251 Organizacija, V olume 58 Issue 3, August 2025 Research Papers 1 Received: 19th December 2024; Accepted: 27th February 2025 How Workplace Friendships Impact Burnout among Social Care Leaders: A Job Demands-Resources Framework Analysis Réka SCHUTZMANN 1,4* , Anna KOZÁK 1 , Klára SOLTÉSZ-VÁRHELYI 2 , Katalin NISTOR 1,3 , Beáta DÁVID 1,5 1 Semmelweis University, Faculty of Health and Public Services, Institute of Mental Health, Doctoral Division of Men- tal Health Sciences, Budapest, Hungary, réka.schutzmann@gmail.com (* Corresponding author) 2 Institute of Psychology, Pázmány Péter Catholic University, Budapest, Hungary 3 Data-Driven Health Division of National Laboratory for Health Security, Health Services Management Training Cen- tre, Semmelweis University, Budapest, Hungary 4 Archiescopical College of Veszprém, Veszprém, Hungary 5 HUN-REN Centre of Social Sciences, Institute for Sociology, Budapest, Hungary Background and purpose: The purpose of this study, guided by the Job Demands-Resources Model, is to investigate the role of workplace friendships in mitigating burnout. This research is notable for its unique focus on a relatively rare sample: social care leaders. These individuals play a crucial role in shaping and influencing social services, making their insights invaluable for understanding the challenges and opportunities within this sector. Methods: Using a cross-sectional and quantitative design, data were collected from a convenience sample of Hungarian social care leaders, including sociodemographic information, the Copenhagen Psychosocial Questionnaire (COPSOQ II), and profes - sional core discussion network (pCDN) questions. The analysis of 449 Hungarian social care leaders employs a saturated model of moderated mediation (controlling for age and gender) to examine how stress mediates the relationship between quantitative demands and burnout and how workplace friend - ships moderate this mediation effect. Results: The results indicate that stress significantly mediates the relationship between quantitative demands and burnout, with workplace friendships acting as a buffer under moderate stress levels. Having at least one workplace friend reduces the impact of stress on burnout; however, this protective effect diminishes under higher stress intensities. Conclusions: These findings underscore the importance of fostering quality and balanced workplace friendships rather than merely increasing the number of supportive relationships. Given the systemic challenges in Hungarian social care, these insights are particularly relevant for leaders seeking to improve workforce resilience and well-being. Keywords: Workplace friendship, Burnout, Stress, Quantitative demands, Social care leaders, Social work DOI: 10.2478/orga-2025-0015 1 Introduction The field of social work stands out as particularly vulnerable to burnout, affecting not only frontline social workers but also leaders in social care services (Gimén- ez-Bertomeu et al., 2024). There is a notable gap in re- search regarding burnout among social care leaders. This research gap exists primarily because most studies concen- 252 Organizacija, V olume 58 Issue 3, August 2025 Research Papers trate on frontline social workers (Maddock, 2023; Watson and Begun, 2024; Giménez-Bertomeu et al., 2024). In contrast, research focusing specifically on social care lead- ers is rare (Kozák et al., 2025; Mahara et al., 2024). This knowledge gap extends to the Hungarian context as well (Győri & Ádám, 2024; Győri & Perpék, 2021; Kopasz et al., 2024), limiting our understanding of how burnout im- pacts individuals in leadership roles. Effective caregiving systems rely heavily on competent social care leaders who navigate the complexities of managing social institutions, especially under challenging conditions. These leaders are entrusted with the responsibility of fostering the well-be- ing of employees who face heightened exposure to stress and burnout. The effects of burnout in leaders go beyond personal consequences, potentially harming the entire organization and negatively affecting coworkers. Hence, exploring factors that may act as buffers or protective mechanisms against burnout among social care leaders is imperative. Burnout, stemming from prolonged exposure to high levels of stress, manifests in emotional exhaustion, dep- ersonalization, and diminished personal accomplishment (Maslach & Leiter, 2016). Existing literature suggests that personal resources, such as workplace social support from colleagues or superiors, can mitigate the impact of feelings of isolation and lack of support, both antecedents of burn- out (Jenaro et al., 2007; Zeijen et al., 2024). Additionally, job resources like social support and performance feed- back could mitigate the adverse effects of job demands, such as cognitive challenges and interpersonal conflicts, on psychological distress and burnout (Bakker & Demer- outi, 2024). Research on post-transition of the political system changes in Hungary revealed a shift in ego network pat- terns from kin ties to non-kin ties dominating core dis- cussion networks (Albert et al., 2021). Additionally, it emphasized workplace friendships as a primary source of personal relationships in Hungary (Utasi, 1990). This highlights the significance and prevalence of workplace friendships in contemporary Hungarian society, impacting both professional and personal spheres. Workplace friendships, characterized by affective connections within formal organizational settings, are recognized as a common phenomenon with significant implications for employee performance and organizational outcomes (Methot et al., 2016; Chen et al., 2024). While numerous studies emphasize the benefits of workplace friendships for employees, we posit that the role of work- place friendships among leaders, particularly supportive relationships with colleagues, could significantly mitigate burnout. This research paper utilizes the Job Demands-Resourc- es (JD-R) framework to examine the interplay between workplace friendships, quantitative demands, stress, and burnout among social care leaders. It aims to contribute to the existing literature on the role of workplace friendships, specifically in social care leadership. 2 Literature Review This literature review uses the JD-R model as a frame- work to investigate previous research about workplace friendships, quantitative demands, stress and burnout. Our review highlights how workplace friendships can influence burnout and focuses on the importance of this dynamic for individuals and organizations. 2.1 The Job Demand-Resources Model The JD-R model is a psychological framework intro- duced by Bakker and Demerouti in the early 2000s and has since been commonly used in occupational and organiza- tional psychology (Bakker et al., 2024; Schaufeli, 2017). It is a comprehensive conceptual framework ó to measure burnout and subsequently expanded to incorporate work engagement (Schaufeli, 2017). The JD-R Model is based on the idea that every job has specific demands and re- sources, and the relationship between these demands and resources can significantly impact employees’ well-being and job outcomes (Bakker, 2024). The model proposes two main categories of factors in the work environment: job demands, which require effort and can be physically, psychologically, or emotionally taxing for employees, and job resources, which support employees in dealing with job demands and achieving their goals. Job resources can be tangible, such as access to training and equipment, or intangible, such as social support from co-workers and supervisors, autonomy, and opportunities for skill devel- opment. Job resources help employees cope with job de- mands, reduce stress, and enhance their motivation and job satisfaction (Bakker, 2024). The JD-R model suggests two underlying psychological processes initiated by demands and resources: the health impairment process, where job demands exceed job resources, leading to adverse out- comes such as stress, burnout, and health problems, and the motivational process, where sufficient job resources lead to positive outcomes such as increased job performance, well-being, and job satisfaction (Demerouti & Bakker, 2011). The model suggests that high job demands can lead to adverse outcomes unless employees have sufficient re- sources to cope effectively. Conversely, high job resources may facilitate positive outcomes despite high job demands (Berthelsen et al., 2018). The JD-R model’s interaction be- tween demands, resources, and adverse outcomes is wide- ly examined in various settings (Schaufeli, 2017; Bakker et al., 2003; Hu et al., 2011). Several studies compare the positive and negative effects of resources and demands on work outcomes. It is commonly found that resources 253 Organizacija, V olume 58 Issue 3, August 2025 Research Papers have a more powerful impact than demands by buffering or reducing work outcomes (Huang et al., 2022). All of this supports the idea that the JD-R model is a well-established theoretical framework for examining the roots of burnout and understanding the relationships between them. 2.2 Quantitative job demands and stress Increased work pressure is a global phenomenon (Van Veldhoven, 2024). In everyday language, quantitative de- mands are usually described by several other terms with slightly different associations or connotations, such as work pressure, workload, or speed. To address the lack of clear definitions, Van Veldhoven suggests employing the following working definition for quantitative demands: “Quantitative job demands constitute those elements of the work environment that concern the amount and speed of work to be performed and require physical and/ or psycho- logical effort” (Van Veldhoven, 2013, p.121). Thus, quan- titative job demands refer to the specific aspects of a job that can be quantified or measured in terms of the amount of work, effort, or output required from an employee (Van Veldhoven, 2024). The level of quantitative job demands in a role can vary significantly from one job to another and within different industries. While some degree of quantitative demand is a normal part of most jobs, excessive demands in these areas can lead to stress and burnout if employees do not have the necessary resources and support to cope with them (Mette et al., 2018; Montgomery et al., 2006). According to Lazarus and Folkman (1984), stress is present when we lack the resources to overcome difficult situations and events. (Demerouti & Bakker, 2011). The conservation of resources theory posits that individuals strive to acquire and maintain valuable resources (Zhang et al., 2019). The theory suggests that people experience stress or threats when faced with these resources’ actual or potential loss (Anthony-McMann et al., 2016). Similarly, the World Health Organization defines work-related stress as a response to work demands and pressures that exceed a person’s knowledge, skills, or ability to manage effectively (WHO, 2019). It has been suggested that job demands pos- itively impact, while job resources negatively affect work stress (Frank et al., 2017). Additionally, there is a consen- sus on the positive linkage between stress and burnout, and burnout is considered a severe feature of prolonged stress (Lloyd et al., 2002; Gorgievski & Hobfoll, 2008; Antho- ny-McMann et al., 2016). 2.3 Burnout According to the International Classification of Dis- eases (ICD-11), burnout is an occupational phenomenon caused by chronic stress that has not been effectively man- aged at the workplace (WHO, 2019). This definition is similar to Maslach’s theory (Maslach & Jackson, 1981). It explains that burnout arises when there is a prolonged mismatch between the individual and one or more work di- mensions, such as workload, control, rewards, community, fairness, and values (Maslach & Leiter, 2016). Burnout is one of the most common and severe psy- chosocial occupational hazards (Schaufeli et al., 2009; De- merouti, 2024). It is generally conceptualized as the result of permanent and chronic occupational stress and failure to cope with it (Guglielmi, 2001). It can range from acute fatigue to chronic exhaustion and associated problems (Edú-Valsania et al., 2022). While burnout is an individual response (Maslach & Leiter, 2016), it also has a detrimen- tal effect at the organisational level (Fernet et al., 2013). Burnout can have a wide range of consequences. It can lead to various physical health problems, such as type 2 diabetes, chronic fatigue, insomnia, headaches, and gas- trointestinal issues. Burnout is also often associated with mental health issues, including depression and anxiety, dissatisfaction with life, low self-esteem, and increased alcohol and tobacco consumption (Edú-Valsania et al., 2022). Individuals experiencing burnout may also develop feelings of helplessness and hopelessness. Additionally, it can decrease job performance, increase absenteeism, and reduce job satisfaction and engagement. Moreover, one adverse outcome of burnout may be increased turnover, with employees choosing to leave their organizations (Edú-Valsania et al., 2022). Social work is a highly demanding profession (Ran- tonen et al., 2016), associated with a high risk of burnout (Lloyd et al., 2002; Sánchez-Moreno et al., 2014). While frontline social workers are known to experience burnout (Maddock, 2023; Maslach & Leiter, 2016; Watson and Begun, 2024), research involving only managers is rare (Erera, 1992). Nowadays, the focus is more on the rela- tionship between leadership styles, attitudes, and employ- ee burnout (Kim & Lee, 2009; Maddock, 2023; Padín et al., 2021). Measuring burnout among leaders in the social care sector can be challenging because they primarily work with employees rather than clients. Nevertheless, they still work with people, just like frontline social workers. In addition, selecting an appropriate measurement tool for burnout also presents a persistent obstacle in academic re- search, mainly due to the significant variations observed across diverse occupational settings (Kristensen et al., 2005). The most widely recognized tool for evaluating burn- out is the Maslach Burnout Inventory (MBI) (Maslach & Jackson, 1981) and its subsequent versions. Maslach in- itially identified three dimensions of burnout: emotional exhaustion, depersonalization and reduced personal ac- complishment (Maslach & Leiter, 2016). The current for- 254 Organizacija, V olume 58 Issue 3, August 2025 Research Papers mat of the measurement tool remains three-dimensional; however, “emotional exhaustion” has been replaced by “exhaustion” and “depersonalization” by “cynicism” to better fit roles without direct human contact. In addition, “personal accomplishment” now measures “professional efficiency,” highlighting the challenges tied to professional competence (Kristensen et al., 2005). One-dimensional questionnaires such as the Copen- hagen Burnout Inventory (Kristensen et al., 2005) or the Shirom Melamed Burnout Measure (Shirom & Melamed, 2006) and the Burnout Measure (Malakh-Pines et al., 1981) narrow the focus by reducing burnout to a single core dimension: exhaustion. Nevertheless, despite the lack of definition and the debate around the construct (Schaufe- li et al., 2020), some common threads could be recognized, such as that exhaustion is the crucial and central compo- nent of the construct (Kiss et al., 2018). 2.4 The importance of workplace friendships The workplace is recognized as a social space (Sias et al., 2011), constituting a significant part of individuals’ lives and fostering friendships (Methot et al., 2024). De- spite the absence of a precise definition, attributed to the subjective perception influenced by individuals’ value sys- tems (Ibrahim & Dickie, 2010) and the blended nature of workplace settings and personal connections (Zarankin & Kunkel, 2019), various studies have identified key char- acteristics distinguishing workplace friendships from oth- er forms of workplace relationships (Methot et al., 2024; Nielsen et al., 2000; Zarankin & Kunkel, 2019;; Sias et al., 2003; Colbert et al., 2016). Workplace friendships, characterised by formal inter- actions layered with affective relationships, are a common organizational phenomenon that can impact employee performance and organizational outcomes (Methot et al., 2016; Chen et al., 2024). Defining workplace friendships can be challenging, which is why there is no widely agreed-upon definition of leader workplace friendships. One notable difference between employee and leader networks is the number of professional contacts outside work (Carroll & Teo, 1996). This suggests that leaders’ workplace friendships may transcend the confines of the physical workplace. In our study, we therefore measured the workplace friendships of leaders in a broader framework outside the physical boundaries of the workplace. While we often cannot choose our colleagues, we can choose our friends (Sias et al., 2003). Research indicates that workplace friendships evolve through informal and voluntary interactions involving genuine personal connec- tions beyond professional relationships (Zarankin & Kun- kel, 2019; Rumens, 2016; Sias, 2009), and organizational mechanisms can facilitate their development. Several or- ganizational factors, such as the physical proximity of the workplace and a climate conducive to supervisor support, collaborative behavior, tasks, and employee participation, can nurture workplace friendships (Zarankin & Kunkel, 2019). Consequently, workplace friendship is recognized as a multifaceted phenomenon (Methot et al., 2024) and is distinct from instrumental relationships, such as those be- tween supervisors or subordinates (Nielsen et al., 2000; Zarankin & Kunkel, 2019; Cao & Zhang, 2020; Dobel, 2001). Among other things, workplace friendship could positively impact task performance (Methot et al., 2016; Berman et al., 2002; Chen et al., 2024). It could increase well-being, job satisfaction, commitment and involvement, and team performance. It can also reduce stress and turn- over intention. Workplace friendship could also positively influence emotions (Zarankin & Kunkel, 2019). Simulta- neously, workplace friendship helps individuals to achieve their work goals and find access to potential job resources (Methot et al., 2016), which may also result in positive job outcomes. For example, workplace friendship can enhance instrumental and emotional support among employees, thus leading to higher employee job effectiveness (Yan et al., 2021). In addition, Colbert and co-authors (2016) dis- covered that friendship exhibited the strongest correlation with positive emotions experienced during work hours among different types of workplace relationships. Workplace friendship is also associated with high ac- cess to resources, a more diverse sources of information, better control in task accomplishment, and supportive in- teractions, e.g., emotional support, reliable personal feed- back, or even career strategizing (Methot et al., 2016). Ad- ditionally, it could also aid individuals in managing better work-related problems and stressful situations and nurture positive work attitudes (Yan et al., 2021). Furthermore, workplace friendship results in a stronger sense of belong- ing at work (Fasbender et al., 2023). Therefore, workplace friendship may be a key determinant of organizational ef- fectiveness (Yan et al., 2021). However, recent research indicates that workplace friendships may affect outcomes differently depending on their quality and type (Zarankin & Kunkel, 2019; Pille- mer & Rothbard, 2018). Developing friendships within the workplace (besides its positive impact) can lead to detrimental outcomes, such as excluding other colleagues and creating tensions with organizational policies. These adverse effects may arise due to the involuntary nature of such relationships and the exchange norms and instru- mental goals that may be involved (Pillemer & Rothbard, 2018). A growing number of researchers support that workplace friendship has a dual nature and that the adverse effects of workplace friendship can be harmful to organi- zations (Methot et al., 2016; Fasbender et al., 2023; Sias et 255 Organizacija, V olume 58 Issue 3, August 2025 Research Papers al., 2004; Pillemer & Rothbard, 2018; Choi & Ko, 2020). According to Methot and colleagues (2016), multiplex research, workplace friendship is a “mixed blessing” as it has both advantageous and detrimental outcomes, and they postulate that an inverted U-shape could depict workplace friendship and performance relationship: “Up to a certain point additional friendships correlate with better perfor- mance; but at that point, performance starts to decline due to the emotional labor and possible exhaustion from maintaining all these friendships.” (Zarankin & Kunkel, 2019, p. 52). Nevertheless, despite the opposing sides of friendship (workplace nepotism, gossip, or even disrup- tive behaviors (Jones & Stout, 2015), friendship has more advantages than disadvantages in the workplace (Song & Olshfski, 2008). As workplace friendship nurtures positive workplace outcomes inspired by the JD-R model, it could be cate- gorized as a valid social job resource (Yan et al., 2021). Additionally, as it is suggested that friendship and burnout are inversely related (Kruger et al., 1995; Doolittle, 2020), we assume that workplace friendship could impact burn- out directly and indirectly through various resources. 2.5 Hypothesis development Based on the existing literature, our research question is as follows: How do workplace friendships influence the relationship between quantitative job demands, stress, and job-related burnout among leaders? In line with the JD-R framework, we have developed six hypotheses to explore the interactions between workplace friendship, quantita- tive demands, stress and burnout. In our proposed mod- el, quantitative demands are the predictor variable, stress functions as the mediator, burnout is defined as the out- come, and workplace friendship is identified as a modera- tor (see Figure 1). Every job has an optimal level of quantitative de- mands. When these demands are too high or the recov- ery time is inadequate, they can adversely affect workers’ health, well-being, and job performance (Van Veldhoven, 2024). Furthermore, excessive job demands can increase stress (Mette et al., 2018; Montgomery et al., 2006), sug- gesting a clear connection between quantitative demands and stress levels (Frank et al., 2017). Therefore, the first hypothesis is that there is a positive relationship between the quantitative demands placed on leaders and their stress levels (H1) (see Figure 1, path a). The relationship between stress and job burnout is widely recognized. There is consistent agreement on the positive correlation between stress and burnout, with burn- out often seen as a significant consequence of prolonged and chronic stress (Lloyd et al., 2002; Gorgievski & Hob- foll, 2008; Anthony-McMann et al., 2016). Consequently, the second hypothesis proposed that there is a positive re- lationship between leaders’ stress levels and their job-re- lated burnout (H2) (refer to Figure 1, path b). Figure 1: The conceptual model of the moderated mediation between Quantitative Demands and Burnout with Stress as medi- ator and Workplace Friendship as moderator; Age and Gender were included as control variables in the model 256 Organizacija, V olume 58 Issue 3, August 2025 Research Papers The relationship between quantitative job demands and burnout is assumed to become more nuanced with var- ious mediators, particularly stress, as excessive demands paired with inadequate resources are widely recognized to generate significant stress (Van Veldhoven, 2024), a direct antecedent to burnout (Lloyd et al., 2002). Therefore, the third hypothesis posits that stress mediates the relationship between quantitative job demands and burnout (H3) (see Figure 1, path ab). Research findings suggest that workplace friendships can buffer stress through social and emotionally support- ive aspects and access to valuable resources (Methot et al., 2024; Fasbender et al., 2023). It is also postulated that workplace friendships are key determinants of orgainazia- tional effectiveness (Yan et al., 2021). This evidence sug- gests that workplace friendships could ease the pressure of high quantitative demands and prevent high stress levels. Consequently, the fourth hypothesis posits that workplace friendships negatively moderate the relationship between quantitative demands and stress (H4) (see Figure 1, path xw). Workplace friendship provides numerous advantages to employees, such as increased well-being, job satisfac- tion, stronger engagement, and improved team perfor- mance (Methot et al., 2016; Berman et al., 2002; Zarankin & Kunkel, 2019; Yan et al., 2021). Additionally, an inverse relationship is proposed between work-related burnout and friendships in private life (Kruger et al., 1995; Doolittle, 2020). All this supports the assumption that workplace friendship can be considered a protective factor against burnout. Therefore, the fifth hypothesis proposes that there is a negative relationship between workplace friendships and job-related burnout (H5) (see Figure 1, path w2). Research also indicates that these friendships could reduce the consequences of stress (Methot et al., 2024; Fasbender et al., 2023), which may moderate the effect of stress on burnout. Given the moderating role of workplace friendship in the relationship between stress and burnout, several questions arise regarding the nature of this effect. Research indicates that the impact of workplace friendship on work-related outcomes varies based on their quality and type (Zarankin & Kunkel, 2019; Pillemer & Rothbard, 2018). Therefore, the sixth hypothesis proposes that work- place friendship moderates the relationship between stress and burnout (H6) (see Figure 1, path mw). 3 Research methodology 3.1 Data collection Power analysis was performed using Monte Carlo sim- ulation (Zhang & Mai, 2023) to estimate the sample size required for the appropriate statistical power for the mod- erated mediation analysis with the following settings: α = .05, target Power = .80. A moderate strength (β = .30) was assumed for the regression coefficients included by the in- direct path, and the strength of the direct path was consid- ered to be zero. Additionally, moderate (β = .30) direct ef- fects of the moderator variable and weak interaction terms (β = .15) were assumed. A sample size of 345 participants was determined to be sufficient to detect the interactions on the indirect pathway. The questionnaire was distributed to Hungarian social care leaders, who participated in the training of the Sem- melweis University Health Services Management Training Centre in 2019 (Bálity et al., 2019; Ethical Approval ID: SE RKEB: 61/2019). All adult participants provided writ- ten consent, and participation was voluntary and anony- mous. Data was obtained through paper-based questionnaires among 667 social care leaders, the response rate was 82%. In addition to general demographic data, the questionnaire also included questions regarding organizational opera- tion-related data and a social network name generator. The scales of the (Quantitative demands, Stress and Burnout) Hungarian version of the COPSOQ II questionnaire (Nis- tor et al., 2015) were applied to measure Quantitative de- mands, Stress, and Burnout. The questionnaire was anon- ymous and voluntary. 3.2 Sample The total sample consisted of 547 Hungarian leaders of social care organizations. However, due to unusable re- sponses (where participants did not mention any confiden- tial professional relationships) and other missing values in relevant variables (considered to be Missing Completely at Random), the final sample size was reduced to 449. The leaders managed organizations with an average of 54.3 (SD = 98.3) employees. The gender ratio of the sample shows female predominance (female = 85.3%; male = 14.7%). 98.4% of the respondents have a higher education degree: 50.3% have a bachelor’s degree, 27.8% have a master’s degree, 19.8% have a postgraduate degree, and 0.4% have a doctorate. The average age of the leaders is 45.4 years (SD = 7.0, Min = 23, Max = 60), with an average of 21.5 years SD = 8.9) of work experience and 10.2 years (SD = 6.8) of management experience. 3.3 Measures The psychosocial work factors were measured by the Copenhagen Psychosocial Questionnaire (COPSOQ) II middle version (Pejtersen et al., 2009). The Hungarian version of the questionnaire was validated by Nistor and colleagues (2015). In the current study, we use the scales of Quantitative Demands as the predictor, Burnout as the outcome, and Stress as a mediator variable. Items were 257 Organizacija, V olume 58 Issue 3, August 2025 Research Papers measured by a five-point Likert scale (converted to 0-100) (Pejtersen et al., 2009). In our sample, the reliability of all scales was excellent: for Quantitative Demands Cron- bach’s α = .81 for Stress α = .87; and for Burnout α = .90. The study regarding the nature of workplace friend- ships was executed through the utilization of social net- work tools. Individuals’ network data was collected using a social network name generator (Burt et al., 2012). The name generator was recall-based (Pustejovsky & Spillane, 2009) and focused on the personal network’s professional, confidential relationship subset (Marin & Hampton, 2007) marked as pCDN. Respondents were asked to record the number of persons with whom they had confidentially discussed their workplace professional problems and con- flicts in the last six months, with the alter number limited to five persons (Merluzzi & Burt, 2013). The analysis cat- egorized the indicated persons as professional confidential relationships (pCDN). The name generator was combined with name interpreter (Stark, 2017) questions (recording the gender, age, type of connection, and the type of shared problems to characterize the relationship with the indicat- ed persons). More precisely, the respondents could mark whether they consider the indicated person their friend. The Workplace Friendship variable (WPF) represents the percentage ratio of friendships among the pCDNs en- tered in the name generator. 464 of the 547 respondents indicated at least one confidential professional relationship (M = 2.7, SD = 1.3), with an average WPF of 42.1% (SD = 40.3). Respondents without any confidential professional relationship were dropped from further analyses. 3.4 Data analysis After reviewing the descriptive statistics, Pearson cor- relation analysis was performed to explore the relation- ships between the WPF and the COPSOQ II variables. Following this, a moderated mediation analysis was con- ducted with Quantitative Demands as the predictor vari- able, Stress as the mediator, and Burnout as the outcome variable. The WPF variable was used as a moderator on paths “a” and “b” (see Figure 1). Age and Gender were included as control variables in the model. The distribution of the variables was found to be suitable for conducting a moderated mediation analysis. For the COPSOQ II var- iables and age, skewness ranged from -0.30 to 0.25 and kurtosis from -0.21 to 0.11, indicating a fairly normal dis- tribution. The distribution of the WPF variable was slight- ly platykurtic (K = -1.47), with a skewness of 0.31, which is within acceptable limits. Nonetheless, we chose Max- imum Likelihood as parameter estimator and Percentile Bootstrap to make the analysis robust to this slight devi- ation from the normal distribution. Standardized variables were used for the mediation analysis. All analyses were performed using JASP (0.18.3) (JASP Team, 2023). 4 Results Participants indicated that on average slightly more than a third of their confidential professional relation- ships are friendships (Median = 33.33, Mean = 42.10, SD = 40.25), bearing in mind that the distribution of WPF is somewhat flat with 38.5% of the sample reporting the pro- portion of friends as zero percent, and 24.1% as one hun- dred percent. They also reported moderate Quantitative Demands, Stress and Burnout (see Table 1). Pearson correlation was used to examine the correla- tions between the WPF variable and the COPSOQ II var- iables. The WPF variable showed a very weak significant negative association with Stress and a weak significant negative association with Burnout. Significant positive correlations were observed between each of the measured COPSOQ II variables (Table 1). Variable M SD 1 2 3 4 5 1. WPF 42.10 40.25 — 2. Quantitative demands 46.88 17.71 .01 — 2. Stress 42.55 18.99 -.12 * .25 *** — 4. Burnout 49.18 20.83 -.27 *** .14 ** .45 *** — 5. Age 45.43 6.96 -.10 * .05 -.05 -.03 — 6. Gender - - .05 .05 -.03 -.01 .05 Note. M and SD are used to mean and standard deviation, respectively. Values indicate Pearson correlation coefficients. For Gen- der 1 = male, 2 = female. * indicates p < .05; ** indicates p < .01; *** indicates p < .001. Table 1: Means, Standard deviations and Pearson correlations between WPF, Quantitative Demands, Stress and Burnout, Age, and Gender 258 Organizacija, V olume 58 Issue 3, August 2025 Research Papers Table 2: Results of the moderated mediation model To examine the influence of WPF on the relationship between Quantitative Demands, Stress, and Burnout, a mediation analysis was conducted, with Quantitative De- mands as the predictor, Burnout as the outcome, and Stress as the mediator. WPF was incorporated as a moderator on the mediated pathway (Fig. 1). Age and Gender were in- cluded as control variables in the model. In our model (Table 2), it was observed that Quantita- tive Demands exert a significant positive effect on Stress (see Figure 1, path a). Furthermore, the analysis confirmed that Stress has a significant positive effect on Burnout (see Figure 1, path b). The direct effect of Quantitative Demands on Burnout, after controlling for Stress, was found to be nonsignificant Path Variables β SE(β) p CI₉₅ a QD → Stress .25 .05 <.001*** .16; .34 w1 WPF → Stress -.13 .05 .005** -.22; -.03 xw QD*WPF → Stress .02 .04 .633 -.07; .11 c’ QD → Burnout .04 .04 .345 -.05; .13 b Stress → Burnout .42 .04 <.001*** .33; .52 w2 WPF → Burnout -.23 .04 <.001*** -.31; -.15 mw Stress*WPF → Burnout .09 .04 .031* -.01; .18 Conditional effect of Stress on Burnout at different values of WPF Path WPF β SE(β) p CI₉₅ bº 0% .33 .06 <.001*** .20; .46 33.3% (Median) .40 .04 <.001*** .31; .50 50% .44 .04 <.001*** .35; .54 100% .55 .08 <.001*** .37; .72 Conditional indirect effect of QD on Burnout through Stress at different values of WPF Path WPF β SE(β) p CI₉₅ abº 0% .07 .03 .004** .02; .14 33.3% (Median) .10 .02 <.001*** .05; 15 50% .11 .02 <.001*** .07; .16 100% .16 .05 <.001*** .07; .25 Note. QD indicates Quantitative Demands; WPF indicates the ratio of Workplace Friendship. Age and gender were included as covariates with no significant effect. N=449. β indicates the standardized estimate. Standard Errors and 95% Confidence Inter - vals of the standardized estimates were calculated with Bootstrapping. Values in bold are statistically significant and discussed in text. * indicates p < .05; ** indicates p < .01; *** indicates p < .001. (see Figure 1, path c’). However, the relationship between Quantitative Demands and Burnout was significantly me- diated by Stress (path ab). Additionally, WPF demonstrat- ed a weak yet significant negative effect on Stress (w1), as well as a significant negative effect on Burnout (w2). Regarding the interaction effects, the analysis revealed that WPF does not have a significant moderating effect on the relationship between Quantitative Demands and Stress (see Figure 1, path xw). However, the analysis revealed that WPF negatively moderates the effect of Stress on Burnout (see Figure 1, path mw). To examine the significant interaction between WPF and Stress, we calculated the effect of Stress on Burnout at different values of WPF (bº). Examination of these condi- 259 Organizacija, V olume 58 Issue 3, August 2025 Research Papers tional effects revealed that higher ratio of WPF is associ- ated with a steeper slope. Similarly, higher WPF is associ- ated with a stronger indirect effect (abº). The direction of moderation may seem surprising at first, but Fig. 2 clarifies this. It shows that with a low ratio of workplace friends, burnout can occur even at low stress levels, whereas with a higher ratio of workplace friends, burnout only increases when stress is high. However, when stress is high, the pro- tective power of WPF is weakened, and individuals with both low and high WPF are at high risk of burnout. This evidence challenges the original reasoning behind our hypothesis (H6), which generally assumed that the moderating effect of workplace friendships acts as a buffer against the negative effects of high levels of stress. Instead, the findings suggest a more nuanced relationship: the ab- sence of workplace friendships increases vulnerability to burnout even at low stress levels, while the presence of workplace friendships serves as a protective factor under moderate stress. However, as stress intensifies, the buff- ering effect of workplace friendships diminishes, leaving individuals with both low and high levels of workplace friendships at increased risk of burnout. 5 Discussions and further directions 5.1 Discussions Using a saturated model of moderated mediation (con- trolling for age and gender), this study aimed to explore the role of workplace friendships in the relationship between quantitative demands, stress, and burnout among social care leaders guided by the JD-R framework. The findings offer novel insights into the dynamics of stress and burnout and their implications for leaders in high-demand roles. According to the results, quantitative demands sig- nificantly predict stress, supporting Hypothesis 1 (H1). This aligns with existing literature, emphasizing that high quantitative demands increase pressure and workload, thus increasing stress levels (Van Veldhoven, 2013; Mette et al., 2018; Montgomery et al., 2006; Frank et al., 2017). Furthermore, stress was found to have a significant posi- tive effect on burnout, supporting Hypothesis 2 (H2). This outcome highlights stress’s critical role in exacerbating job-related burnout, consistent with the well-established Figure 2: Relationship between Stress and Burnout moderated by WPF Note. No friendship among pCDNs indicates 0% WPF; Some Friendship among pCDNs indicates 0