Neglected empathy: A study of the mechanisms of social workers’ empathic competence on work withdrawal

Abstract

Background

Empathy is a core professional competence in social work. Managing the emotional labor arising from empathy is vital to social workers’ sustainable development and occupational health.

Objective

Based on the research findings, we offer practical reflections and recommendations. Our goal is to relieve the emotional burden caused by empathy and support the long-term wellbeing of social workers.

Methods

We widely distributed the “Empathy Capacity Scale for Social Workers” and surveyed 526 practitioners across China. Using structural equation modeling (SEM), we analyzed the relationships among empathy, emotional labor, burnout, and work withdrawal.

Results

Results show that empathy significantly and positively affects emotional labor, burnout, and work withdrawal. Emotional labor and burnout serve as a sequential mediator in the path from empathy to withdrawal. Perceived organizational support moderates the effects of empathy on both emotional labor and burnout.

Conclusions

Social workers operating in empathy-intensive and emotionally demanding environments often endure prolonged psychological and physical strain. This increases the likelihood of work withdrawal and undermines overall wellbeing. However, meaningful support from organizations and society can significantly improve this situation. Further research is needed to examine social workers across different cultures and service contexts, especially those in Western countries.

Keywords

social workers empathy burnout professional emotional stress perceived social support mental health

Introduction

Social workers constitute a crucial force in modern social governance and public service delivery.^1,2 According to the China Civil Affairs Statistical Bulletin, the social work sector employed over 11.52 million professionals by the end of 2023.³ As social governance modernization accelerates globally, social workers worldwide—including those in China—increasingly serve on the frontlines addressing diverse societal challenges. Specifically, they fulfill a “last-mile” function in community services, poverty alleviation, and psychological mediation. Due to their constant exposure to vulnerable populations and complex social issues,^4,5 empathy has consequently emerged as a critical factor influencing client satisfaction, burnout levels, well-being, and career longevity among social workers.⁶

Research reveals that empathy, signifying emotional investment and affective resonance with others,⁷ serves as the driver of personal emotional engagement. It fosters prosocial behaviors,^8,9 enhances interpersonal relationships,^10,11 and elevates subjective well-being.¹² These critical traits position social workers as both “advocates” for vulnerable groups and “bridges” connecting governments and citizens. However, functioning in high-empathy contexts as dual-role agents accelerates the depletion of emotional resources required for accurately identifying and responding to clients’ needs.^13–15 When confronted with stagnant salaries, limited career advancement, and workloads disproportionate to available resources, they often end up burning the candle at both ends, leaving them highly susceptible to chronic empathic fatigue and emotional exhaustion. This insidiously compromises occupational health and triggers work withdrawal behaviors. Long-term exposure to these conditions intensifies talent attrition within the social work sector, culminating in a structural impasse characterized by surging enrollment in social work qualification examinations alongside persistent workforce shortages. Consequently, this dilemma not only diminishes service quality but also perpetuates internal pressure cycles within the profession. Urgent interventions are therefore required to enhance professional identity and job efficacy, thereby stabilizing the workforce and ensuring sustainable, efficient social work practice.

Against this backdrop, this study adopts empathy capacity as the core construct to investigate its impact on burnout and well-being through emotional labor pathways, ultimately triggering work withdrawal. Furthermore, it examines the moderating role of perceived organizational support. While numerous studies have focused on the intensity of emotional labor among social workers and its associations with burnout or turnover intention, a significant gap persists in comprehensively exploring empathy's fundamental mechanisms. Consequently, this research not only provides novel theoretical perspectives on mental health mechanisms in social workers, but also delivers evidence-based insights and practical recommendations for policymakers and administrators regarding talent development and occupational safeguards.

Research on empathy competence among social workers

Social work is fundamentally an emotionally intensive practice centered on “life influencing life.” While this defines its professional distinctiveness, it concurrently poses significant risks to practitioners’ occupational health, endowing the profession with a characteristic double-edged nature.¹⁶ As a core competency in this field, empathy encompasses not only the innate capacity to perceive and sensitively respond to others’ emotional states, but also an intrinsic motivation to care for others’ well-being.⁷ Consequently, empathy is widely recognized as a pivotal driver of job performance.¹⁷ Specifically, when engaging with patients or vulnerable populations, social workers must demonstrate acute detection of emotional states coupled with appropriate responsiveness. This capacity extends beyond cognitive understanding (“knowing others’ feelings”) to include affective resonance (“feeling with others”) and behavioral responses (“responding appropriately”).¹³ Crucially, however, echoing social work's double-edged nature, empathy does not invariably constitute a positive resource; its excessive activation may consequently induce empathy fatigue.

Although society highly values the empathic labor of social workers, its practice inevitably demands substantial personal effort and sacrifice—costs frequently overlooked in public expectations.¹⁸ These pressures are particularly pronounced in high-emotional-intensity settings such as eldercare facilities and medical social work, where practitioners confront clients with elevated emotional density and compressed recovery windows.¹⁹ Research indicates that while highly empathic social workers achieve superior service outcomes, those lacking effective self-regulation strategies experience significant emotional depletion and psychological exhaustion following repeated deep emotional engagements.^20,21 More critically, persistent exposure to clients’ trauma narratives may transform empathy into an “emotional toxin,” consequently triggering dual crises of burnout and secondary traumatic stress.^22,23 This paradox is especially acute in medical social work: studies confirm that high empathy reduces cancer patients’ suicidal ideation by up to 30%, yet simultaneously risks rendering practitioners “second victims”.²⁴ The Person-Environment Resilience Theory provides a crucial framework for understanding this dilemma, positing that empathy and psychological resilience must be jointly cultivated. Without resilience training, excessive empathy undermines rather than sustains occupational stability, resulting in continuous erosion of mental well-being.²⁵

Research on emotional labor

Emotional labor, as the operational manifestation of empathy, denotes the process of displaying specific emotions to achieve occupational objectives. For social workers, their empathic capacity necessitates more frequent and intensive emotional labor. This encompasses both surface acting (e.g., feigning smiles) and deep acting, wherein workers authentically internalize clients’ emotions.²⁶ Consequently, social workers are expected to consistently exhibit patience, empathy, and care—even while managing internal distress.²⁷

Research demonstrates that such sustained role demands and emotional suppression progressively deplete emotional resources, thereby inducing cognitive dissonance and exhaustion.²⁸ Notably, female social workers frequently shoulder heavier emotional burdens due to gendered caregiving expectations.²⁹ A poignant qualitative vignette illustrates this paradox: a female practitioner vomited after counseling a domestic violence survivor, yet resumed leading a children's group with a smile within five minutes.³⁰ If such emotional dissonance persists unchecked, it may propagate psychosomatic symptoms. Clinical evidence reveals that chronic emotional suppressors among social workers exhibit significantly elevated rates of chronic fatigue (82.6%), non-specific back pain (63.4%), and sleep disorders (71.2%) compared to the general population—symptoms strongly correlated with emotional labor intensity.³¹ Disturbingly, a 2023 survey of Hangzhou social workers found 74.6% experienced moderate-to-severe burnout, with the peak prevalence occurring among 30–40-year-olds—precisely their professional prime.³²

Research on occupational burnout

Occupational burnout denotes a state of physical and mental exhaustion stemming from prolonged engagement in high-intensity, high-demand service professions. Its core manifestations include emotional depletion, depersonalization, and diminished personal accomplishment.³³ For social workers, sustained exposure to intensive emotional labor significantly elevates burnout risk,^34,35 particularly in environments lacking organizational support and positive reinforcement.⁴ The consequences of burnout can be harmful to workers, the people who they service, their families and careers.³⁶

Crucially, social work burnout rarely originates from isolated incidents; rather, it emerges from the cumulative interplay of persistent emotional labor, psychological strain, and insufficient organizational backing. The ramifications are extensive: in eldercare settings, burnout arising from poor social support not only compromises practitioners’ physical and psychological health, but also directly impairs resident care quality and institutional development. This manifests as reduced work motivation, emotional detachment from elderly clients, and deteriorating personal health.³⁷ Frontline child welfare social workers in St John's, NL, Canada also regularly experience varied feelings and emotions of burnout and utilize a variety of self-care strategies to manage.³⁸ U.S. child welfare studies confirm that work-family conflict, role ambiguity, and role conflict contribute to burnout, which subsequently predicts work withdrawal and turnover intentions. Job demands precede burnout, which progressively erodes work engagement over time.³⁹ Furthermore, robust evidence links burnout severity to physical and psychological sequelae—including chronic fatigue, sleep disturbances, and headaches. Social workers often transmute psychological stress into somatic symptoms, consequently undermining both quality of life and job stability.^31,32 Therefore, comprehensively understanding burnout's multifaceted etiology and developing effective interventions is imperative. This is vital not only for preserving practitioners’ well-being and career continuity, but also for ensuring the long-term quality and resilience of social service systems.

Research on work withdrawal

Work withdrawal denotes negative workplace behaviors aimed at avoiding tasks or work situations within organizational contexts.⁴⁰ Its prevalence and detrimental effects are empirically established: Bennett et al. (2000) found 31% of employees deliberately delayed tasks, 33% arrived late without authorization, and over half (52%) misused break times.⁴¹ These behaviors not only erode organizational efficiency but also, alongside other counterproductive conduct, impose substantial financial burdens. Murphy (1993) quantified this burden at $200 billion annually (inflation-adjusted).⁴² Critically, Viswesvaran (2002) and Zimmerman et al. (2009) consistently demonstrated that work withdrawal significantly impairs employee performance.⁴³ Whether in fast-paced retail, precision manufacturing, or creative knowledge teams, this detrimental effect persistently undermines organizational effectiveness.

In social work—an emotionally intensive field—withdrawal risk is particularly acute. A 2024 UK Community Care survey of 716 social workers revealed over 40% considered leaving due to extreme stress, with 10% nearing psychological breakdown. This progression from fatigue to helplessness and disengagement epitomizes sector-specific withdrawal. Concurrently, a 2022 UNISON report warned social workers were at a “breaking point,” with approximately half exhibiting withdrawal behaviors like emotional detachment and task avoidance. Consequently, emotional exhaustion in this high-demand profession frequently evolves into withdrawal, avoidance, and detachment—patterns marked by insidious destructiveness.

Among drivers of social work withdrawal, accumulated negative emotions from emotional labor constitute a core psychological mechanism. Professor Liu Chao, a leading mainland Chinese researcher, has demonstrated a direct effect of emotional labor on withdrawal behaviors.⁴⁴ In the Western organizational behavior literature, Elfenbein et al.⁴⁵ have highlighted the dangers of surface acting—suppressing true feelings and merely adjusting external expressions. Such inauthentic performance fails to resolve negative affect and instead intensifies burnout. Once trapped in this self-reinforcing negative spiral, employees often substitute problem-solving with avoidance impulses—neglecting challenges or engaging in perfunctory task completion rather than pursuing solutions. Grounded in this “negative emotion → avoidance tendency” mechanism, empirically validated evidence indicates emotionally overwhelmed employees face significantly higher withdrawal risks than emotionally stable counterparts.

Research on the moderating role of perceived organizational support

Perceived Organizational Support (POS) encompasses employees’ perceptions of their organization's emotional and material valuation, recognition, and care.⁴⁶ For social workers specifically, organizational care extends beyond financial incentives to include emotional supervision, professional development opportunities, and flexible management strategies⁴⁷ —critical resources for mitigating occupational stress and sustaining motivation.

Empirical evidence consistently demonstrates POS's buffering effect on emotional labor-induced burnout. For instance, during the COVID-19 pandemic, Canadian frontline social workers reporting high burnout and secondary traumatic stress frequently attributed their resilience to organizational support, exemplified by statements such as, “support from my agency kept me going”.⁴⁸ Across diverse cultural contexts:U.S. healthcare workers with higher POS showed significantly lower burnout scores (40% reduction risk).⁴⁹ Jordanian medical staff exhibited reduced anxiety and stress levels with increased perceived support.⁵⁰ Post-pandemic Chinese studies confirmed job resources buffered the impact of job demands on burnout and improved overall well-being.^51,52 Collectively, these findings indicate POS not only moderates burnout intensity triggered by emotional labor but also functions as a protective barrier against its progression to work withdrawal. This moderating mechanism elucidates why some high-empathy social workers engaged in intensive emotional labor avoid proportional burnout or withdrawal behaviors.

Methods

Theoretical foundation

Job demands–resources model (JD-R model)

As a leading theoretical framework in organizational behavior, the Job Demands–Resources (JD-R) model offers a systematic lens to explore the dynamic tension between employee exhaustion and engagement. It is especially applicable to the emotionally intense and complex work ecology faced by social workers. The model was originally introduced by Evangelia Demerouti. A few years later, Wilmar B. Schaufeli and colleagues expanded it in 2004 by incorporating work engagement into the framework. Later, Despoina Xanthopoulou further extended the model by integrating personal resources. In 2015, Schaufeli also included participative leadership as a component of the JD-R model.

According to the JD-R model, all job characteristics can be classified into two broad categories: job demands and job resources. Demerouti et al. (2001, p.501) define job demands as “those physical, psychological, social, or organizational aspects of the job that require sustained effort and are therefore associated with certain physiological and psychological costs.” In simpler terms, job demands represent the “negative” aspects of work that deplete energy, such as workload, role conflict, time pressure, and job insecurity. In contrast, job resources are considered the “positive” aspects of work. They are defined as those physical, psychological, social, or organizational factors that serve one or more of the following functions:

facilitating the achievement of work goals;

reducing job demands and the associated psychological and physiological costs;

stimulating personal growth, learning, and development. Examples include colleague support (which helps achieve goals), job autonomy (which may reduce demands), and performance feedback (which may enhance learning) (Figure 1).

Figure 1.

JD-R model.

Emotional Labor Theory

Emotional labor, a concept first introduced by sociologist Arlie Hochschild in 1983, refers to the process by which workers regulate and manage their emotions to meet job expectations. This process typically involves three strategies: surface acting (faking emotions), deep acting (modifying inner feelings), and genuine expression (aligning authentic emotions with role expectations). The core mechanisms of emotional labor include the following:

Commodification of emotion: In the service industry, emotions are treated as commodities that can be purchased and regulated by organizations. Workers are expected to display specific emotional expressions as dictated by their roles—for instance, flight attendants must smile at all times.

Risk of self-alienation: Prolonged emotional masking can result in a split between one's authentic self and professional persona, often leading to identity confusion and psychological dissonance.

Resource depletion model: Emotional labor consumes psychological energy. Without compensatory resources such as fair compensation or recognition, this depletion can lead to emotional exhaustion, including burnout, depression, and withdrawal from work.

Research hypotheses

As a core trait of professional competence, empathy drives positive service performance among social workers and enhances interaction and understanding with clients. However, it may also increase emotional labor due to deep emotional involvement,²⁸ which in turn can deplete their psychological and physical resources.^22,23 Prior studies have shown that higher levels of empathy are associated with a greater tendency to adopt high-investment strategies such as deep acting, which may lead to increased job stress and health risks.³⁶ Based on this, the following hypotheses are proposed:

H1a: Empathy has a significant positive effect on emotional labor among social workers.

H1b: Empathy has a significant positive effect on job burnout among social workers.

H2: Empathy has a significant positive effect on work withdrawal among social workers.

Emotional labor, as a depleting process of emotional regulation, poses a persistent psychological challenge in social work. Prolonged engagement often leads to resource exhaustion and emotional fatigue, eventually resulting in job burnout.²⁸ To meet organizational or professional expectations, social workers frequently engage in surface or even deep acting. The accumulated emotional tension inevitably translates into psychological fatigue over time.²⁶ Meanwhile, once burnout reaches a certain threshold, its impact extends beyond emotional strain. It may trigger work withdrawal through stress-related mechanisms and has been identified as a key predictor of turnover intention.⁴⁵ Accordingly, the following hypotheses are formulated:

H3: Emotional labor has a significant positive effect on job burnout.

H4a: Emotional labor has a significant positive effect on work withdrawal.

H4b: Job burnout has a significant positive effect on work withdrawal.

Building on existing theoretical foundations and empirical discussions,^20–23^,28,44–45 this study further proposes a chained mediation effect. It posits that empathy enhances emotional labor, which then leads to burnout, and ultimately results in work withdrawal. This pathway echoes the ongoing discourse on “compassion fatigue” in social work and offers a new lens for understanding the psychological health of social workers. On this basis, the following hypothesis is proposed:

H5: Emotional labor and job burnout play chained mediating roles in the relationship between empathy and work withdrawal among social workers.

Given the buffering effect of organizational context on psychological stress, and its potential to mitigate the negative outcomes of empathy,^48–50 this study introduces perceived organizational support (POS) as a moderating variable. POS not only strengthens employees’ resilience in emotion regulation but may also serve as a protective buffer in the transition from burnout to withdrawal. Prior research indicates that under high POS conditions, employees maintain greater psychological resilience, experience less depletion, and demonstrate milder burnout responses.⁵¹ Therefore, the following hypotheses are proposed:

H6a: POS moderates the relationship between empathy and emotional labor among social workers.

H6b: POS moderates the relationship between empathy and job burnout among social workers.

A research model is constructed based on the above hypotheses, as shown in Figure 2.

Figure 2.

Hypothetical research model.

Research design

Scale design

To achieve the research objectives, this study developed and widely distributed the “Empathy Capacity Questionnaire for Social Workers,” based on a comprehensive review of existing domestic and international instruments. The questionnaire included basic demographic information and in-depth questions on empathy capacity and its consequences. It measured five key dimensions: empathy capacity, emotional labor, job burnout, work withdrawal, and perceived organizational support. All items were rated on a 5-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). The scale dimensions were adapted from established instruments, including the Interpersonal Reactivity Index (IRI) by Davis (1983) for empathy, the Emotional Labor Scale by Brotheridge and Lee (2003), the Maslach Burnout Inventory-Human Services Survey (MBI-HSS) by Maslach and Jackson (1981), the Survey of Perceived Organizational Support (SPOS) by Eisenberger et al. (1986), and the Work Withdrawal Scale by Hanisch and Hulin (1990). This ensured the scientific rigor and authority of the measurements.

Prior to full-scale distribution, a pilot test was conducted within a selected social work agency to ensure the questionnaire's validity. Follow-up interviews were conducted with a subset of participants. Based on the feedback collected and preliminary data analysis, we refined the ordering of items in the questionnaire for greater clarity and logical flow. The finalized version was suitable for large-scale empirical investigation.

Data collection and sample characteristics

This study employed a stratified random sampling approach.Initially, the country was divided into three regional strata—eastern, central, and western China—based on distinct socioeconomic development characteristics.Within each stratum, a number of social work service agencies were then independently and randomly selected with support from local social work associations, resulting in a total of 23 participating organizations.Finally, within each selected agency, a subset of frontline social workers was randomly chosen as the final sample to complete the survey.Data collection was conducted from March to June 2025. A total of 574 questionnaires were distributed, with 554 returned. After data cleaning and exclusion of invalid responses, 526 valid questionnaires were retained, yielding a response rate of 96.5% and a valid response rate of 91.6%.

Descriptive statistics indicated that the respondents were primarily social workers. The gender distribution was relatively balanced, with 56.7% male and 43.3% female participants. Nearly half of the respondents (49.6%) had less than two years of work experience. Those with 2–6 years made up 33.5%, and 18.9% had six years or more. This sample composition aligned well with the target population and was appropriate for empirical analysis.

Results

Data analysis was conducted using SPSS 24.0 and AMOS 24.0. First, the reliability and validity of the scales were tested. Reliability was measured using Cronbach's alpha and composite reliability (CR). Validity was assessed using the Average Variance Extracted (AVE). Then, a structural equation model (SEM) was employed to test hypothesized paths within the conceptual framework. Finally, bootstrapping was used to examine mediation effects for specific hypotheses.

Factor analysis

Reliability Testing

The KMO and Bartlett's tests were conducted for all variables. Results showed KMO values above 0.7 and significant p-values below 0.05, indicating that factor analysis was appropriate. Reliability was assessed using Cronbach's alpha. All five variables had alpha coefficients above 0.75, indicating an acceptable level. This suggests strong internal consistency and suitability for empirical analysis.

Validity Testing

Validity includes both convergent and discriminant validity. Convergent validity was examined using Average Variance Extracted (AVE) and Composite Reliability (CR). AVE values above 0.5 and CR values above 0.7 are considered evidence of good convergent validity. Discriminant validity is confirmed when the square root of each construct's AVE exceeds its correlations with other constructs. Tables 1 and 2 demonstrate that the sample data achieved good validity. (Note: p-values are reported as < 0.01, < 0.05, or < 0.1 throughout the paper.)

Table 1.

Scale reliability and validity tests.

Dimensions	Question items	Unstandardized coefficients	Standard errors	Z-values	P-values	Standardized coefficients	Cronbach's α coefficients	Combined reliability	Mean variance precipitation
Empathy capacity for social workers	A1	1				0.927	0.939	0.94	0.839
	A2	1.037	0.027	37.838	***	0.934
	A3	0.947	0.029	32.967	***	0.887
Emotional labor	B1	1				0.899	0.949	0.95	0.863
	B2	1.027	0.029	36.025	***	0.94
	B3	1.018	0.028	36.633	***	0.947
Job burnout	C1	1.042	0.029	36.547	***	0.925	0.961	0.961	0.861
	C2	0.991	0.026	38.539	***	0.941
	C3	0.982	0.026	37.7	***	0.934
	C4	1				0.912
Work withdrawal	D1	1				0.938	0.945	0.946	0.854
	D2	1.023	0.025	41.24	***	0.939
	D3	1.04	0.029	35.287	***	0.895
Perceived organizational support	E1	1				0.842	0.872	0.878	0.706
	E2	1.127	0.047	24.02	***	0.9
	E3	0.962	0.05	19.35	***	0.792

Table 2.

Discriminant validity.

	Perceived organizational support	Empathy capacity for social workers	Emotional labor	Job burnout	Work withdrawal
Perceived organizational support	0 . 84
Empathy capacity for social workers	0.286	0.916
Emotional labor	0.241	0.736	0.929
Job burnout	−0.382	0.769	0.702	0.928
Work withdrawal	−0.317	0.747	0.688	0.863	0.924

Table 3.

Results of hypothesis testing of the model.

Hypothesis	Impact path	Unstandardized regression coefficients	S.E.	C.R.	P	Standardized regression coefficients	Test results
H1a	Empathy capacity for social workers→Emotional labor	0.792	0.041	19.446	***	0.736	Significant positive impact
H1b	Empathy capacity for social workers→Job burnout	0.525	0.046	11.424	***	0.55	Significant positive impact
H4a	Emotional labor→Job burnout	0.263	0.041	6.351	***	0.297	Significant positive impact
H2	Empathy capacity for social workers→Work withdrawal	0.163	0.048	3.367	***	0.157	Significant positive impact
H4a	Emotional labor→Work withdrawal	0.097	0.039	2.469	***	0.1	Significant positive impact
H4b	Job burnout→Work withdrawal	0.728	0.05	14.689	***	0.672	Significant positive impact

Table 2 shows that perceived organizational support is significantly negatively correlated with job burnout (r = -0.382, p < 0.001) and work withdrawal (r = -0.317, p < 0.001). It is positively correlated with empathy capacity (r = 0.286) and emotional labor (r = 0.241). This indicates that perceived organizational support functions as a crucial buffering moderator. Higher perceived organizational support is associated with lower levels of burnout and withdrawal. Enhancing organizational support can mitigate the cascade from empathy to emotional labor, burnout, and ultimately withdrawal.

Structural equation model fitting and hypothesis validation

The study employed maximum likelihood estimation in AMOS 24.0 to evaluate the structural equation model. Model fit indices were as follows: χ² = 158.281, df = 59, χ²/df = 2.683, SRMR = 0.0163, RMSEA = 0.057, CFI = 0.988, IFI = 0.988, TLI = 0.985, GFI = 0.956. All fit indices met recommended thresholds, indicating good model fit and overall adequacy.

Table 3 presents the hypothesis testing results. (1) Empathy capacity significantly and positively affected emotional labor, burnout, and withdrawal. (2) Emotional labor significantly increased both burnout and withdrawal; burnout also positively influenced withdrawal.

Mediation effect testing

Based on the SEM results, bias-corrected percentile bootstrap (1000 samples) was used to test mediation effects of emotional labor and burnout. Results are presented in Table 4.

Table 4.

Tests for multiple chain mediation effects.

Path Relationships		Point Estimates	Bootstrap Test
			bias-corrected 95% CI		percentile 95% CI
			Lower Limit	Upper Limit	Lower Limit	Upper Limit
Mediation effects, direct effects, and total effects tests
M1 Specific intermediary	Empathy capacity for social workers→Emotional labor→Work withdrawal	0.076	0.000	0.164	−0.002	0.160
M2 Specific intermediary	Empathy capacity for social workers→Job burnout→Work withdrawal	0.382	0.277	0.500	0.273	0.499
M3 Remote mediation	Empathy capacity for social workers→Emotional labor→Job burnout→Work withdrawal	0.152	0.093	0.228	0.088	0.222
M4 Total indirect effects	M1 + M2 + M3	0.611	0.493	0.745	0.485	0.743
M5 Direct effects	Empathy capacity for social workers→Work withdrawal	0.163	0.024	0.295	0.024	0.296
M6 Total effects	M1 + M2 + M3 + M4 + M5	0.773	0.685	0.880	0.682	0.875
Comparison of intermediation effects
DIFF1	M1 vs.M3	−0.075	−0.181	0.024	−0.178	0.028
DIFF2	M2 vs.M3	0.231	0.087	0.382	0.084	0.380
DIFF3	M1 vs.M2	−0.306	−0.439	−0.156	−0.445	−0.161
Percentage of intermediary effects
P1	M3/M4	0.249	0.147	0.363	0.146	0.362
P2	M1/M4	0.125	−0.003	0.245	−0.003	0.246
P3	M2/M4	0.626	0.476	0.771	0.481	0.775
P4	M4/M6	0.790	0.635	0.968	0.634	0.967
P5	M5/M6	0.21	0.032	0.365	0.033	0.366

The results revealed that the direct, total indirect, and total effects of empathy capacity on withdrawal were all positive and significant (95% confidence intervals excluded 0; p < 0.001). Three specific mediation paths were tested: “Empathy → Emotional Labor → Withdrawal,” “Empathy → Burnout → Withdrawal,” and “Empathy → Emotional Labor → Burnout → Withdrawal.” All three mediation effects were significant, as their 95% confidence intervals did not include zero, supporting Hypothesis H5. The direct effect of empathy on withdrawal remained significant. Emotional labor, burnout, and their combined sequential path partially mediated the impact of empathy on withdrawal.

Moderation effect testing

Using the PROCESS macro developed by Hayes (2012), a moderated chain mediation model (Model 71) was tested. The results revealed the following: (1)

Organizational support enhances the positive transformation of empathy resources.

Along the emotional labor path, perceived organizational support had a significant enhancing moderation effect (β = 0.15, p < 0.001). When social workers perceived high organizational support, their empathy was more strongly translated into emotional labor investment (β changed from 0.83 to 0.53). This supports the Resource Gain effect: organizational support systems—such as emotional backing and value alignment—can activate empathy resources, making highly empathic workers more willing to engage emotionally.

As shown in Figure 3, perceived organizational support significantly amplified the positive effect of empathy on emotional labor. Under high perceived support (solid line), increased empathy led to a more efficient conversion into emotional labor. Highly empathic workers (right end) showed an emotional labor score of 4.16 in low-support settings, which rose significantly to 4.68 in high-support settings—an increase of 12.5%. This indicates that organizational support can activate emotional resources, helping empathic traits convert more fully into service engagement. Notably, among those with low empathy (left end), the enhancing effect of organizational support was weaker (2.84 to 3.02). This confirms the empathy-support matching principle: the value of support is most evident among individuals with high empathy traits. (2)

Organizational support buffers the risk of empathy-related exhaustion.

Figure 3.

Moderating effect of perceived organizational support on the empathy → emotional labor path.

Along the burnout path, perceived organizational support exerted a significant buffering effect (β= −0.22, p < 0.001). In low-support environments, highly empathic social workers faced a high risk of burnout (β= 0.93). High perceived support reduced this risk by 47% (β= 0.49). This aligns with Conservation of Resources (COR) theory and its “resource shield” mechanism: organizational support, via psychological safety and autonomy, prevents empathy from leading to emotional exhaustion.

As shown in Figure 4, perceived organizational support buffered the risk of empathy-related exhaustion. Among high-empathy individuals (right end), low support led to burnout levels as high as 4.18, nearing the scale's critical threshold. In contrast, high organizational support (dashed line) lowered burnout to 3.24—a 22.5% reduction—forming a protective buffer. This moderating effect supports COR theory: organizational support interrupts the vicious cycle of “empathy overload → emotional exhaustion” through the provision of supportive resources. Among those with low empathy (left end), support also reduced burnout (2.32 to 1.98), though the effect was weaker. This suggests the primary protective role of organizational support lies in shielding high-empathy individuals.

Figure 4.

Moderating effect of perceived organizational support on the empathy → burnout path.

Together, the two figures demonstrate a dual-path moderating effect of perceived organizational support. Among highly empathic workers, organizational support both boosts emotional labor through resource transformation (Figure 3) and reduces burnout through risk buffering (Figure 4). This “dual-path synergy” offers valuable insights for social work management. Building a high-support environment can help highly empathic workers achieve a balance between maximizing service engagement and minimizing burnout. This helps resolve the “double-edged sword” dilemma of empathy. In conclusion, all proposed hypotheses in this study were supported by the empirical data.

Discussion

Discussion and recommendations

Within social work, work withdrawal and turnover are frequently attributed to excessive emotional labor or burnout stemming from prolonged hours and suboptimal working conditions. This study corroborates these established perspectives, aligning with extant literature. However, a critical question persists: why do social workers chronically engage in emotional over-investment? When controlling for external factors like workload and environment, why do significant variations in burnout and withdrawal behaviors endure?

This study makes a novel contribution by revealing empathy capacity as the underlying driver of emotional labor, burnout, and withdrawal—beyond superficial “overwork” explanations. While this ostensibly positive trait facilitates client understanding and high-quality care, it paradoxically morphs into a risk factor for emotional depletion without organizational safeguards, exhibiting classic double-edged sword dynamics. Crucially, absent institutional buffers, empathy's positive force may transform into a systemic burden for both practitioners and the profession. Consequently, to safeguard mental health and establish sustainable emotional labor systems, this study proposes targeted interventions across three dimensions: (1)

Individual-level interventions: Strengthen emotional regulation to establish psychological resilience for sustainable empathy.

Cultural values significantly shape empathy expression and regulation. Notably, in Chinese societies emphasizing collectivism and self-sacrifice, high social expectations often blur emotional boundaries—potentially inducing over-empathy, emotional suppression, and burnout. Conversely, Western social workers typically maintain clearer professional boundaries, prioritize self-care, and advocate “sympathy” over complete emotional immersion. Consequently, practitioners globally may integrate Western trauma-informed practices,⁵³ which emphasize both recognizing client trauma and monitoring secondary trauma risks during empathic engagement.

Therefore,"You can't pour from an empty cup”, enhancing individual emotion regulation serves as a critical safeguard. Evidence-based approaches include self-compassion training through mindfulness meditation, self-care journaling, and breath-focused exercises—all proven to bolster emotional resilience. Simultaneously, developing cognitive flexibility in emotional labor strategies is essential. Within complex service contexts, social workers must dynamically shift between natural expression and deep acting, balancing authenticity with professional detachment. Empirical research confirms deep acting fosters greater emotional consistency than surface acting while significantly reducing exhaustion.⁵⁴ Therefore, it is recommended to enhance education and simulation training on emotional labor strategies. Creative research methods should be applied to deepen social workers’ understanding of empathy in practice.⁵⁵ This approach will enable practitioners to flexibly select the most suitable emotion regulation techniques across diverse service scenarios.Ultimately, these efforts will strengthen their strategic flexibility and professional efficacy. (2)

Organizational Level: Establishing Systemic Buffers Against Empathy-Driven Exhaustion

If organizations fail to provide institutional responses to emotional labor and burnout, even highly empathetic and emotionally skilled social workers will struggle to sustain empathy in the long run. In other words, the more an organization acknowledges the value of emotional labor and actively addresses its associated exhaustion, the better it can protect workers’ well-being and long-term career engagement.

First, the system of restorative labor should be further improved. Unlike physical labor, emotional labor induces subtler, persistent burnout patterns. Untreated emotional depletion severely compromises health and workforce stability. Consequently, governments and professional bodies should implement rotation systems and psychological recovery periods for high-intensity roles. Special leave policies—such as “emotional recovery days” or “supervisory mental health leave"—should be established with psychological coaching support. Simultaneously, organizations must create recovery spaces (e.g., relaxation rooms, counseling hubs, mindfulness retreats) to provide regular emotional regulation channels, thereby enhancing professional appeal.

Furthermore, scientific mental/physical health monitoring systems enabling closed-loop interventions are essential. The U.S. Employee Assistance Program (EAP) exemplifies this approach,⁵⁶ strengthening organizational capacity to detect emotional strain through systematic mental health support. Withdrawal behaviors—identified here as subtle exhaustion indicators—remain frequently overlooked in practice. Hence, implementing early-warning systems is imperative. We recommend biannual mental health assessments across social service organizations, coupled with personalized interventions (counseling, job reassignment, short-term leave) for at-risk staff to establish early identification-intervention-recovery cycles. (3)

Professional Level: Institutionalizing Protections for Empathic Labor to Optimize Sector Ecosystems

The social work profession must urgently integrate “empathic labor protection” into its professional discourse and foster emotionally supportive service environments that formally recognize practitioners’ invisible contributions. Social work is inherently emotion-intensive; though intangible, this labor profoundly impacts mental health and career sustainability. International precedents substantiate this: the UK's Professional Capabilities Framework explicitly designates “emotional awareness and regulation” as a core competency, identifying crisis-level emotional professionalism as a hallmark of senior practitioners.⁵⁷ Absent such recognition, marginalized empathic labor traps individual and organizational efforts in siloed inefficiencies, structurally eroding professional dignity and stability.

Concurrently, when procuring public services, governments should implement policy and fiscal mechanisms ensuring stable occupational safeguards.⁵⁸ Service contracts should incorporate dedicated “emotional support funds,” embedding emotional care costs within project budgets. Specifically, preallocating resources for group stress-relief activities and mental health assessments would institutionally validate emotional labor. This approach not only addresses providers’ fundamental rights but also constitutes critical investment in service quality and workforce retention.

Furthermore, to counter pervasive issues—short project cycles, unstable funding, and insecure employment—a multidimensional “service rating system” merits exploration. Government-funded projects could be evaluated using weighted metrics: client satisfaction, service outcomes, and practitioner well-being support. Projects achieving three-star ratings or higher would receive procurement priority. Thereby incentivizing organizations to strengthen internal support systems, this enhances practitioner resilience and belonging while advancing the sector toward high-quality, human-centered development with dual performance-wellness benefits.

Limitations and future directions

This study has several limitations that should be addressed and refined in future research. For instance, the sample was regionally concentrated, mainly involving social workers from China. Although the sample size was relatively large, differences in culture, organizational management, and policy context may lead to external validity bias. Therefore, the generalizability of the findings in international contexts should be interpreted with caution; This study employed a cross-sectional survey approach, which can effectively identify significant associations between variables. However, it cannot fully rule out the possibilities of reverse causality or long-term effects. Future studies are encouraged to use longitudinal designs or experimental methods. These approaches would help to clarify the causal pathways more accurately.

Conclusion

First, social workers’ empathy capacity exerts a significant positive influence on emotional labor levels, indicating that highly empathic practitioners invest greater emotional regulation to fulfill their helping roles. Simultaneously, empathy significantly predicts job burnout, suggesting sustained emotional investment without effective regulation readily culminates in exhaustion. Furthermore, empathy directly and substantially forecasts work withdrawal occurrence, paradoxically implying this strength may jeopardize physical health.Second, emotional labor substantially increases burnout likelihood, demonstrating that frequent emotion management intensifies psychological fatigue. Moreover, both emotional labor and burnout significantly contribute to work withdrawal, revealing accumulated emotional strain and mental fatigue as primary triggers for somatic symptoms. Third, emotional labor and burnout collectively form a significant chain-mediating pathway between empathy and work withdrawal: heightened empathy first amplifies emotional labor, subsequently induces burnout, and ultimately manifests in somatic symptoms. Finally, perceived organizational support demonstrates notable moderating effects. Crucially, high organizational support significantly attenuates emotional labor's impact on burnout. Conversely, such support buffers burnout's negative influence on work withdrawal, confirming its critical buffering role in safeguarding well-being and mitigating occupational risks.

Social work functions as a vital “human connection hub” bridging the government and the public. For the sustainable development of this profession, it is essential to address the double-edged nature of empathy. On the one hand, empathy serves as the cornerstone of professional relationships. It enables social workers to accurately perceive the emotional needs of clients, thereby improving the quality of services, increasing satisfaction, and fostering trust. On the other hand, when intense empathic emotional labor is overlooked or exploited over time, it can easily result in burnout and health-related crises, ultimately leading to work withdrawal or even resignation. Therefore, the findings of this study offer a timely lens through which to examine the intricate interplay among empathy, emotional labor, burnout, work withdrawal, and perceived organizational support. These insights underscore the urgent need to strengthen institutional safeguards for social workers. Empathy should not be left to personal endurance—it must be structurally supported. After all, even the strongest backs can break when overloaded. With adequate organizational backing, empathy can become a sustainable professional strength, not a source of self-depletion. Only then can social work remain a calling that “meets the world with compassion without being consumed by it.”

Footnotes

ORCID iD

Chen Wenxuan

Ethical approval (name of institute and number)

This study involving human participants was reviewed and approved by the Academic Committee of the School of Public Administration, Shandong Normal University (Approval No. 010420251029).

Informed consent

Written informed consent was obtained from all participants prior to their inclusion. Participants consented to participate in the study and for the publication of the findings. They were informed that the article would be openly accessible online. All participant data have been anonymized.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Xin

. Functions and capacity building of social workers. Chin Women's Mov 2010: 45–46.

Hare

. Defining social work for the 21st century: the international federation of social Workers’ revised definition of social work. Int Soc Work 2004; 47: 407–424.

Ministry of Civil Affairs of the People's Republic of China. Statistical bulletin on the development of civil affairs in 2023. 2024. https://www.mca.gov.cn/n156/n2679/c1662004999980001204/attr/355717.pdf (accessed 23 July 2025).

Nyashanu

Pfende

Ekpenyong

. Exploring the challenges faced by frontline workers in health and social care amid the COVID-19 pandemic: experiences of frontline workers in the English midlands region, UK. J Interprof Care 2020; 34: 655–661.

Sklar

Ehrhart

Aarons

. COVID-related work changes, burnout, and turnover intentions in mental health providers: a moderated mediation analysis. Psychiatr Rehabil J 2021; 44: 219–228.

Wagaman

Geiger

Shockley

, et al. The role of empathy in burnout, compassion satisfaction, and secondary traumatic stress among social workers. Soc Work 2015; 60: 201–209.

Meng

Yang

. A review of compassion fatigue: theoretical implications for social work. J Southwest Pet Univ (Soc Sci Ed) 2021; 23: 104–112.

Morelli

Rameson

Lieberman

. The neural components of empathy: predicting daily prosocial behavior. Soc Cogn Affect Neurosci 2014; 9: 39–47.

Decety

Bartal

IBA

Uzefovsky

, et al. Empathy as a driver of prosocial behaviour: highly conserved neurobehavioural mechanisms across species. Philos Trans R Soc Lond B Biol Sci 2016; 371: 20150077.

10.

Eisenberg

Miller

. The relation of empathy to prosocial and related behaviors. Psychol Bull 1987; 101: 91–119.

11.

Hoffman

. The contribution of empathy to justice and moral judgment. In: Eisenberg

Strayer

(eds) Empathy and its development. Cambridge: Cambridge University Press, 1987, pp.47–80.

12.

Hablado

Clark

. A state of de-stress: examining the relationship between subjective well-being, self-compassion, and empathy. Adv Soc Sci Res J 2020; 7: 412–426.

13.

Cuff

BMP

Brown

Taylor

, et al. Empathy: a review of the concept. Emot Rev 2016; 8: 144–153.

14.

Luo

. Ecological empathy and ecological empowerment: emotional accumulation and rational reflection of environmental social workers. Study Explor 2023: 38–46.

15.

Sessions

Robertson

Jones

, et al. Self-compassion in qualified and student social workers: a systematic literature review and meta-analysis. J Evid Based Soc Work 2025; 22: 76–103.

16.

Wang

. Professional emotions and rational practice in social work. Chin Soc Work 2020: 46.

17.

Prasetyo

Perez

JPA

Gumasing

MJJ

, et al. The effects of communication, empathy, encouragement, growth, and rewards on employee performance: a structural equation modeling approach. WORK 2023; 76: 749–758.

18.

Cameron

Hutcherson

Ferguson

, et al. Empathy is hard work: people choose to avoid empathy because of its cognitive costs. J Exp Psychol Gen 2019; 148: 962–976.

19.

Moudatsou

Stavropoulou

Philalithis

, et al. The role of empathy in health and social care professionals. Healthcare (Basel) 2020; 8: 26.

20.

Gerdes

Segal

. Importance of empathy for social work practice: integrating new science. Soc Work 2011; 56: 141–148.

21.

Byrne

Campos

Daly

, et al. The current state of empathy, compassion and person-centred communication training in healthcare: an umbrella review. Patient Educ Couns 2024; 119: 108063.

22.

Bride

Radey

Figley

. Measuring compassion fatigue. Clin Soc Work J 2007; 35: 155–163.

23.

. Understanding compassion fatigue among social workers: a scoping review. Front Psychol 2025; 16: 1500305.

24.

Ningxi

Yuting

Zhibo

, et al. Influence of social workers’ empathy ability on suicidal ideation of cancer patients. Front Public Health 2022; 10: 925307.

25.

Kumpfer

. Factors and processes contributing to resilience: the resilience framework. In: Glantz

Johnson

(eds) Resiliency and development: positive life adaptations. New York: Kluwer Academic, 1999, pp.179–224.

26.

Grandey

. When “the show must go on”: surface acting and deep acting as determinants of emotional exhaustion and peer-rated service delivery. Acad Manage J 2003; 46: 86–96.

27.

Hennig-Thurau

Groth

Paul

, et al. Are all smiles created equal? How emotional contagion and emotional labor affect service relationships. J Mark 2006; 70: 58–73.

28.

Jung

Baek

. A structural equation model analysis of the effects of emotional labor and job stress on depression among nurses with long working hours: focusing on the mediating effects of resilience and social support. WORK 2020; 66: 561–568.

29.

Walton

. Women in social work. 1st ed. London: Routledge, 1975.

30.

. Research on the emotional labor of female social workers . PhD Thesis, Hangzhou Normal University, 2022.

31.

Guo

Zhang

Zhu

, et al. Clinical characteristics and related factors of depressed patients with work withdrawal in different age stages. J Neurosci Ment Health 2022; 22: 539–546.

32.

. Research on the current situation and causes of occupational burnout among social workers . PhD Thesis, Zhejiang Gongshang University, 2023.

33.

Chen

Wang

. Job burnout: connotation, measurement and formation mechanism. Foreign Econ Manag 2019; 41: 86–99.

34.

Wang

. The impact and mechanism of emotional labor on occupational burnout of social workers. J East China Univ Sci Technol (Soc Sci Ed) 2021; 36: 66–77. 90.

35.

Mihai

Crivoi

Alecu

, et al. Stress and burnout among social workers – a relation mediated by coping styles. Humanit Soc Sci Commun 2025; 12: 1430.

36.

Ratcliff

. Social workers, burnout, and self-care: a public health issue. Dela J Public Health 2024; 10: 26–31.

37.

Fan

Yang

. The governance of “service poverty” for rural empty-nesters and the intervention of social work. Acad J Zhongzhou 2020: 79–84.

38.

Fuseini

. Suffering in silence”: how social workers in child welfare practice experience and manage burnout. Child Youth Serv Rev 2024; 166: 107939.

39.

Travis

Lizano

Mor Barak

. ‘I'm so stressed!': a longitudinal model of stress, burnout and engagement among social workers in child welfare settings. Br J Soc Work 2016; 46: 1076–1095.

40.

Yang

Chen

. A review of research on work withdrawal behavior. Hum Resour Dev China 2014; 17: 43–49.

41.

Bennett

Robinson

. Development of a measure of workplace deviance. J Appl Psychol 2000; 85: 349–360.

42.

Murphy

. Honesty in the workplace. Pacific Grove, CA:Brooks/Cole Pub. Co., 1993.

43.

Viswesvaran

. Absenteeism and measures of job performance: a meta-analysis. Int J Sel Assess 2002; 10: 12–17.

44.

Liu

Wang

, et al. A cross-level examination of the relationship between emotional labor, emotional state, and work withdrawal behavior. Chin J Manag 2013; 10: 545–551.

45.

Elfenbein

Curhan

Eisenkraft

, et al. Are some negotiators better than others? Individual differences in bargaining outcomes. J Res Pers 2008; 42: 1463–1475.

46.

Eisenberger

Huntington

Hutchison

, et al. Perceived organizational support. J Appl Psychol 1986; 71: 500–507.

47.

Huang

. The impact of social workers’ perceived organizational support on job burnout: The chain mediating role of public service motivation and emotional labor . PhD Thesis, Guangdong University of Foreign Studies, 2021.

48.

Ashcroft

Sur

Greenblatt

, et al. The impact of the COVID-19 pandemic on social workers at the frontline: a survey of Canadian social workers. Br J Soc Work 2022; 52: 1724–1746.

49.

Prasad

McLoughlin

Stillman

, et al. Prevalence and correlates of stress and burnout among US healthcare workers during the COVID-19 pandemic: a national cross-sectional survey study. EClinicalMedicine 2021; 35: 100879.

50.

Alnazly

Khraisat

Al-Bashaireh

, et al. Anxiety, depression, stress, fear and social support during COVID-19 pandemic among Jordanian healthcare workers. PLoS One 2021; 16: e0247679.

51.

Zhou

Wang

, et al. Burnout and well-being of healthcare workers in the post-pandemic period of COVID-19: a perspective from the job demands-resources model. BMC Health Serv Res 2022; 22: 284.

52.

Xie

Huang

Cheung

, et al. Job demands and resources, burnout, and psychological distress of social workers in China: moderation effects of gender and age. Front Psychol 2021; 12: 741563.

53.

Levenson

. Trauma-informed social work practice. Soc Work 2017; 62: 105–113.

54.

Huang

. Emotional labor: which is more effective, surface acting or deep acting? Acta Psychol Sin 2006; 38: 262–270.

55.

Lynch

. Applying creative research methods to develop understanding of empathy in social work: a relational practice gallery and relational practice poems. Eur Soc Work Res 2025; 3: 2–20.

56.

Hartwell

Steele

French

, et al. Aiding troubled employees: the prevalence, cost, and characteristics of employee assistance programs in the United States. Am J Public Health 1996; 86: 804–808.

57.

Higgins

. How has the professional capabilities framework changed social work education and practice in England? Br J Soc Work 2016; 46: 1981–1996.

58.

Zhan

Lin

. The risks of government purchasing public services and the construction of its control mechanism. Jianghai Acad J 2021: 138–145, 254–255.