The moderating role of gender in the effect of self-monitoring personality trait on emotional labor strategies

Abstract

This study investigated the effect of self-monitoring on emotional labor strategies (surface acting and deep acting) and the moderating role of gender in development. In anchoring the relationship between self-monitoring and emotional labor, the conservation of resources theory was used, while the moderating role of gender was based on sociocultural theory. A sample of 312 employees from the health and communication sectors provided the analyzed data. Data analysis was based on regression analysis (complemented with the PROCESS tool). The results showed positive self-monitoring (surface and deep acting) effects on emotional labor. More so, gender was found to moderate the impact of self-monitoring on surface acting strategy but not on deep acting strategy. The practical implication of this study is to provide information on the value of self-monitoring in improving, especially in men, the ability to express emotional labor.

Keywords:

• self-monitoring
• emotional labor
• surface acting
• gender
• conservation of resources theory

PUBLIC INTEREST STATEMENT

Emotional labor as a two-dimension variable of surface-acting and deep-acting strategies is gaining empirical confirmation as a valuable factor in individual performance and, consequently, effective organization functioning. Emotional labor refers to the effort, planning, and control required to express the most needed emotions during interpersonal communications. To fully exploit the value of emotional labor, its dispositional antecedents need to be identified. Therefore, this paper examined the usefulness of self-monitoring personality traits in predicting emotional labor. Self-monitoring refers to the extent to which individuals are willing and able to engage in the expressive control of their self-presentation in public. Self-monitoring improves, essentially among men, the ability to express emotional labor; therefore, it should be considered in the recruitment, selection, and placement of service employees. Since personality and emotions are widely discussed to differ based on gender, this study also examined whether the impact of self-monitoring on emotional labor is different for males and females. The results of this study offered relevant insight into human resource practices.

1. Introduction

Any organization’s behavior as expressed through the workforce is substantially determined explicitly or implicitly by the necessity to meet the needs and expectations of its stakeholders, especially the customers. In the literature, such demand on the employee, essentially related to the customers (M.-S. Lee & Kim, 2020), is widely termed emotional labor. Emotional labor is defined as the effort, planning, and control required to express the most needed emotions during interpersonal communications (McShane & Von Glinow, 2018). In the work environment, emotional labor (EL) has been identified to have beneficial and damaging effects on employees and organizations (Eneizan et al., 2019; Ogunsola et al., 2020; Yao et al., 2019; Zhao et al., 2020). However, the existing literature concludes that many of the adverse effects of EL are on the individual (Jeung et al., 2018; Nikmah & Sulistyarini, 2017; Ogunsola et al., 2020; Yao et al., 2019). Compared to other studies, positive effects are substantial for the organization (Eneizan et al., 2019; Kamassi et al., 2019; Tunç et al., 2019; Zhao et al., 2020). The literature indicates that the emotional labor strategy determines the outcome of the EL for the employee or the organization. The surface acting strategy is more associated with a negative consequence for the individual, while the deep acting strategy is more related to positive results (Yao et al., 2019; Yoo & Jeong, 2017).

The accumulating findings that EL has many negative consequences for the individual called for an inclusive search for EL antecedents. However, the literature on EL is building vigorously; several studies exist (see, e. g., Karimi et al., 2020; Xu et al., 2020; Yeh et al., 2020) in its antecedents. However, the literature has not substantially examined self-monitoring personality traits as an antecedent of EL. Self-monitoring (SM) describes how individuals monitor and modify how they present themselves verbally and nonverbally to others (Gulliford et al., 2019). The lack of substantial research on SM in the EL nomological network is a gap in the EL literature for some reasons. Conceptual SM and EL share specific characteristics in common. They essentially represent individual acts that often do not agree with their internal state. They serve as adaptive mechanisms in a given setting, are externally induced from the individual, and represent the capability to understand situational clues and act as desired. These features point to the possible interaction between SM and EL. Despite this, only Scott et al. (2012) and Huang et al. (2019) had examined SM as an EL antecedent. And this implies the absence of study on the relationship between SM and EL.

Furthermore, several personality traits, mainly those associated with the Big Five model (Khalil et al., 2017; Yeh et al., 2020) and related variables, such as emotional intelligence (Hjalmarsson & Dåderman, 2020; Pindar et al., 2020), are examined along with EL, and valuable insights have emerged from such studies. The valuable results support the need for an inclusive search for the precursor of EL. Consequently, this study examines the effect of SM on EL. Several researchers (e.g., L. Lee & Madera, 2019; M.-S. Lee & Kim, 2020) have noted that existing studies have failed to explore the mechanism in the relationship between EL and its predictors. And in their separate literature, SM and EL are widely influenced by demographic variables such as gender (Alberto, 2020; Cetin et al., 2018; Dahling & Perez, 2010; Frazier & Fatis, 1980; Lammers, 1991; Rasheed-Karim, 2020). Such findings indicate that gender is a common cause of SM and EL. The usual cause status implies that demographics can influence when and how SM relates to EL. And some related studies (e.g., Cheung & Tang, 2010; Javed & Batool, 2021; Scott & Barnes, 2011) have assessed and identified gender as a moderating variable. Consequently, this study also examined the moderator role of gender in the effect of SM on EL.

2. Theoretical review and hypotheses development

2.1. Self-monitoring personality traits and emotional labor

Self-monitoring is a personality trait that refers to the extent to which individuals are willing and able to engage in the expressive control of their self-presentation in public (Wilmot, 2020). The underlying postulation of the self-monitoring concept is that individuals differ in how they monitor (that is, observe, regulate, and control) their public appearance, which they exhibit in the social milieu, and create and manage their interpersonal relationships. SM is generally classified in terms of high and low. High self‐monitors regulate the public presentation of their expressive behavior to fit the situational context. High-monitors regulate their expressive self-presentation for the sake of desired public appearances and are highly responsive to social and interpersonal cues of situational appropriateness. Low self-monitors seem less willing and able to engage in such appearances. Low monitors remain true to their internal dispositions and attitudes, displaying behavioral consistency across situations (Charlier, 2020; Westfall, 2020). Besides the classification into high and low, SM is widely discussed as a two-component variable of ability to modify self-presentation and sensitivity to expressive behavior of others (Lennox & Wolfe, 1984).

EL refers to the effort, planning, and control required to express the most needed emotions during interpersonal communications (McShane & Von Glinow, 2018). It is essentially the process of managing feelings and expression to achieve the required emotion needed in a job. The emotion displayed leads to the achievement of expected performance (Ogunsola et al., 2020). A two-dimensional model of EL strategies (surface and deep acting) dominates the empirical literature (Kamassi et al., 2019). Surface acting represents displaying emotions that one does not feel to others. The surface acting would involve a change in observable features, such as gestures, facial expressions, or vocal tones. On the contrary, deep acting involves changing their noticeable features and latent inner states (Yao et al., 2019). Deep acting is managing one has feelings, including efforts to truly change one’s emotional state to match the emotions required by the job (Landy & Conte, 2019).

A review of the empirical literature indicates the absence of studies on the relationship between SM and EL, except Scott et al. (2012) and Huang et al. (2019). The former authors reported that SM influences the level and variability of surface acting and moderates the effects of surface acting variability on job satisfaction and work withdrawal. The latter authors observed a positive relationship between SM and EL and a stronger relationship between SM and surface acting. Although there is an apparent scarcity of studies on SM and EL, there are a few studies in other models of personality and EL. Similarly, several studies are also available on SM and some other emotions. These studies have produced findings that provide insight into personality and emotion. For example, there is a negative relationship between workplace resilience and surface acting strategy and a positive association between resilience and frequency of emotional labor (Delgado et al., 2020). Several Big Five personality traits moderate the relationship between surface acting and sleep problems (Yeh et al., 2020). A positive personality relates to deep acting, while a negative personality relates to surface acting (Khalil et al., 2017). Surface acting was significantly predicted by neuroticism, and deep acting was considerably predicted by openness to experience (Basim et al., 2013). In Mróz and Kaleta’s (2016) study, neuroticism and extraversion found correction with emotional labor. Conscientiousness is positively related to deep and surface acting (Diefendorff et al., 2005). Extraversion, conscientiousness, agreeableness, and openness positively influence emotional labor (Sohn & Lee, 2012). And core-self evaluations had both direct and indirect effects on depersonalization through employees’ perceptions and reactions to emotional labor (Pujol-Cols et al., 2021).

The conservation of resources (COR) theory explained the plausible positive relationship between SM and EL. The COR theory (Hobfoll, 1989) proposes that employees are motivated to have, sustain, and maintain resources, particularly when faced with work demands. Resources are anything a person values (Joyner & Leake, 2018). SM is a resource, while EL is a work demand. SM as a personal characteristic is a resource that helps employees engage in effective EL. Some studies (see, e.g., Gursoy et al., 2011) adopted COR theory as a framework to investigate the effect of personality on emotional labor. Therefore, it was hypothesized that:

H1: SM positively predicts surface acting strategy

H2: SM positively influence deep acting strategy

2.2. Gender as moderators in the SM and EL relationship

Demographic differences in personality, emotion, and behavior are well documented in the theoretical and empirical literature. Sociocultural theory (Wood & Eagly, 2012) explained gender differences in personality and emotion. This theory proposes that gender differences arise from social, cultural, psychological, and other environmental forces. Like other gender theories, sociocultural theory acknowledges the roles of biological and learned influences in gender differences. Studies have shown that female workers exhibit a more intensive EL than men (Cetin et al., 2018). Gender and age affected the frequency and kind of emotional labor experienced (Rasheed-Karim, 2020). Gender, seniority, and institution type impact the emotional labor process, including strategies and how others received and responded to the strategy (Mahamad, 2018). The gender and position of the employees moderated the effect of emotional intelligence on emotional labor (deep acting and surface acting), and surface acting was significantly higher for female employees than for male employees (Jung & Yoon, 2014). Men tend to score higher on SM measures than women (Day et al., 2002; Westfall, 2020; Wilmot, 2011). The review indicates that males and females differ in their expression of SM and EL. Therefore, it was hypothesized that:

H3: Gender moderates the positive effect of SM on surface acting strategy

H4: Gender moderates the positive impact of SM on deep acting strategy

3. Method

3.1. Sample and design

The sample comprises 312 employees drawn from the health and communication sectors in Delta State, Nigeria. Two hundred and ten participants came from the health sector, while one hundred and two participants were drawn from the communication sector. Participants were sampled from public and private organizations. The sample size met the maximum sample-to-item ratio of 5–1 and the sample-to-variable ratio of 20-to-1 (Memon et al., 2020; Tsang et al., 2017). Participants comprise 61.2% females, 63.8% married, 84.2% hold a first degree and lesser certificates, while 15.8% have certificates above the first degree. The participants’ age means was 34.27 years (SD, 9.92). All employee categories were represented in the sample.

The study design was cross-sectional, as data was collected at one point in time. A convenience sampling technique was adopted to distribute the questionnaires as organizations and participants were used based on availability. Nonrandom samples are a common feature in organizational studies, particularly in this research location, where sampling frames are often unavailable or extremely difficult to access. Regression analysis (complemented with Hayes’s PROCESS tool for SPSS) was used to test the hypotheses. Regression is a statistical tool for data analysis was appropriate as this study predicted EL from SM. More so, the PROCESS tool for SPSS is gaining much popularity among researchers and has been adjudged “the best way to tackle moderation and mediation” (Field, 2018). Regression is a parametric statistical test; therefore, the design and preliminary data analysis observed several assumptions associated with its usage. For example, the collected data were independent, which met the separate response requirement. The adopted Likert scale format met the demand for interval scaling, and scatter plots on the data revealed that the variables were linearly related. Data were analyzed with SPSS version 27.

On verbal approval of the participating organizations and some organizations’ administrative staff, the research questionnaires were distributed to the participants at their workplaces. In total, 360 questionnaires were distributed; within six weeks, 320 filled questionnaires were received. On physical examination of the returned questionnaires, it was observed that eight copies did not have sufficient information. Therefore, data analysis was conducted on responses from 312 participants. The response rate is considered satisfactory (Khalil et al., 2017).

3.2. Measures

3.2.0.1. Self-monitoring measure

Lennox and Wolfe’s (1984), pp. 13-item scale assessed SM. The scale has two dimensions (the ability to modify self-presentation and the sensitivity to the expressive behavior of others). The former dimension has seven items, while the latter has a six-item one. However, in this study, SM was analyzed as a one-dimensional variable because the two dimensions complement each other. Complementary characteristics are reflected in the significant positive correlation coefficient obtained on the two dimensions. Lennox and Wolfe (1984) reported a coefficient alpha of 0.77 for the seven items, 0.70 for the six items, and 0.75 for the total scale. Other scale users (e.g., Gulliford et al., 2019) reported satisfactory psychometric properties.

3.2.0.2. Emotional labor measure

Yang et al.’s (2019), pp. 12-Item scale on EL was adopted. The scale assessed four EL strategies that covered surface acting (3 items), deep acting (3 items), expression of naturally felt emotions (3 items), and emotion termination (3 items). The authors reported good psychometric properties on the scale. However, this study did not adopt natural emotions and emotion termination dimensions. The dimension of naturally felt emotions does not represent emotional labor, as it does not require effort or involve emotional strain on the employee. The emotion termination dimension seems to be unique to the Chinese context, as it is absent in all other extant EL scales. It essentially accounts for the dominant adoption of surface acting and deep acting strategies in the extant literature (see, e.g., Kamassi et al., 2019; Ogunsola et al., 2020).

Data analyzes were conducted on the mean scores of the participants. Data were collected at the individual level, and all items were scored on a 5-point Likert scale (strongly agree = 5, agree = 4, undecided, = 3, disagree = 2, and strongly disagree = 1), which is less confusing to respondents and help increase the response rate. The mean score was obtained by dividing the total score on each scale by the number of valid responses. This procedure yielded scores that ranged from 1 to 5.

3.2.0.3. Common-method variance

Data were collected with self-report measures, some common-method variance (CMV) control procedural measures were infused into the study design. The procedures include ensuring participants’ anonymity and confidentiality, which attenuate evaluation apprehension and consequently improve honesty in response (Rodríguez-Ardura & Meseguer-Artola, 2020; Tehseen et al., 2017). The focal variables (SM and EL strategies) were presented to the participants on a separate sheet of paper. It achieved a physical gap that cuts the flow of thought from one variable to another. Employees self-evaluated SM and EL, as they are subjective and abstract experiences that others cannot easily and objectively measure. Only the individual directly involved can tell the “truth” (Kamassi et al., 2019). Therefore, the employees’ self-report of these variables was considered appropriate.

4. Results

4.1. Reliability and validity

Two statistical procedures (the Harman single-factor test and the correlation matrix) were applied to the data. The procedural control of common-method variance incorporated in the design and statistical procedures (the Harman single-factor test and correlation matrix) was applied to the data. From the Harman single-factor test, the factors with an eigenvalue equal to 1 and explained above 74.08% of the total variance. However, the first factor explained 15.63% of the total variance. Since the variance explained by the first factor did not represent most of the total variance, there is a weak CMV in the data. More so, assessing from the correlation matrix procedure, CMV was not an issue with the data as correlations between the variables examined were considered within the acceptable range (Martínez-Córcoles & Zhu, 2020; Rodríguez-Ardura & Meseguer-Artola, 2020). Cronbach’s alpha was used to test the internal consistency reliability of the measures. The coefficients obtained are presented in Table 1. Alpha statistics within the range of 0.75 and 0.91 are widely considered satisfactory (Field, 2018). Their acceptance and adoption established the face validity and content validity of the measures in the literature (Karimi et al., 2020; Mirjana et al., 2018). Construct validity, an estimate of the extent to which variance in the measurement reflects variance in the underlying construct (Jha et al., 2019), was tested with convergent validity and discriminant validity. The twin aspects to the validity assessment. Cronbach’s alpha statistics obtained in this study offered support for the convergent validity of the measures (Field, 2018). Discriminant validity was tested with a factor structure (cross-loading). It was observed that for the items, the loading in their construct was higher than their cross-loading. It supports discriminant validity for the measures, as such a pattern of item loading indicates that they belonged exclusively to their factors (Makhijaa & Akbarb, 2019). The obtained Durbin-Watson test statistics were 1.98 and 1.76, acceptable for autocorrelation. The variance inflation factor (VIF) statistics were below 10, while the tolerance statistics were above 0.2. The two sets of statistics indicate the absence of collinearity in the data sets (Field, 2018).

Table 1. Mean, the standard deviation on the researched variables

	Gender	N	Mean	Sd
Self-monitoring	Males	118	3.76	.62
	Female	191	3.71	.64
	Males & Females	309	3.73	.63
Surface-acting strategy	Male	110	2.68	1.01
	Female	186	2.45	.96
	Males & Females	296	2.54	.99
Deep-acting strategy	Males	110	3.30	1.11
	Females	185	2.95	1.06
	Males & Females	295	3.08	1.10

Table 1 shows the mean, standard deviations for the variables studied. The means observed on a five-point Likert scale format could be adjudged moderate. The SM mean score for men was (3.76, SD, 0.62); for women, it was (3.71, SD, 0.64). The surface acting strategy means that men’s scores were (2.68, SD, 1.01) and women (2.45, SD, 0.96). The deep acting strategy means that the score for males was (3.30. SD, 1.11) and for females was (2.95, SD, 1.06). The t-test statistics showed a significant difference between men and women in the surface acting strategy (t = 2.02 (df, 307; p < 0.05), but not in SM (t = 0.68 (df, 294; p > 0.5), and the deep acting strategy (t = 2.72 (df, 293; p > 0.05).

Table 2 shows zero-order correlations on the variables under study. The zero-order correlation revealed a significant positive relationship between the variables in the study. However, the highest relationship was between surface and deep acting strategies. The modest zero-order correlation between the predictors and the criterion variables indicates the absence of multicollinearity in the model.

Table 2. Alpha and zero-order correlations of the researched variables

		Alpha	No. of item	1	2	3	4
1	Gender			1
2	Age			.12*	1
3	Self-monitoring	.75	13	−.03	−.09	1
4	Deep-acting strategy	.83	3	−.14*	−.05	.19**	1
5	Surface-acting strategy	.91	3	−.12*	−.30**	.13*	.42**

** <.001; * < .05

4.2. Test of direct effect

Table 3 shows a simple regression analysis predicting SM’s two EL strategies (surface and deep acting). The statistics in the table offered support for hypothesis 1 and hypothesis 2. For hypothesis 1, SM significantly predicted surface acting, (β = 0.20, 95% CI [0.02 0.37], t = 2.26, p < 0.05). The observed B-value indicates that a one-unit increase in SM brings about a 0.20-unit increase in surface acting. R² indicates that SM accounts for approximately 2% variance in deep acting, and R² of 0.017 indicates a small effect. The analysis of variance test, F (1; 297) = 11.56, p < 0.001, indicated that the regression was statistically significant, meaning that El can be predicted from SM (good model). The small difference between R² (0.017) and adjusted R² (0.014), which is 0.003, indicates good cross-validation; this model can be applied to other samples from the same population.

Table 3. Simple regression on self-monitoring and the EL strategies

	Surface-acting						Deep-acting
	B	se	β	P	R^2	Adj R^2	B	se	β	P	R^2	Adj R^2
Constant	1.88	.36		.001			1.79	.33		.001
SM	.20	.08	.13	.02	.017	.014	.32	.09	.19	.001	.038	.034

For Hypothesis 2, SM significantly predicted deep acting (β = 0.19, 95% CI [0.14 − 0.52], t = 3.40, p < 0.001). The observed β-value indicates that a one-unit increase in SM brings a 0.32 unit increase in deep acting. The R² also indicates that SM account for about 4% variance in deep acting, and the R² of 0.034 indicates a small effect. The analysis of variance test, F (1; 298) = 5.12, p < 0.05, indicates that the regression was statistically significant, which means that El can be predicted from SM (good model). The small difference between R² (0.038) and adjusted R² (0.034), which is 0.004, indicates good cross-validation; this model can apply to other samples from the same population.

4.3. Test of the moderation effect

Table 4 shows statistics on the test of Hypotheses 3 and 4. Hypothesis 3 was supported as the interaction of SM and sex on surface acting was statistically significant (β = 0.40, p < 0.05). However, Hypothesis 4 was not supported as the interaction of SM and gender on deep acting was not statistically significant (β = 0.33; p > 0.05). Statistics from the conditional effect of SM on the surface acting strategy show that when gender is high (male), there is a significant positive relationship between SM and the surface acting strategy (β = 0.47, 95% CI 0.18, 0.75, t = 3.16, p < .001), and that when gender is low (females) there is a nonsignificant positive relationship between SM and the surface acting strategy, (β = 0.07, 95% CI {-0.15, 0.29}, t = 0.63, p > 0.53). Figure 1 is a simple slope analysis on the moderating role of gender in the effect of SM on the surface acting strategy.

Table 4. Analysis of gender as a moderator in the effect of SM and EL strategies

		Surface-acting				Deep-acting
		B	Se	T	P	B	se	T	p
Constant		2.53{2.42;2.64}	.06	45.17	001	3.08{2.96;3.20}	.06	49.90	001
SM		.21{.04;.39}	.09	2.45	.01	.32{.13;.51}	.10	3.28	001
Gender		.24{-.47; −.02}	.12	−2.11	.04	−.33{-.58; −.08}	.13	−2.57	.01
SM × Gender		.40{-.76; −.04}	.18	−2.20	.03	.33{-.08;.71}	.20	1.58	.11
	R = .22; R^2 = .05					R = .26; R^2 = .07	R = .26; R2 = .07

Figure 1. Conceptual model.

Figure 2. Interaction between self monitoring and gender in predicting surface acting strategy.

5. Discussion

This study investigated the effect of SM on EL strategies (surface and deep acting) and the moderating role of gender in effect. Descriptive statistics revealed moderate SM, deep acting, and surface acting strategies and a statistically significant zero-order correlation between the variables. The highest correlation coefficient among the variables studied was between the deep and surface acting strategies. The observed coefficient (0.42) suggests that the dimensions relate positively. The significant positive correlation possibly implies that individuals engaged in deep acting are more likely to engage in surface acting and vice versa. The coefficient’s R² (0.12) indices indicate a medium effect size. It suggests that the degree of relationship between the two EL strategies is practical importance. The descriptive statistics showed that the participants were moderate in the expression of SM and EL. A plausible explanation for this observation is that although expressing SM and EL may be required by the sampled organizations, how much they are described is mainly within the control of the individual. Since extreme expression of the two forms of behavior could be harmful to social and psychological well-being, individuals are likely to keep them at the restrained level. However, male and female participants recorded higher mean scores for SM, surface acting strategies, and deep acting strategies. For SM, the observation is consistent with the results of Day et al. (2002), Westfall (2020), and Wilmot (2011). However, for EL, the observation was inconsistent with Cetin et al. (2018) and Jung and Yoon (2014). Furthermore, analyses show that men and women differ significantly in the surface acting strategy but not in the SM and deep acting strategy. The statistics showing significant gender differences in surface acting and nonsignificant differences in deep acting are possible because individuals are more likely to be aware of their engagement of surface acting strategy than their deep acting strategy.

The model in which SM positively and significantly predicts the surface acting strategy was good, which supported hypothesis 1. SM positively influencing surface acting is expected and consistent with the related literature (e.g., Khalil et al., 2017; Scott et al., 2012; Yeh et al., 2020). Similarly, the model that SM positively and significantly predicts deep acting strategy was good, which supported Hypothesis 2. This finding has been supported in related existing related studies (e.g., Basim et al., 2013; Sohn & Lee, 2012). As noted in the Introduction, conceptually, self-monitoring and EL share specific characteristics in common. They primarily represent individual acts that are essentially make-up. They serve as an adaptive mechanism in each setting. They are externally induced from the individual and embody the capability to understand situational clues and act as desired. Since the two variables share much in common, they interact much like one. The similarity is an essential factor in attraction (Morry, 2005).

Hypothesis 3, which tested for the moderating role of gender in the effect of SM on surface acting, was supported. A few related studies (e.g., Jung & Yoon, 2014) reported similar results. This finding indicates the amount of SM trait an individual possesses, and the gender of the individual determines the level of surface acting strategy expressed. SM and gender worked together to determine the degree of surface acting strategy in an individual. So understanding how much an individual can be self-monitored is not sufficient to know how much surface acting the individual would express. Another implication of this finding is that any effort to improve surface action among employees through SM would require taking gender into cognizance. One possible explanation for why gender moderates the relationship between SM and the surface acting strategy and not the deep acting strategy is because the surface acting strategy operates within the level of consciousness. The deep acting strategy is used more at the unconscious level and much more outside the individual control. Thus, it is more amenable to the control of the individual than the deep acting strategy (Morsella & Poehlman, 2013).

Furthermore, the interaction between SM and the surface acting strategy differs for men and women. A high level of SM leads to a significant increase in the number of surface-acting strategies for men and a nonsignificant increase in the number of surface acting strategies for women. A possible explanation for this observation is that the link between personality and behavior would be more vital for men than for women. The former group is more at liberty to express its individuality than the latter group (Furnham & Henderson, 1981; Naveed et al., 2020).

Hypothesis 4, which tested for the moderating role of gender in the effect of SM on deep acting, was not supported. It implies that SM interaction with males or females does not affect how deep acting is expressed. It is unexpected as several studies (Cetin et al., 2018; Rasheed-Karim, 2020; Westfall, 2020; Wilmot, 2011) have reported gender differences in SM and EL. This result has a foundation in the nonsignificant difference between men and women in SM, as revealed in the t-test analysis.

5.1. Contribution

In several ways, this study made theoretical and practical contributions. The existing literature lacks studies on SM and EL. Consequently, this study pioneered an investigation into the relationship between the two variables. It is vital as SM has a much theoretical connection with EL. Existing studies on other personality traits, such as the five big models with EL, have provided insight that is of great value. Second, this study examined the interaction among the three variables of different domains: demographic, personality, and emotion. An interactionist approach to knowledge has received substantial recommendation and confirmation. The findings that SM positively predicts EL add to the extent of studies that have offered support and confirmation to the conservation of resources theory, which proposes that employees are motivated to have, sustain and maintain resources, particularly when faced with work demands. This study also builds empirical evidence that gender is not a differentiating factor in SM. Yang et al. (2019) emotional scale adopted in this study is relatively new in the literature. It has not been widely and substantially subjected to the validity and reliability of psychometric tests. Consequently, this study has contributed to the literature on the validity and reliability test of the measure.

The results of this study have some practical implications. SM positively influenced surface and deep acting strategies. The finding indicates that SM contributes to the ability of employees to express EL, and in various discussions, EL is presented as a necessity for service organizations. SM enhancing expression of EL has implications in making EL less laborious and, therefore, in minimizing the negative consequences of EL on the individual. Therefore, in-service organizations, the process of recruitment, selection, and placement of employees, essentially those in face-to-face interaction with customers, should recognize the personality traits of SM. Several accurate scales can help to achieve the process. Additionally, several programs, such as emotional intelligence training, and sensitivity training, should be made available to service employees to enhance SM ability. Therefore, service organizations must apply a mixture of the two approaches.

5.2. Limitations and suggestions for future research

One limitation of this research is adopting a single source of data collection and cross-sectional design. There is the possibility of some common-method bias impacting the results despite the precaution in design. There is also the likelihood of reversed causality, as cross-sectional research design has no control over that. Therefore, multisource and longitudinal research designs are recommended in future research. In this study, the effect of SM on EL was examined with Grandey’s (2003) two-dimensional (surface and deep acting) model of EL. Future studies should examine SM and other EL models (see, e.g., Morris & Feldman, 1996). The moderating of gender in the effect of SM on EL was contrary to expectation. Since this study is a pioneer in SM and EL relationship, more studies on the two very are recommended

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The authors received no direct funding for this research.

Notes on contributors

Sunday Samson Babalola

Sunday Samson Babalola is a Professor of Management with over 20 years in teaching, research, and administration in higher educational institutions. His focus is on applying psychological processes to diverse work behaviors with contributions to theory building and methodology. His research benefited from a solid background in quantitative analysis, with working knowledge in qualitative analysis.

Chiyem Lucky Nwanzu

Chiyem Lucky Nwanzu is a Ph.D. holder in Industrial/Organizational Psychology and a Senior Lecturer in the Department of Psychology, Delta State University, Abraka, Nigeria. His research interests are workplace attitude and behavior, organization sustainability, and change management.