, 2, 3, 4, 5, and 6). For the positive scale (items 7 to 12), a higher score indicates a higher degree of well-being; for the negative scale (items 1 to 6), a higher score indicates a lower degree of well-being, with reverse scoring for the statistical analysis (see Table 2). To measure emotional disorder, this study used 5 items developed by Morris and Feldman [48]. For the positive scale (items 2, 3, and 5), a higher score indicates a higher degree of emotional disorder, while for the negative scale (items 1 and 4), a higher score indicates a lower degree of emotional disorder, with reverse scoring in statistical analysis (see Table 2 for details).

The survey questionnaire was administered to the respondents from June 10 to July 10, 2019. When participants clicked on the survey link, they first accessed a page that described the study purpose, confidentiality, and privacy protocols. A total of 335 questionnaires were collected, and 6 questionnaires were eliminated because the responses were incomplete. Thus, the final sample size was 329.

3.2. Data Analysis

The data analysis in this study was divided into five steps. First, descriptive statistics were analyzed to determine the characteristics and distribution of the measured variables, and the reliability of the scale was examined. Second, a measurement model with all constructs/dimensions was established using confirmatory factor analysis (CFA) in an attempt to test the fit of the measures. In the next step, a baseline path model was developed using structural equation modeling (SEM) to test H1 to H5. SEM was used to test the relationship between the independent variable and the dependent variable. For mediation to occur, four criteria must be met: (1) the independent variable should be significantly associated with the dependent variable; (2) the independent variable should be related to the mediator; (3) the mediator should be related to the dependent variable; and, (4) the association between the independent variable and the dependent variable must be reduced when the mediator is partially omitted [43]. All variables (surface acting, deep acting, emotional disorder, and well-being) were included in the measurement model (as in Figure 1), and the goodness of fit was tested with SEM to explore the path of influence and the extent to which the dependent variable can be explained. Finally, a structural model was developed, and the regression weights were compared to test the hypotheses.

4. Results

4.1. Respondents’ Sociodemographic Profiles

More than half of the participants (57.1%) were female, and 42.9% were male (Table 1). The mean age and years of working experience of the respondents, respectively, were 32.4 years (SD = 9.8) and 4.0 years (SD = 3.3), respectively. Most of the respondents (67.5%) had a bachelor’s degree or less, and the rest had a master’s degree (19.5%) or a doctoral degree (13.1%). Approximately two-thirds (59.3%) of the respondents were single, and 40.7% were married. In terms of their work experience, more than half of the respondents (57.1%) had three or fewer years of work experience, and approximately one-third (33.7%) had four to six years of work experience.

4.2. Descriptive Statistics and Correlations for the Measured Variables

The Cronbach’s α tests showed a high reliability for the measures of surface acting (α = 0.94) and deep acting (α = 0.88), which were higher than the suggested 0.80 [43]. The assumptions in the Exploratory Factor Analysis (EFA) were met, as the Kaiser–Meyer–Olkin (KMO) values for the analyses were 0.95, 0.85, and 0.91, which were higher than 0.80, and the Bartlett test of sphericity was significant at the 0.001 level. The results of the reliability analysis indicated high internal consistency for destination quality. The AVEs (average variance extracted) for all four factors were 0.67, 0.64, 0.68, and 0.64, which were higher than the suggested value of 0.50 [49]. Thus, the convergent validity of the scale was acceptable. The respondents were likely to agree with the following statements about emotional labor: “To serve customers the right way, I show the right emotions and hide my true feelings” (M = 3.88), “When communicating with others, I adjust and modify my emotions” (M = 3.83), “I try to feel the emotions I need to express toward customers at work” (M = 3.83), and “I try to cultivate the emotions I need to express toward customers” (M = 3.81). All items for emotional labor were above 3.0 on average, which shows that bank workers are essentially putting in such efforts in their customer service roles.

Additionally, the measure of emotional disorder had a high reliability (α = 0.90). The Bartlett test of the sphericity statistic for the well-being scale was 3640.11, with a significance level below 0.001, and the KMO value was 0.97, which showed a high validity (

Table 3. Correlations between Variables.

	Surface Acting	Deep Acting	Emotional Disorder	Well-Being
Surface Acting	1.00
Deep Acting	−0.76 **	1.00
Emotional Disorder	0.69 **	−0.61 **	1.00
Well-being	−0.77 **	0.59 **	−0.84 **	1.00

Note: All factor loadings were significant at the 0.001 levels. * p < 0.05 and ** p < 0.01.

). The respondents were likely to agree with the following items: “The norms of service attitudes at work require me to express emotions that differ from my true feelings” (M = 3.68); “Usually, the emotions I express toward customers are consistent with my true feelings” (M = 3.65); “At work, I often hide my true feelings” (M = 3.64); and “The emotions I express at work are not consistent with my true feelings” (M = 3.57).

The Cronbach’s α tests also showed a high reliability for the measure of well-being (α = 0.96). The Bartlett test of the sphericity statistic for the well-being scale was 3640.11, with a significance level below 0.001, and the KMO value was 0.97, which showed high validity (Table 3). Compared with the scores for the other two scales, the score for the well-being scale was lower. The respondents were likely to agree with the following items: “Work makes me feel fulfilled” (M = 3.71) had the greatest agreement, followed by “Work makes me happy” (M = 3.59), “Work makes me feel nervous” (M = 3.57), and “Work makes me feel restless” (M = 3.56), while the item “I feel frustrated at work” had the lowest level of agreement.

The Cronbach’s α value for each variable was greater than 0.80, which indicated that the variables had a high internal consistency and that the questionnaire had good reliability (

Table 4. Results of the confirmatory factor analysis on emotional labor.

Items	Factor Loading	Error Variances	Ave	Cronbach’s α	Bartlett Sphericity Test (Df)	KMO
Surface Acting			0.67	0.94	2560.58 (55) ***	0.95
To serve customers the right way, I show the right emotions and hide the true feelings.	0.82	0.41
2. At work, when I communicate with others, I display a good mood, even if I do not feel it in my heart.	0.85	0.38
3. Exhibiting appropriate emotions at work is like acting to me.	0.81	0.45
4. I simply pretend to display the emotions I need to show at work.	0.82	0.42
5. I show the emotions I need to at work, but I do not change my true feelings.	0.81	0.44
6. In my work, the emotions I express in front of others are different from the real feelings I feel inside.	0.80	0.44
7. When communicating with others, I adjust and modify my emotions.	0.84	0.39
Deep Acting			0.64	0.88
1. I try to experience the emotions that should be expressed at work rather than just only pretending to feel them.	0.83	0.36
2. I try to express my feelings to customers from my heart rather than only pretending to feel them.	0.77	0.38
3. I try to feel the emotions I need to express toward customers at work.	0.85	0.28
4. I try to cultivate the emotions I need to express toward customers.	0.75	0.40

Notes: 1 All factor loadings were significant at the 0.001 levels. * p < 0.05, ** p < 0.01, and *** p < 0.001. 2 Model fit indices: n = 329, χ2 = 594.995, df = 344, CFI = 0.968, GFI = 0.875, NFI = 0.927, and RMSEA = 0.047, AVE: average variance extracted; KMO: Kaiser–Meyer–Olkin; CFI: comparative fit index; GFI: goodness-of-fit index; NFI: normed fit index; RMSEA: root mean square error of approximation.

). According to the significance level and the KMO metrics, the correlation coefficient matrix and the identity matrix of all variables were significantly different, and the original variable-factor analysis was appropriate. With the principal component analysis, two factors were extracted from the emotional labor scale, one factor was extracted from the emotional disorder scale, and two factors were extracted from the well-being scale.

We analyzed the correlations between the four variables and found the following relationships: surface acting and well-being had a significant negative correlation (r = −0.77, p < 0.01), surface acting and emotional disorder had a significant positive correlation (r = 0.69, p < 0.01), deep acting and emotional disorder had a significant negative correlation (r = −0.61, p < 0.01), deep acting and well-being had a significant positive correlation (r = 0.59, p < 0.01), and emotional disorder and well-being had a significant negative correlation (r = −0.84, p < 0.01) (Table 3). We established a baseline structural equation model below to explore their relations in depth.

4.3. Measurement Model

In the second step, the fits of the measures were assessed using a CFA. In this research, model fits were evaluated by means of the comparative fit index (CFI), the goodness-of-fit index (GFI), the normed fit index (NFI), and the root mean square error of approximation (RMSEA). Based on Byrne’s [50] and Bollen’s [51] suggestions, a model is regarded as acceptable if the CFI exceeds 0.93, if the NFI and the GFI exceed 0.90, and if the RMSEA is less than 0.80.

The CFA results showed that the initial measurement model, consisting of 11 items for two factors (surface acting and deep acting), 12 items for one factor (emotional disorder), and 5 items for one factor (well-being), had acceptable fit indices (n = 329, χ2 = 594.995, df = 344, CFI = 0.968, GFI = 0.875, NFI = 0.927, and RMSEA = 0.047).

As all regression weights were significant (p < 0.001), the measurement model was further refined, as standardized residuals greater than 2.57 are considered statistically significant [47], and large modification indices (those greater than 3.84) are considered statistically significant [43].

Furthermore, the discriminant validity of the measures was assessed by comparing the squares of the correlations between each pair of factors with their AVEs [49]. The correlation coefficients between the four factors ranged from 0.28 to 0.62. Since the AVEs for all seven latent factors were higher than the squares of all correlation coefficients, the discriminant validity was considered to be acceptable. As shown in Table 4;

Table 5. Results of the confirmatory factor analysis on well-being and emotional disorder.

Items	Factor Loading	Error Variances	Ave	Cronbach’s α	Bartlett Sphericity Test (Df)	KMO
Perceived Well-Being			0.68	0.96	1005.66 (10) ***	0.85
1. Work makes me feel terrible. Nothing makes me happy.	0.84	0.39
2. Work makes me feel nervous.	0.84	0.42
3. Work makes me feel fidgety.	0.85	0.41
4. Work makes me feel hopeless.	0.84	0.43
5. Work frustrates me.	0.85	0.42
6. Work makes me feel worthless.	0.85	0.43
7. Work makes me feel happy.	0.80	0.48
8. Work makes me feel energetic.	0.82	0.44
9. Work makes me very happy.	0.84	0.45
10. Work makes me feel calm.	0.83	0.46
11. I am satisfied with my work.	0.82	0.44
12. Work makes me feel full of life.	0.76	0.55
Emotional Disorder			0.64	0.90	3640.11 (66) ***	0.91
1. The emotions I display at work are consistent with the true feelings in my heart.	0.83	0.43
2. The emotions I express at work are not consistent with my true feelings.	0.77	0.60
3. The norms of service attitudes at work require me to express emotions that differ from my true feelings.	0.84	0.39
4. Usually, the emotions I express toward customers are consistent with my true feelings.	0.82	0.46
5. At work, I often hide my true feelings.	0.74	0.62

Notes: 1 All factor loadings were significant at the 0.001 levels. * p < 0.05, ** p < 0.01, and *** p < 0.001. 2 Model fit indices: n = 329, χ2 = 594.995, df = 344, CFI = 0.968, GFI = 0.875, NFI = 0.927, and RMSEA = 0.047, AVE: average variance extracted.

, the composite reliability values for surface acting, deep acting, emotional disorder and well-being were 0.94, 0.88, 0.96, and 0.90, respectively. All the values were higher than the suggested value of 0.80 [47]. The AVEs for all seven factors were equal to (for employee characteristics) or higher than the suggested value of 0.50 [49]. Thus, the convergent validity of the scale was acceptable.

4.4. Baseline Model

In the next step, H1–H5 were tested by establishing a baseline structural model (n = 329) that included emotional labor as an exogenous variable and emotional disorder and well-being as endogenous variables. To simplify the process of the data analysis, the factor scores for the four factors (surface acting, deep acting, emotional disorder, and well-being) were used in the process of the structural equation modeling. As shown in

Table 6. Results of SEM.

Effects	Direct Effects	Indirect Effects	Total Effects
H1: Surface Acting → Well-being	−0.404 ***	−0.024	−0.428
H2: Deep Acting → Well-being	0.085 ***	0.092	0.177
H3: Surface Acting → Emotional Disorder	0.173 **		0.173
H4: Deep Acting → Emotional Disorder	−0.078		−0.078
H5: Emotional Disorder → Well-being	−0.137 *		−0.137

Note: 1 * p < 0.05, ** p < 0.01, and *** p < 0.001; 2 Model fit indices: n = 330, χ2 = 594.995, df = 344, CFI = 0.968, GFI = 0.875, NFI = 0.927, and RMSEA = 0.047.

and Figure 2, the baseline structural model had acceptable fit indices (n = 329, χ2 = 594.995, df = 344, CFI = 0.968, GFI = 0.875, NFI = 0.927, and RMSEA = 0.047). The SEM results revealed that surface acting had a significant direct effect on emotional disorder (β = −0.404; p < 0.001) and well-being (β = −0.173; p < 0.01), so H1 and H3 were supported. Deep acting had a significant direct effect on well-being (β = 0.085; p < 0.001), supporting H2, and emotional disorder had a significant direct effect on well-being (β = 0.137; p < 0.01), supporting H5. However, deep acting had no significant direct effect on emotional disorder (β = −0.078). These findings supported H1, H2, H3, and H5.

5. Discussion and Conclusions

In conclusion, the purpose of this study was to investigate the relationship between emotional labor and well-being among the front-line staff who have the most frequent customer contact in the banking field and to investigate the moderating effect of emotional disorder. Our results indicate that emotional labor has a significant impact on employee well-being: surface acting has a significant negative impact on employee well-being, while deep acting has a significant positive impact on employee well-being. Moreover, emotional disorder played a mediating role between emotional labor and employee well-being, and emotional disorder was positively correlated with surface acting and negatively correlated with deep acting.

5.1. Discussion and Implications

These research findings have the following theoretical implications. First, most previous papers have discussed the relationship between emotional labor and employee well-being, whereas the current study explored the mechanism of the mediating role of emotional disorder, representing a useful addition to the relevant theoretical research. According to the COR theory, people are eager to retain the resources they value for their own individual development, and one of these resources is energy, including emotion, attraction, passion, and so on [46]. When an individual’s psychological resources are consumed without timely replenishment, emotional disorder and low performance can result [16]. Specifically, the present research demonstrated the intrinsic mechanism through which emotional disorder plays a mediating role between emotional labor and well-being, at least among the front-line bank staff in the context of China. The research findings also suggest that employees and organizations should pay more attention to psychological resources and to their replenishment, since doing so can reduce emotional disorder and improve well-being. In this sense, future studies can find out how to increase and replenish psychological resources of employees so as to improve their well-being. Researchers can consider and incorporate the variables, such as individual traits, organizational support, as well as explore their influencing mechanism, in order to design and provide more accurate employee assistance programs.

This research also clarified the underlying mechanism regarding the two types of emotional labor and their different effects on employee well-being. While previous studies of emotional labor demonstrated that deep acting was much more valuable than surface acting [46,52,53], the present research explored the underlying mechanism for this relationship. Employees’ emotional labor can cause psychological fatigue, which consumes psychological resources [9]. If employees’ emotional experiences are contrary to what is required of them, they will experience a cognitive imbalance. They will try to compensate for this inappropriate emotional expression, making them more prone to emotional imbalance [17]. Without timely replenishment, employees experience high work pressure, low self-efficacy, and emotional exhaustion, leading to job burnout and low job satisfaction. Deep acting is a means of obtaining resources that can help employees adjust their internal and external emotions and make them more consistent, thus reducing emotional disorder and facilitating sincere and enthusiastic communication with customers, which can improve their job satisfaction and well-being. By contrast, surface acting was not helpful for replenishing resources, leading to low well-being.

Third, our results proved that deep acting had a negative impact on employees’ emotional disorder, which played an important role in improving well-being. The front-line bank staff are bound by established organizational rules, such as “service with a smile,” and they must spend a certain amount of psychological resources expressing the emotions required by the organization. Employees are accustomed to using surface acting, especially in the Chinese culture, which encourages implicit and mild expressions and the concealment of one’s true feelings. China was ranked number 93 for overall well-being in 2019 despite its strong economic strength, and one cause may be overly repressed feelings without good emotion regulation [54,55].

Lastly, this study responded to calls for an improved understanding of the context specificity of the originally Western construct of emotional labor and of the extent to which it may be generalizable to non-Western settings, such as China. To help fill this gap in the literature, this paper reports the results of a survey of front-line bank staff carried out in China and found that surface acting reduced their well-being, a relation that was lacking in the Western context. Individualism encourages outward displays of emotion in Western culture [56], whereas the group harmony is more valued in the Chinese collectivistic culture than an individual’s expression of emotions [57,58]. Because of the collectivistic culture, Chinese workers often manage and moderate emotions, and in some cases, they repress emotions, which leads to a greater level of anxiety and fear [59,60,61]. For example, when facing conflicts with customers in the service setting, workers are prone to depress their anger in order to compromise. In these situations, emotional disorder may occur more often, thus worsening their well-being. Therefore, developing deep acting skills is very important for the front-line bank staff in China.

Our study results have also provided valuable practical implications. The study results suggested that employees should be encouraged to use more deep acting, maintain consistency between their internal and external emotions, and communicate with customers in a positive manner. Thus, commercial banks should consider implementing employee assistance plans at the organizational level to help employees with disordered thinking to change their views on work. Given that deep acting had a negative impact on employees’ emotional disorder, organizations should provide much more assistance to front-line bank staff, such as training in deep acting or building a psychological center for employees to seek help from counselors. These practices can help employees alleviate negative emotions and approach their work with a positive and optimistic attitude. Their emotional disorder may be reduced, and their well-being will naturally improve. In addition, the human resource departments of banks should match positions according to the characteristics of each employee so employees can better adapt to their positions.

5.2. Limitations and Future Research

This research has several limitations and recommendations for future research. First, this research focuses on the mediating effect of emotional disorder on the relationship between the emotional labor and well-being of the front-line bank staff in the context of China, so its findings may not be generalizable to other countries. Therefore, future research could examine the same research question in other contexts, such as Western culture. Second, the research found that emotional disorder is an important mediator between emotional labor and well-being, but perhaps there are other organizational and individual variables that influence this relationship. Therefore, future research should incorporate other variables, such as individual traits and organizational support, to more deeply explore the mechanism of action between emotional labor and well-being.

In addition, in the future, using the physiological indices of the human body may be a more reliable measure for emotional disorder. Lastly, this study used the self-report method to collect data on emotional labor and disorder; thus, the results may be challenged by potential bias. Given that emotions are mentally represented in the conscious mind, and individuals can communicate these representations using verbal language [62], it is not surprising to employ the self-report method, which is considered as a valuable and irreplaceable source of information about emotions [63]. Additional efforts were made to reduce potential self-report bias. For example, survey questions were kept short and clear, in order to allow participants to easily understand the questions. Participants were also asked to reflect their most recent experiences when answering questions on emotional labor and disorder, in order to keep the time frame short and thus reduce response bias.

In conclusion, the results showed that using emotional labor influenced bank employees’ perceptions of well-being. Surface acting has a significant negative impact on employee well-being, while deep acting has a significant positive impact. Also, emotional disorder was revealed to play a role in mediating emotional labor and employee well-being; it was positively correlated with surface acting and negatively correlated with deep acting. The results suggested that developing deep-acting skills is important for increasing front-line bank staff’s well-being in China, who are accustomed to repressing their emotions, and emotional disorder might occur more often than has been previously believed, which worsens their well-being.