Introduction

Emotional labor refers to the process by which employees are required to regulate and manage their emotions as part of their job, in accordance with organizationally prescribed rules and guidelines [1]. Employees typically adopt emotional labor strategies which are generally categorized into surface acting (i.e., faking or suppressing true feelings) and deep acting (i.e., attempting to genuinely feel the required emotion) [1]. Research has demonstrated that consistently suppressing one’s true emotions to display organizationally desired ones can lead to increased physiological and psychological stress responses [1–5]. Furthermore, a comprehensive report from an occupational safety and health agency revealed that employees engaged in high-intensity emotional labor were 2.44 to 3 times more likely to screen positive for depression compared to those performing low-intensity emotional labor [6]. The study also found that these workers not only experienced a 2.86 to 3.07 times higher prevalence of clinical insomnia, but also demonstrated a 2.02 times greater likelihood of reporting suicidal ideation. These adverse effects and its consequent impact on productivity have prompted researchers to focus on categorizing or subtyping emotional labor workers and identifying vulnerable subgroups within this population [6–9]. In a previous study that categorized male firefighters into high-risk and normal groups based on emotional labor intensity [9], the high-risk group was associated with higher levels of depressed mood and insomnia, as well as lower levels of post-sleep fatigue recovery and resilience. These findings highlight the severe psychological consequences of high-intensity emotional labor, emphasizing the urgency of identifying and supporting these vulnerable groups. To effectively identify at-risk employees and develop appropriate support measures, more studies should explore which aspects of emotional labor are particularly detrimental to psychological well-being and thus should be targeted for intervention. The Job Demands-Resources (JD-R) model may provide a theoretical framework for understanding and addressing these vulnerabilities [10,11]. According to the JD-R model’s health impairment process, chronic imbalance between high job demands (such as emotional labor) and insufficient resources leads to energy depletion, burnout, and serious health deterioration [12]. Therefore, it is necessary to comprehensively assess job demand and resource factors related to emotional labor and examine their effects on psychological well-being.

Identifying at-risk employees is particularly important given a recent report that workers’ emotional labor strategies, especially maladaptive ones that are more detrimental to psychological well-being, tend to remain stable over time [8]. Specifically, the study identified distinct profiles of emotional labor based on employees’ use of different strategies, with profiles characterized by high levels of surface acting exhibiting the lowest job satisfaction and the highest emotional exhaustion and turnover intentions. A within-person stability analysis revealed that 95.4% of this subgroup remained in the same profile after three months, indicating that this group is particularly resistant to transitioning into more adaptive coping styles [8]. The persistence of the use of emotional labor strategy highlights the need for organizational-level screening and targeted interventions to address the enduring negative effects of emotional labor in this high-risk group. To effectively implement these measures, it is crucial to identify the specific aspects of emotional labor that have significant impact on psychological well-being. This targeted approach would enable organizations to develop more precise and effective interventions.

Building upon the need for proactive organizational measures, it is essential to explore specific interventions that can be effectively integrated into the work environments. Among these, meditation stands out as a promising approach that can be implemented in a naturalistic and cost-effective manner. The efficacy of meditation has been well-established through previous research, with studies showing that meditation can alleviate a variety of medical and psychiatric conditions, including cardiovascular disease and depression [13–20]. One of the key mechanisms is that meditation enhances autonomic regulation by promoting parasympathetic over sympathetic activity, thereby facilitating emotional and physiological relaxation [13,14,21,22]. Given that persistent emotional suppression and control during emotional labor can result in prolonged autonomic arousal and consequent cardiovascular and immune dysfunction [1,3,4], interventions targeting autonomic regulation become particularly relevant. For example, workplace-based mindfulness programs have been reported to reduce cortisol production and sympathetic nervous system reactivity, as well as improve autonomic balance and immune function [23]. Besides physiological relaxation, meditation is known to improve positive and negative affect [16]. According to Affective Events Theory (AET) [24], employee attitudes and behaviors are not merely products of stable work environments (e.g., work intensity) but are primarily driven by specific daily work events that trigger internal emotional reactions. Indeed, a study of call center workers revealed that affect mediated the relationship between work characteristics and employee health and job satisfaction [25]. Specifically, work overload induced negative emotions that led to health problems, whereas positive work characteristics such as autonomy, supervisor support, and welfare enhanced job satisfaction through positive affect. These findings suggest that managing employees’ daily affective experiences may be critical for workplace well-being. From this perspective, meditation-based interventions become particularly relevant. In addition, from the JD-R perspective, job resources including institutional and social support become increasingly critical as job demands increase [12,26,27]. This theoretical framework suggests that the protective effects of resources are amplified under high-demand conditions, helping to buffer against work-related strain. Applied to meditation interventions, this may imply that the effects of meditation may differ depending on emotional labor intensity. However, few studies have investigated whether employees with different levels of emotional labor intensity would exhibit differential responsiveness after meditation interventions. Taken together, while research on emotional labor and its intervention strategies has grown substantially, important empirical questions remain unexplored. First, comparative analyses of how specific dimensions of emotional labor differentially affect psychological outcomes in high-risk versus low-risk workers are limited. Second, although vulnerable subgroups have been identified [8], their distinct psychological profiles require further characterization. Third, despite meditation showing promise as a workplace intervention, whether psychological changes after meditation vary across emotional labor risk group remains largely unexplored. As emotional labor becomes a defining feature of more occupations, developing evidence-based, tailored intervention strategies has become critical for protecting worker well-being and organizational sustainability. Addressing these needs, the present study examined how specific emotional labor characteristics relate differently to depressed mood and quality of life in high- vs. low-risk emotional labor groups, and compared their responsiveness to a meditation intervention.

Previous studies have demonstrated that an 8-week mind-body training (MBT) program incorporating movement-based meditation effectively reduces stress and enhances emotional intelligence, resilience, and coping abilities [28,29]. In addition, a 4-week MBT was found to improve functional connectivity within the default mode network [30]. Although it was once believed that intensive and prolonged (months to years) meditation is required for beneficial effects, recent studies showed that brief meditation is sufficient to induce mental, physical, and neuronal changes [31,32]. For example, daily meditation (20 min for 3–5 days) was effective in improving autonomic regulation and cognitive functions including attention, working memory, and executive control [33,34], and enhancing physiological relaxation as evidenced by decreased heart rate [35]. Given its accessibility, cost-effectiveness, and demonstrated benefits, meditation-based interventions could serve as a valuable tool in mitigating the adverse effects of emotional labor. Based on these findings, the present study adapted the previously validated 4-week and 8-week MBT programs [28–30] into a condensed 35-minute single session format that can easily be integrated into the work setting, and applied the program to high- and low-risk emotional labor groups. In particular, we aimed to identify how the subfactors of emotional labor relate differently to depressed mood and quality of life in each group. Given the accumulating evidence on the benefits of MBT [28–30], the present study focused on comparing immediate responsiveness to the MBT between the two groups. In addition, we conducted exploratory physiological analyses to examine which dimensions of emotional labor may be associated with autonomic changes following the MBT.

Materials and methods

Categorizing emotional labor groups

The level of emotional labor was assessed using the Korean Emotional Labor Scale (K-ELS) [36]. The K-ELS was validated through a nationwide validation study conducted by Korean Occupational Safety & Health Agency using a random sample of 1,042 Korean employees [37]. It has been widely used in previous studies, including those examining employees’ suicidal ideation, burnout, sleep disturbance, work-related symptoms of posttraumatic stress disorder, and the effectiveness of stress reduction programs [38–44]. The scale comprises five subscales: five items of emotional demand and regulation (EL1, e.g., “I consciously make an effort not to express negative emotions to customers.”), three items of overload and conflict in customer service (EL2, e.g., “I have to deal with aggressive or demanding customers.”), six items of emotional disharmony and hurt (EL3, e.g., “I feel hurt when I have to hide and suppress my emotions from customers.”), three items of organizational surveillance and monitoring (EL4, e.g., “I am monitored (e.g., through CCTV) to ensure I interact with customers according to workplace requirements.”), and seven items of lack of a supportive and protective system in the organization (EL5, e.g., “There are formal systems and procedures in place at work to help resolve and support issues that arise during customer interactions.”, reverse-scored). It consists of a total of 24 items rated on a 4-point Likert scale, with higher scores indicating greater emotional labor at work. As in previous studies [6,9], the sum of the first three subscales was defined as the intensity of emotional labor. Each subscale and the sum scores were converted to a score out of 100. In this study, the Cronbach’s alpha for the five subscales ranged from 0.669 to 0.878, demonstrating acceptable internal consistency.

A national research report of Korea Occupational Safety and Health Agency used receiver operating characteristic (ROC) curve analysis to determine the cut-off values for emotional labor scores in classifying depression [6]. In this analysis, the Patient Health Questionnaire-9 (PHQ-9) was used as the criterion standard for depression [45]. ROC curves were constructed to evaluate how well emotional labor intensity scores could predict PHQ-9-defined depression. The analysis identified optimal cut-off points—60.36 for males and 66.67 for females—at which the predictive performance was maximized. Individuals scoring above these thresholds were classified as being at high-risk for depression. As in a previous study [9], we adopted these cut-off values in our study to categorize employees into the high-risk (n = 332) and the low-risk (n = 391) emotional labor groups. Table 1 presents detailed demographic and work-related features of the two groups.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Baseline demographic, work-related, and outcome measures of high-risk vs. low-risk emotional labor groups.

https://doi.org/10.1371/journal.pone.0345553.t001

Results

The relationships between emotional labor and psychological well-being

Fig 1 demonstrates a correlation heatmap of the association of emotional labor subscales with depressed mood and quality of life in the two groups. Multiple regression analyses on depressed mood revealed that, in both groups, EL3 was positively associated with depressed mood (Table 2). Notably, EL5 showed a positive association with depressed mood only in the high-risk group (β = 0.11, p = .040). These findings suggest that the factors accounting for depressed mood may differ between the two groups, with lack of a supportive and protective system in the organization playing a more significant role in the high-risk group.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Stepwise multiple linear regression analysis examining associations between emotional labor subscales and psychological well-being in the high-risk and low-risk group.

https://doi.org/10.1371/journal.pone.0345553.t002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. The correlation heatmap of the association of emotional labor subscales with depressed mood and quality of life.

EL1: emotional demand and regulation, EL2: overload and conflict in customer service, EL3: emotional disharmony and hurt, EL4: organizational surveillance and monitoring, EL5: lack of a supportive and protective system in the organization, QOL: quality of life.

https://doi.org/10.1371/journal.pone.0345553.g001

Multiple regression analyses examining the relationship between emotional labor factors and each domain of quality of life revealed distinct patterns in the two groups (Table 2). In both groups, EL3 showed negative relationships with all quality of life domains. In the low-risk group, EL4 was negatively associated with the physical and environmental domains. Notably, in the low-risk group, EL2 was negatively associated with psychological and social domains, suggesting a more pronounced role of work-related stress in this group. Interestingly, EL1 was positively associated with environmental quality of life in the low-risk group, indicating that some aspects of emotional management might be linked to better perceived quality of life in the low-risk group, but not in the high-risk group.

Changes following the MBT

Table 3 summarizes the changes in psychological and physiological outcome measures in each group. There was a significant main effect of Time (pre- vs. post-training) on positive affect (F[1,719] = 343.22, p < .001, η²_p = .32), indicating that positive affect increased following MBT (pre-training: 21.78; post-training: 28.68). The main effect of Group revealed that, regardless of Time, the level of positive affect was greater in the low-risk group (F[1,719] = 9.25, p = .002, η²_p = .013). There was no significant interaction of Time x Group (F[1,719] = 1.01, p = .31, η²_p = .001).

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Changes following the MBT.

https://doi.org/10.1371/journal.pone.0345553.t003

For negative affect, the results showed significant interaction of Time x Group (F[1,719] = 12.60, p < .001; η²_p = .017) as well as significant main effect of Time (F[1,719] = 357.06, p < .001, η²_p = .33) and Group (F[1,719] = 25.30, p < .001, η²_p = .034). Subsequent independent t-test revealed a greater decrease in negative affect seen in the high-risk compared to the low-risk group (t = 3.37, p = .001). However, when controlling for job type and baseline severity, changes in negative affect were comparable between the two groups (F[1,719] =.179, p = .673).

Discussion

This observational study demonstrated that the relationships between emotional labor and psychological outcomes differed between high-risk and low-risk emotional labor groups. In both groups, emotional disharmony and hurt caused by emotional labor was associated with exacerbated depressed mood and reduced quality of life. However, lack of a supportive and protective system in the organization was significantly associated with depressed mood only in the high-risk group, suggesting a need for organizational-level interventions to effectively support the psychological well-being of high-risk emotional laborers. Regarding the program, both groups showed increases in positive affect and decreases in negative affect after MBT, with a greater reduction in negative affect observed in the high-risk group.

Emotional disharmony and hurt, considered to have the greatest impact among emotional labor subscales, has been associated with various negative outcomes including turnover intention and suicidal ideation in previous studies [42,44,63,64]. While emotional labor itself may not be inherently negative, the failure to regulate emotions effectively could be detrimental to psychological well-being. Thus, workplace programs focusing on emotional regulation skills could be beneficial [65–67]. Notably, depressed mood in the high-risk group was also associated with lack of a supportive and protective system in the organization in the present study, suggesting that the positive relationship between organizational support/protection and depressed mood may be more pronounced for workers experiencing high-intensity emotional labor. This aligns with contemporary occupational stress theories including the JD-R model, which suggests that job resources including institutional and social support become increasingly critical as job demands increase [12,26,27]. In high-stress occupational contexts, the protective effect of these resources is considerably enhanced, significantly buffering against work-related psychological strain [12,26,27]. Accordingly, high-risk workers may rely more on these support systems to mitigate psychological strain. Consistent with this, the high-risk group in this study showed a greater decrease in negative affect following the MBT compared to the low-risk group. This differential response in negative affect could indicate that high-risk group may be more reliant on and responsive to organizational support programs.

However, the greater reduction in negative affect observed in the high-risk group was no longer significant after controlling for job type and baseline severity. The high-risk group in our study is characterized by a higher proportion of occupations with low job autonomy (e.g., call-center) and greater baseline psychological burden, indicating that risk grouping reflects a cluster of workers with both demanding occupational contexts and heightened vulnerability. Accordingly, group differences in changes following the MBT should be interpreted in light of these occupational and baseline factors, rather than attributed to risk grouping alone. Future studies with samples balanced for job type and baseline severity are needed to clarify whether emotional labor intensity has additional explanatory power beyond these characteristics. Changes following the MBT by job type are summarized in S4 Table. Taken together, these findings underscore the need to consider emotional labor intensity in conjunction with occupational and baseline vulnerabilities when designing workplace interventions, and to develop tailored strategies that provide organizational support and address the heightened negative affect observed in high-risk workers.

In the low-risk emotional labor group, quality of life scores in the environmental domain increased with higher levels of emotional demand and regulation, while this was not observed in the high-risk group. This suggests that in the low-risk group, higher scores on emotional demand and regulation reflecting individual efforts in emotion regulation and the diversity of emotional expression during customer interactions may be related to a better perception of their environmental well-being. Prior research has similarly noted that emotional labor does not uniformly exert negative influences; for example, some studies have found positive associations between emotional labor levels and job satisfaction [68]. Ashforth and Humphrey (1993) proposed that displaying emotions appropriate to one’s job role can potentially make customer interactions more structured and predictable, helping workers avoid embarrassing interpersonal problems [69]. However, in the present study, such associations were not observed in the high-risk group. It is possible that, under conditions of high emotional labor intensity, organizational support and protection may play a more important role in sustaining workers’ mental well-being. In addition, overload and conflict in customer service and organizational surveillance and monitoring were negatively associated with quality of life only in the low-risk group. This discrepancy may be due to a ceiling effect, as these factors were already higher at baseline in the high-risk group. The high-risk group, comprising more call center workers (47.0%), experiences constant monitoring and restricted autonomy, which may help explain the weaker associations of these factors with quality of life in this group. Taken together, these findings highlight the importance of considering the unique characteristics and needs of different emotional labor groups when designing and implementing interventions to promote psychological well-being in the workplace. Future research should further investigate the complex interplay between emotional labor, organizational factors, and individual coping mechanisms to develop more targeted and effective interventions for promoting psychological well-being among workers in various emotional labor contexts.

Meditation-induced changes in positive and negative affect have been well-documented in the literature [16,28]. A meta-analysis revealed an estimated effect size of 0.34 for the effect of meditation on negative emotions, and of 0.25 for its effect on positive emotions [16]. Moreover, a previous study reported that the use of smartphone-based mindfulness applications resulted in positive affect [70], suggesting that meditation can be used as a cost-effective and convenient tool for improving affect. Consistent with these reports, we observed immediate changes in positive and negative affect after MBT. Prior research grounded in AET has demonstrated through longitudinal studies that employees’ momentary affective experiences at work influence the development of job attitudes, work behaviors, and employee well-being over time [25]. These findings highlight the importance of understanding affective trajectories through longitudinal assessment. Future intervention research could benefit from incorporating longitudinal designs with appropriate control groups to examine patterns of affective experiences over time. Such research on affective dynamics could inform the development of more targeted interventions for emotional labor workers.

Given the well-established effects of meditation on autonomic regulation [71,72], our exploratory physiological analysis focused on identifying which dimensions of emotional labor might relate to physiological responsiveness to MBT rather than demonstrating its physiological effects per se. Interestingly, the decrease in heart rate was positively correlated with baseline scores of emotional demand and regulation, suggesting that employees who exert greater effort to regulate their emotions and display diverse emotional expressions at work exhibited more pronounced physiological relaxation after MBT. One possible explanation is that these employees might experience greater stress relief due to higher baseline emotional strain. Alternatively, these employees may be more practiced in emotion regulation, which could facilitate their understanding and application of the techniques used in MBT. Both interpretations highlight the potential value of meditation-based interventions for employees in emotionally demanding roles. While the small sample limits group comparisons of physiological responsiveness, exploring which dimensions of emotional labor relate to physiological responses may provide potential insights for personalized workplace intervention design. Even within the same risk group, workers may vary in their responsiveness depending on their specific emotional labor characteristics or coping strategies, yet little is known about such moderating factors in emotional labor contexts. Future studies should examine whether specific dimensions such as emotional regulation predict responsiveness to meditation interventions, which could inform the design of more tailored programs and improve prediction of intervention outcomes.

Given the growing adoption of digital wellness programs in the workplace, the adaptability of the current MBT program to online formats warrants consideration. Our previous research demonstrated that a 4-week online MBT program incorporating the core components used in this study (stretching and mindfulness breathing) significantly reduced anxiety and trait anger among healthcare workers [30]. These findings suggest that the current MBT program could be effectively adapted to digital delivery formats. Online programs offer distinct advantages, including greater accessibility, scalability, and the potential to facilitate sustained practice over extended periods rather than single-session interventions. Considering that in-person delivery may include real-time expert guidance, enhanced participant engagement, and a stronger sense of organizational support, a hybrid approach combining initial in-person sessions with digital follow-up may optimize intervention effects.

This study has several limitations. While this study addressed emotional labor from both individual and organizational perspectives, employees’ coping strategies such as surface acting or deep acting were not assessed. Subtyping and profiling based on these coping strategies could potentially offer more insights for developing tailored interventions. Future research incorporating such behavioral measures or qualitative interviews could help elucidate the mechanisms underlying differences in risk groups. In addition, the present study did not assess participants’ subjective experiences with the MBT program. Future studies would benefit from incorporating qualitative measures, such as semi-structured interviews or open-ended survey questions, to explore how participants perceive the intervention, what barriers they encounter in applying the techniques to their daily work routines, and whether they find the program feasible and sustainable in real-world workplace settings. Such qualitative data would provide valuable insights for refining the intervention and enhancing its practical applicability. Second, the observational nature of this study and the absence of a control group limit the ability to attribute the observed changes solely to the MBT program. Other factors, including potential placebo effects, cannot be ruled out. In particular, for physiological measures such as resting heart rate, reductions observed after MBT may partly reflect natural adjustment to the environment rather than the program effect. Future studies with a control group of emotional laborers not receiving MBT intervention would provide stronger evidence for the effectiveness of the MBT intervention. Third, while the study included a large sample for psychological measures, the sample size for physiological measures was relatively small. This subset was selected based on participant availability, workplace scheduling, and limitations in equipment, resulting in a non-random, restricted sample. Consequently, the generalizability of findings on physiological changes and their links to emotional labor subscales is limited. Specifically, the null findings for heart rate variability warrant further investigation given the insufficient statistical power in the present study. Future research would benefit from larger sample sizes with a priori power analyses and randomized sampling procedures to better capture the physiological changes associated with workplace meditation interventions. Fourth, the study sample consisted predominantly of female participants (over 90%), which limits the generalizability of the findings to male emotional laborers. This gender imbalance is likely due to the nature of the occupations included in the study, such as customer service and call center work. Baseline differences between male and female are summarized in S1 Table. Although sensitivity analyses restricted to female employees yielded results consistent with the main findings (S2 and S3 Tables), the small proportion of male participants in this study, together with known gender differences in emotional labor [73], indicates that the applicability of the present findings to male employees is limited. Future studies should recruit more balanced samples to clarify potential gender-specific patterns in emotional labor and to examine whether the present findings can be applied to male workers. Lastly, the relatively low adjusted R² values observed in the regression models suggest that other unmeasured factors may have contributed to depressed mood in emotional labor workers. Future studies should consider incorporating additional psychological, interpersonal, and organizational variables, as well as objective measures, to better account for the variance in psychological outcomes. In addition, due to the lack of multiple comparison corrections, the possibility of false positive associations cannot be ruled out. The regression results should therefore be interpreted with caution. Nevertheless, the present study has the strength of identifying distinct psychological profiles among high-risk emotional laborers, which may provide groundwork data for future controlled studies aimed at investigating vulnerable groups.

Conclusion

This study revealed important differences in psychological profiles and intervention responses between high-risk and low-risk emotional labor groups. Emotional disharmony and hurt was consistently associated with poorer psychological well-being across both groups, while lack of a supportive and protective system in the organization showed a stronger association with psychological difficulties in the high-risk group. Both groups demonstrated reduced heart rates and negative affect following the MBT, though the high-risk group showed significantly greater reductions in negative affect. These findings suggest that detailed consideration of workers’ emotional labor profiles encompassing both individual emotional experiences and organizational support factors should be incorporated when designing individualized workplace stress management programs.

Supporting information

Acknowledgments

We would like to acknowledge the contributions of Global Cyber University.