Corporate Culture, Leadership, and Pathological Relationships: A Moderated Mediation Model of Employees’ Well-Being

Abstract

This study investigates how corporate culture, work atmosphere, leadership, and pathological workplace relationships are linked to employees’ psychological well-being. Drawing on the JD-R framework and COR theory, we conceptualise supportive team dynamics and toxic interpersonal processes as parallel conditions shaping well-being at work. Survey data from 441 Slovak employees were analysed using a dual-moderated mediation model. The results show that corporate culture is positively associated with well-being mainly through its relationship with the work atmosphere, confirming the mediating role of workplace climate. Leadership moderates this pathway by attenuating the indirect association between culture and well-being. Contrary to expectations, pathological relationships did not exhibit a significant moderating effect, indicating that their role may be more complex and context-dependent. These findings extend existing models of employee well-being by incorporating negative social phenomena to an integrative framework. The study underscores the importance of examining both supportive and toxic aspects of organisational life simultaneously. Practically, the results suggest that fostering a positive culture, a healthy work atmosphere, and effective leadership practices can jointly strengthen employees’ psychological well-being.

1. Introduction

Since psychological well-being has a direct impact on employee health, productivity, and overall organisational success, it is crucial in the workplace (Chang, 2024; Meyerhuber, 2020; Michulek, 2021). High psychological well-being is important and socially significant because it promotes better job performance (Bansal, 2022), increased engagement (Sim et al., 2023), and a healthier, more positive workplace atmosphere (Weziak-Bialowolska et al., 2023). Despite existing studies on workplace bullying and psychological well-being, research on the moderated mediation between corporate culture, climate, leadership, and pathological relationships is sparse.

Previous studies (Búgelová, 2002; Branch et al., 2013; Rodic, 2015; Aristidou et al., 2020; Kubáni, 2011) have used a range of terms—such as bullying, mobbing, or harassment—to describe persistent harmful interpersonal behaviour at work. In this study, these phenomena are collectively referred to as pathological relationships. A potential synopsis of the connections between these concepts is illustrated in Figure 1.

Although various forms of toxic workplace behaviour have been studied under different labels such as bullying, mobbing, and harassment, the literature remains fragmented. The lack of a unified conceptualization limits the ability to compare findings and integrate these phenomena into broader theoretical models. To address this gap, the present study adopts the term pathological relationships as an overarching construct that captures the shared features of these negative social processes and examines its role within the JD-R framework.

In the workplace, this issue frequently worsens over time. However, the discussion around one-off incidents remains contentious due to their severity, the potential for chronic harm, and/or repetitive episodes involving different individuals (2020). According to Dujo López et al. (2020), workplace harassment underscores the devastating impact at both human and organisational levels.

According to earlier research, supportive leadership improves workers’ psychological health by encouraging acknowledgement, feedback, and a feeling of safety (Arnold et al., 2007; Ding et al., 2020). Leadership, work environment, and organisational culture all play mediating and moderating roles in shaping employees’ well-being, but these interactions are still little understood. The combined consequences of pathological relationships (Sjögren et al., 2006), corporate culture (Bommakanti & Swamy, 2024), and leadership have not been fully integrated into a unified framework, despite the fact that unfavourable social phenomena are known to negatively impact psychological health (Michulek et al., 2024).

By combining two significant viewpoints—the Conservation of Resources (COR) theory and the Job Demands–Resources (JD-R) model—our study fills this gap. COR emphasises how depletion of resources, such as through toxic workplace relationships, can lead to stress and decreased well-being. In turn, JD-R views supportive organisational elements and leadership as vital assets that reduce stress and improve well-being. Combining these theories, we investigate whether leadership and unhealthy relationships serve as moderators in the moderated mediation between corporate culture and work environment, and psychological well-being. We can better understand whether and how organisational issues influence employee well-being by adopting this dual-theory approach.

This study thus contributes to prior work by introducing the integrative construct of pathological relationships, testing a dual-moderated mediation model, and providing evidence from the underexplored Central European context.

The research aims to determine whether factors such as corporate culture, work atmosphere, pathological relationships, and leadership have an impact on psychological well-being and, at the same time, whether pathological relationships and leadership act as moderators in the mediating relationship between corporate culture, work atmosphere, and psychological well-being. This study addresses this gap by testing a dual-moderated mediation model linking corporate culture, workplace atmosphere, pathological relationships, and leadership to employee well-being.

2. Theoretical Framework

2.1. Corporate Culture and Working Atmosphere

Corporate culture, as an informal social system, shapes employees’ values, behaviour, and well-being through shared norms and relationships (Michulek et al., 2023). A crucial component of human resource management is corporate culture, which can be defined as behaviour patterns that are supported by individuals, procedures, and occasions that bring together the viewpoints of workers and establish what is seen as successful or appropriate in the company (Sadri & Lees, 2001; Graham et al., 2022). For instance, according to Chernukh (2022) and Hsieh et al. (2018), corporate culture encompasses outward manifestations of fundamental cultural values, including language, dress code, workplace design, symbols, beliefs, expectations, norms, and rituals. These subcultures develop because of the varied aims, work procedures, and leadership of various departments, which in turn create norms and values within each group. Communication, teamwork, and an organisation’s overall success can all be impacted by differences in corporate culture. Effective corporate cultures serve as a unifying factor; nevertheless, excessive cultural differences can impede cooperation and foster a sense of alienation (Hofstede, 1989; Wilson, 2001). According to the Job Demands–Resources (JD-R) model, work climate is the closest, collective impression of job resources, whereas corporate culture is an organisation-level resource that influences job resources (such as support, role clarity, and involvement). Therefore, in accordance with JD-R research on resource–well-being relationships, we model it as a mediator between culture and psychological well-being (Pecino et al., 2019). The entire perceptual, sensory, and emotional impression of an area or situation—particularly regarding how individuals feel and act at work—is referred to as the psychosomatic environment. However, it is crucial to keep in mind that it is not just about the physical setting; it is also about the atmosphere or the calibre of the interpersonal connections at work (Bressani & Sprecher, 2019). Subjective experiences, such as individual feelings or group dynamics, are then combined with objective elements, like lighting and organisation, to produce the work atmosphere (Edensor & Sumartojo, 2015). Organisational culture naturally contributes to the overall work climate, as it determines interpersonal relationships and work rules.

H1. 

The level of corporate culture is positively associated with the level of working atmosphere.

Corporate culture influences both individual choices and collective actions, as it encompasses a range of common employee ideas, attitudes, habits, and symbols. These concepts’ origins serve as the foundation for a later pattern of ideas that, in certain circumstances, are shared by all employees. This pattern of ideas is reflected in how employees generally interact with managers, other employees, customers, and so on, and establishes the culture in which the company’s employees work (Lorincová et al., 2020).

Organisational culture and leadership styles vary significantly depending on the situation, according to cross-cultural research. Younger workers in South Korea have criticised corporate practices that are performance-driven and hierarchical for fostering an unsupportive work environment (Postrelova, 2019; Kim et al., 2020). On the other hand, Croatian studies highlight the significance of job security, supervisor support, and cooperative relationships as essential components of a positive workplace culture (Pozega et al., 2013).

The findings of Alipour and Khoramian (2023) revealed that biophilic components (as part of organisational culture) have significant positive correlations with employee well-being, but no significant relationship was discovered between biophilic components and employee performance. We postulate the following, considering the mounting evidence that a robust and positive business culture promotes the psychological well-being of employees:

H2. 

The level of corporate culture is positively associated with employees’ psychological well-being.

Cultural values moderate the relationship between job demands and resources in terms of employee engagement and burnout. The significance of context in implementing the JD-R model is highlighted by the fact that, for instance, various cultural aspects (using Hofstede’s framework) can either strengthen or decrease these interactions (Rattrie et al., 2020; Brough et al., 2013).

Time pressure, competition, and job insecurity are common outcomes of the constant change, globalisation, and rising performance demands that organisations and people must deal with (Houtman et al., 2007). While a supportive work environment can improve employees’ psychological and physical well-being, corporate culture is crucial in helping them overcome these obstacles and accomplish their objectives (Shang, 2023; de Bruin et al., 2020). We postulate the following given the obvious correlation between psychological outcomes and supportive versus stressful work environments:

H3. 

The level of working atmosphere is positively associated with employees’ psychological well-being.

Systems are in place for employees to report poor mental health. Safety cues provide information about services employees can use to obtain respite or relief from risk, and management prioritises employee well-being over performance in a psychosocially safe environment (Law et al., 2011). Work-related pressure is the leading cause of stress in the United States, posing a severe threat to employee health and well-being (Aikens et al., 2014). Companies seek “high productivity and work quality.” It only works if employees can be promised a “good quality of life and well-being” (Tempel et al., 2004). For instance, Halabi and Kumral (2022), Kudryavtseva et al. (2018), Tempel et al. (2004), Sjögren et al. (2006), and others examined the work environment and psychological health of employees.

The JD-R model offers a robust framework for examining the interplay between leadership, resources, and job demands. By encouraging leadership that supports flexible work schedules and promoting physical spaces, this framework provides organisations with doable ways to protect the psychological health of their workforce (Rasak et al., 2024). One way to think of the office environment is as a combination of environmental needs and resources, encompassing both the physical space and the social climate. Workers actively influence their surroundings to reduce demands and optimise resources, which enhances the business environment (Rasak et al., 2024; Roskams & Haynes, 2021).

2.2. Pathological Relationships

The phrase “pathological relationships” is used in this study as a generalisation to describe a variety of unfavourable interpersonal dynamics in the workplace, such as bullying, mobbing, harassment, rudeness, and similar actions. Although the specific causes, manifestations, and possible legal ramifications of these phenomena vary, they all share one commonality: they pose psychosocial risks that threaten the psychological well-being of employees and disrupt normal organisational operations. Our decision to choose a more comprehensive, integrative term is both methodological and conceptual. From a conceptual standpoint, it enables us to recognise how these categories overlap and are frequently applied inconsistently in various disciplinary and cultural contexts. In terms of methodology, it mirrors our approach to measurement, as the Luxembourg Workplace Mobbing Scale (LWMS) measures broad trends of defamation, isolation, and criticism, rather than specific legal or cultural differences. As a result, the primary goal of this article is to examine how these concepts impact employees’ well-being, rather than to establish their definitional boundaries. However, we acknowledge that this integrative method may weaken theoretical precision, and as a result, we advise that future studies continue to separate the various causes and effects of toxic workplace behaviour.

Conceptually, pathological relationships differ from constructs such as corporate culture or work atmosphere, which reflect supportive organisational resources, whereas pathological relationships represent resource-depleting social dynamics.

Pathological relationships at work can be a stressful situation that has an impact on mental health (Sjögren et al., 2006). Specifically, a study by Annor and Amponsah-Tawiah (2020) examines the relationship between pathological relationships and subjective well-being to determine if high resilience moderates the effect of pathological relationships. According to a study by Hsu et al. (2019), pathological relationships at work have a detrimental impact on the well-being of hotel staff members. Research conducted in 2021 by Hayat and Afshari (2021) showed that workplace pathological relationships not only directly affect worker well-being but also, through raising worker burnout, indirectly lower worker well-being.

Resilience moderated the favourable effects of pathological relationships on felt victimisation and further moderated the negative impacts of perceived victimisation on employee well-being, according to Gupta and Bakhshi’s (2018) moderation analyses. Sprigg et al. (2019) discovered that witnessing pathological relationships harmed employees’ well-being (work-related depression and anxiety). The impact of pathological attitudes at the workplace on psychological well-being was discussed, for example, Koç and Keklik (2020); Manotas (2015); Minárová et al. (2020); Fiabane et al. (2015), and others.

The Conservation of Resources (COR) hypothesis posits that, unless workers have access to sufficient organisational or personal resources to offset these demands, resource depletion leads to decreased well-being (Barling & Frone, 2017; Kuriakose et al., 2019; Lesener et al., 2019).

Employees with greater exposure to workplace pathological relationships reported lower levels of quality of life and occupational health (Yoo & Lee, 2018). Given the abundance of research demonstrating the detrimental impact of pathological relationships on mental health, we postulate:

H4. 

The occurrence of pathological relationships is negatively associated with employees’ psychological well-being.

Job satisfaction alone partially mediated the relationship between workplace pathological relationships and task performance (Devonish, 2013). Finchilescu et al. (2019) found that higher levels of pathological relationships are associated with poorer work satisfaction and a greater likelihood of leaving.

Every culture is strong if it is desirable, employees understand it, and, most importantly, enterprise management practices it daily. A positive company culture’s performance is dependent on the behaviour of the enterprise’s management and successful mutual communication. It includes oppression, mobbing, physical aggression, and/or sexual harassment, all of which have a negative influence on an enterprise’s ability to foster a great culture (Minárová et al., 2020). Since the beneficial effects of corporate culture may be overshadowed by harmful workplace interactions, we hypothesise the following:

H5. 

Pathological relationships are expected to moderate the association between corporate culture and employees’ psychological well-being.

The investigation showed that organisational and social resources—like friendly coworkers and a positive workplace culture—consistently mitigated the negative impacts of pathological relationships. Personal resources, on the other hand, had minimal influence as modifiers (Farley et al., 2023). These results suggest that normally helpful organisational processes may be hampered by toxic social dynamics. To put it another way, even if a business cultivates a positive culture, the existence of unhealthy connections may interfere with the process by which the workplace culture converts cultural values into psychological advantages. Thus, our hypothesis is:

H6. 

Pathological relationships are expected to moderate the indirect association between corporate culture and employees’ psychological well-being through work atmosphere.

According to COR theory, stress is primarily caused by the loss of resources (such as job security, social support, or self-esteem), and the gain of resources becomes particularly significant after a loss has occurred (Hobfoll, 2001). Social stresses, such as pathological relationships at work, have a particularly detrimental effect on well-being. Social pressures frequently have more detrimental effects on resource loss than on resource gain (Hobfoll, 2001; Meier et al., 2023).

2.3. Leadership

Regardless of the approaches they take or the social support networks they participate in, leaders have little impact on the long-term effects on their well-being that result from being subjected to pathological relationships. This realisation is noteworthy, as past research has demonstrated that leadership is crucial in managing pathological relationships and promoting employee well-being (Inceoglu et al., 2018; Woodrow & Guest, 2017).

The results of the data analysis showed that instances of unethical behaviour had a stronger influence on employee well-being than ethical behaviours. By highlighting the importance of resource loss over resource gain in determining employee well-being, these findings reinforce the theoretical approach of resource conservation (Ahmad et al., 2020). According to Perko et al. (2016), employees with higher levels of well-being across time reported more favourable leadership behaviours at both time points, and changes in employee well-being were reflected in changes in perceived leadership.

According to research, leadership—particularly supportive, empowering, and authentic styles—is a critical job resource in the JD-R paradigm (Tummers & Bakker, 2021; H. J. Wang et al., 2017; Schaufeli, 2015). In addition to directly providing resources, leaders also provide psychological safety, cushion job demands, and facilitate job crafting, all of which improve organisational results and employee well-being.

Han et al. (2023) discovered that empowered leadership was positively associated with leader recovery, which mediated its favourable effect on leaders’ psychological well-being. Zeike et al. (2019) found that improved digital leadership skills were strongly connected with higher well-being. The findings of Samul (2024) demonstrated that spiritual leadership has an indirect impact on workers’ professional engagement through their spiritual well-being. According to Nielsen and Daniels (2012), transformational leadership should be researched as a group phenomenon, considering how group-level perceptions may affect followers’ well-being, and leaders should recognise that there is an “I” in groups as well as a “we” in groups. The relationship between leadership and psychological well-being has been investigated by Weiss et al. (2018), Kang et al. (2023), and Curral et al. (2023), among others. We make the following hypothesis, considering the substantial body of research showing the beneficial effects of different leadership philosophies on workers’ psychological health:

H7. 

Leadership is positively associated with employees’ psychological well-being.

The theory of transformational leadership, comprising four dimensions—inspirational motivation, intellectual stimulation, individual consideration, and idealised influence—has a positive impact on the psychological well-being of employees (Arnold, 2017). Some research has shown that transformational leadership improves well-being through mediators such as empowerment (Hasanein & Elrayah, 2025), meaningful work (Arnold et al., 2007), and team effectiveness (Nielsen et al., 2009). Individual characteristics and organisational culture may have different effects (Lindert et al., 2022). These results suggest that leadership may relate how corporate culture translates into well-being outcomes, in addition to having a direct impact on psychological well-being. Consequently, our hypothesis is:

H8. 

Leadership is expected to moderate the association between corporate culture and employees’ psychological well-being.

The work environment may also be shaped by leadership, which could strengthen or diminish the relationship between psychological health and company culture. Transformational leaders, for instance, establish welcoming and inspiring environments that promote collaboration and clear goals (Dextras-Gauthier et al., 2023; Monteiro & Joseph, 2023; Clarke et al., 2025; Khan et al., 2024). Therefore, we hypothesise the following:

H9. 

Leadership is expected to moderate the indirect association between corporate culture and employees’ psychological well-being through work atmosphere.

The proposed model assumes that the relationship between corporate culture and psychological well-being operates indirectly through work atmosphere and is further conditioned by leadership and pathological relationships as moderating factors.

3. Materials and Methods

3.1. Data Collection and Sample Size

To investigate unhealthy workplace relationships and their effects on workers, a questionnaire survey was given to employees of businesses in the Slovak Republic. An online email poll that guaranteed privacy and voluntary participation was used to gather data. Although online surveys generally have lower response rates, this strategy was selected due to its effectiveness and affordability (Michaelidou & Dibb, 2006). In accordance with accepted guidelines for online survey design, participants were briefed on the study’s objectives, confidentiality, and feedback opportunities in order to enhance the quality of the data (Bosnjak & Batinic, 2002; Ganassali, 2008). Purposive sampling was used to select individual companies that were part of the research. The largest companies operating in the Slovak Republic, based on the number of employees, were identified from the Finstat website.

The questionnaire consisted of three parts in total, namely a general part, a part focused on the work environment, and a part devoted to pathological relationships in the workplace. In the general part, the respondents were asked six basic questions focused on the gender, age, and experience of the respondents, the classification of the company in the category according to SK NACE (Slovak Statistical Classification of Economic Activities), and size criteria, and a question focused on the gender of the head employee of the work group. The second part was devoted to characterising employees’ perceptions of satisfaction with selected work environment factors (4 factors) and LWMS factors (5). Likert scales with five options were used to express satisfaction or the occurrence of individual factors, either in the LWMS or the work environment. According to Steffgen et al. (2016), the LWMS is a general measure that can be applied in various work contexts because it is not influenced by respondents’ age or the industry in which they work. It made it feasible for us to apply the LWMS in our study. In line with our research, which has shown that individual factors function as moderators in the relationship between working conditions and psychological well-being in the workplace, the authors also demonstrated a mutual correlation between individual factors and factors related to the quality of work, as well as the level of psychological risk. Consistent with our research focus on the occurrence of toxic interpersonal dynamics, LWMS responses were recoded into a binary indicator (0 = no occurrence; 1 = any occurrence). This threshold-based operationalization aligns with the JD-R/COR view that demands become salient once present and was preferred given the floor-skewed distribution of LWMS in our data. We note that dichotomization improves interpretability but limits scale-based validity assessments; this trade-off is acknowledged among the study’s methodological limitations.

The last part of the questionnaire survey focused on pathological relationships in the workplace, specifically whether respondents had witnessed such behaviour, what type of pathological relationships were present in the workplace, their manifestations, the consequences for the victim, and the solutions to pathological relationships within the company.

The questionnaire was created using the Google Forms application, used for online data collection. Since HR departments handled internal distribution independently, it is impossible to pinpoint the exact number of businesses contacted and people reached. Nonetheless, the poll link is believed to have reached more than 1000 businesses across various industries. A purposive sampling approach was used to recruit employees from large Slovak companies operating across various industries. A total of 441 employee responses were gathered, and all the replies were retained for analysis, as no exclusion criteria were applied. The largest companies addressed came from the Finstat database. Therefore, the number of responses is relevant to the survey conducted. The data were collected from 8 January 2024, to 15 February 2024. All questionnaires were completed correctly by the respondents; therefore, no questionnaires were excluded.

3.2. Ethical Considerations

We are aware of the importance of ethics in conducting scientific research; therefore, ethical considerations are carefully addressed throughout the entire research. Respondents were informed, before completing the questionnaire, that the results would be used for scientific purposes; they were also informed about the research objectives and assured of the anonymity, confidentiality, and voluntary nature of their participation. Respondents could decide to withdraw from the research at any time while answering the questionnaire survey questions without penalty. This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of UNIZA (Directive 207, dated 31 May 2024).

3.3. Moderated Mediation

The social and behavioural sciences have increasingly used moderated mediation analysis over the last decade. It is crucial to ensure that the moderated mediation analysis does not produce erroneous results, given its extensive use (Ng et al., 2024). While there are numerous intricate causal relationships, the presence of mediators or moderators can be taken into consideration while conducting a mediation or moderation analysis. Furthermore, a mediation-moderation study can be carried out in the presence of both mediators and moderators. Numerous disciplines, including social science, medical science, and natural science, among others, address these issues (Yoon, 2020).

When a moderator variable Z or W varies in such a way that it influences the direction or strength of the mediation impact of X on Y through M, this is known as moderated mediation (L. Wang & Preacher, 2015). According to the mentioned, we created

Table 1. Variables of moderated mediation.

Variable Name	Variable Type	Note
Corporate Culture	Independent variable	X
Psychological Well-being	Dependent variable	Y
Working Atmosphere	Mediator variable	M
Pathological Relationships	Moderation variable	W
Leadership	Moderation variable	Z

Source: own elaboration.

, which lists the variables used in the research, their types, and notes.

We tested a dual-moderated mediation model using PROCESS v4.0 (Hayes, 2018; Model 17). In this model, corporate culture (X) influences psychological well-being (Y) indirectly through working atmosphere (M). Pathological relationships (W) and leadership (Z) were specified as moderators of the indirect effect. Based on COR and JD-R theory, we specified a dual-moderated mediation model (Figure 2), in which corporate culture influences psychological well-being via working atmosphere, with pathological relationships and leadership moderating the indirect path.

For estimating indirect and conditional effects, we used the PROCESS macro in SPSS with bootstrapping procedures. Specifically, 5000 bootstrap resamples with 95% bias-corrected confidence intervals were generated. Bootstrapping is particularly recommended in moderated mediation analyses, as it does not rely on the assumption of normality of the indirect effects (Hayes, 2018). This approach thus provides more robust and reliable estimates compared to traditional parametric methods.

3.4. Assumptions of Moderated Mediation

Moderated mediation has several assumptions necessary for performing regression analyses. These assumptions involve ensuring the independence of observations, that error values are normally distributed, and the homoscedasticity of error values. Independence of observations was assumed due to cross-sectional design.

3.4.1. Normal Distribution

To verify normality, we will use the histogram shown in Figure 3, which utilises regression-standardised residuals and the dependent variable of psychological well-being. We are checking whether the error terms of the residuals follow a normal distribution or not. As shown in Figure 3, it can be considered very close to a normal distribution. There is no indication that the error term deviates from normality. The regression-standardised residual more or less follows the normal distribution.

The probability plot (p-p plot) shows us the expected cum prob and the observed cum prob. The data points should lie along the 45-degree line. If the data lie close to this line, we can consider it a normal distribution. According to Figure 4, there is some deviation from normality at the bottom and middle of the line.

According to Švábová et al. (2022), if the set has a large range (441 measurements, which is significantly more than 50), the test requiring normality is sufficiently robust against the violation of the mentioned assumption, and the results can therefore be considered correct.

Bootstrapping with 5000 bootstrap samples and 95% bias-corrected (BC) confidence intervals (CIs) was employed to interpret direct and indirect effects, as this method enables direct testing of indirect effects while remaining robust to violations of traditional assumptions, such as normality and linearity between variables (Hayes, 2018).

3.4.2. Homoscedasticity of Error Values

We will use the Levene test to verify homoscedasticity. According to

Table 2. Levene’s Test of Equality of Error Variances.

Levene’s Test of Equality of Error Variances
Dependent variable: Psychological well-being
F	df1	df2	Sig.
3.383	124	316	0.000
Tests the null hypothesis that the error variance of the dependent variable is equal across groups.

Source: own elaboration.

, the significance of Levene’s test is F (124, 316) = 3.383, p < 0.05. The value of Levene’s test is 0.000, which is less than level of defined significance level of 0.05. Thus, the null hypothesis is maintained, and there is difference between the variances of the three groups. Thus, the three samples come from populations with the same variance. However, Rieder et al. (2019) found the data to be heteroscedastic with non-normal residuals in their research, so they used the PROCESS HC3 estimator macro to allow for heteroscedasticity-consistent inferences.

To account for possible heteroskedasticity, we used robust standard errors based on the HC4 estimator (Cribari-Neto, 2004). This approach ensures consistent and unbiased parameter inference even when the homoscedasticity assumption is not met.

3.5. Data Analysis

A total of 262 (59.41%) respondents were female, and 179 (40.59%) respondents were male. The respondents’ ages were classified into five categories. The most numerous age group was the 36–45 age group (150; 34.01%). A total of 109 (24.72%) respondents represented the second-largest age group, aged 26–35. It was followed by the 46–55 age group with 102 (23.13%) respondents. The least represented age groups were 56 and over (60; 13.61%) and 18–25 (20; 4.54%). The majority of respondents were women aged 36–45 (95; 36.26%), while the least represented age group was 18–25, with 16 women, which accounted for 6.11% of the total number of women. For men, the most numerous categories were, as for women, the group aged 36–45 years, with 55 respondents, representing 30.73% in relative terms. The smallest group was again respondents aged 18–25, namely 4 (2.23%). In terms of length of experience, most respondents had 19 or more years of experience (179), which represents 40.59% of the total share of respondents. Subsequently, 74 people had 4–8 years of experience, and 14–18 years (16.78%). Seventy-one respondents had 9–13 years of experience (16.10%). The fewest people, 43 (9.75%), had 0–3 years of experience. Women most often had 19 or more years of experience (92; 35.11%) and 4–8 years (52; 19.85%). Among men, respondents with experience of 19 and more (87; 48.60%) and 14–18 years (35; 19.55%) prevailed. In terms of age, there is a direct correlation, i.e., the older the age, the more years of experience one has.

Steffgen et al. (2016) established the Luxembourg Workplace Mobbing Scale (LWMS), which has five dimensions—criticism, ignoring, pointless chores, ridiculing, and conflicts—as the basis for measuring abnormal relationships. Using straightforward questions with a 5-point Likert scale, other dimensions, such as leadership, company culture, work environment, and psychological well-being, were also assessed.

First, we conduct a reliability analysis to assess the internal consistency of the data. The results of the reliability analysis (

Table 3. Results of reliability analysis.

Reliability Statistics
Cronbach’s alpha	N of items
0.935	5

Source: own elaboration.

) indicate that the data can be considered reliable, as the Cronbach’s alpha result is 0.935. Statistical authors consider values higher than 0.65 or 0.8 to be reliable, depending on the author. The result of the analysis meets both stated values. It is true that the closer the value of Cronbach’s alpha is to 1, the higher the reliability. We can also see from this table that we examined five variables.

The majority of the model’s constructs—leadership, corporate culture, working environment, and psychological well-being—were assessed using single-item indicators. These single-item measures are therefore inapplicable to conventional reliability indicators such as Cronbach’s alpha and convergent/discriminant validity measures (e.g., AVE or Fornell-Larcker criterion). Simplified scales for measuring factors such as corporate culture, leadership, work environment, and psychological well-being were used mainly to maintain an acceptable questionnaire length and, at the same time, prevent “response fatigue,” as many respondents discourage completing long questionnaire surveys (Jeong et al., 2022; Neuert, 2021). There are several scales to measure the above constructs, such as the Corporate Safety Culture Scale (Colarossi, 2012), the Organizational Culture Scale (Ghosh & Srivastava, 2014), and the Corporate Entrepreneurship Index (Bau & Wagner, 2015), for corporate culture. Leadership can be measured using the Multifactor Leadership Questionnaire (Chakrabarty, 2014) and, Wise Leadership Questionnaire (Hassi & Storti, 2023). The Work Environment Scale (Abraham & Foley, 1984) or the Group Atmosphere Questionnaire (Posthuma, 1970) can be used to assess the work atmosphere. Ryff’s Scales of Psychological Well-being (Cui, 2005), the WHO-5 Well-Being Index (McDowell, 2010), and the COMPAS-W (Gatt et al., 2014) can be used to measure psychological well-being. Some scales are designed for specific industries or specific cultural contexts of countries and therefore cannot be applied to our research. Another problematic area is that some scales contain factors that are subsequently part of another scale (with differing names), which would lead to inconsistent responses for the same respondent across different measurement scales. For the same respondent, this may result in conflicting or non-replicable results from several instruments (Walkey & McCormick, 1985; L. Xu, 2019).

It has been demonstrated that single-item assessments of psychological well-being at work are reliable and useful substitutes for longer scales, particularly when concision is needed. Research shows that single-item measures of well-being are equivalent to multi-item scales in both students and working adults, and they can predict significant variance in well-being outcomes. For quick evaluations or when survey length is a concern, these metrics are particularly helpful (Williams, 2014; Williams & Smith, 2016). There is a lack of well-defined single-item measurements for broad constructs such as workplace atmosphere and company culture. According to certain research, single-item operationalizations are occasionally used to measure organisational practices and climate; however, multi-item scales are generally considered more reliable (Finsel et al., 2021). According to Knox et al. (2018), single-item questions about work climate have been used in extensive surveys and demonstrate concurrent validity with established measures; nonetheless, they do not fully reflect the complexity of the concept.

The study’s methodology was limited using single-item indicators rather than validated multi-item scales to measure several dimensions, including corporate culture, leadership, work climate, and psychological well-being. Single-item measures may be appropriate for well-defined constructs, according to prior research (Wanous et al., 1997; Bergkvist & Rossiter, 2007), although this method limits the capacity to evaluate validity and reliability. Thus, established multi-item measures like the Multifactor Leadership Questionnaire (Hunt, 1999; Lowe et al., 1996) for leadership, Newman (1977) for work environment, Ryff and Keyes (1995) for psychological well-being, and Kalliath et al. (1999) are well-established, widely used in organisational research, and highly cited in the scientific literature, which further underlines their methodological robustness and suitability for future studies.

Five items from the Luxembourg Workplace Mobbing Scale (LWMS) were used to measure the construct of unhealthy connections, and the results showed good internal consistency (Cronbach’s alpha = 0.935). Only this multi-item construct was subjected to reliability testing because it was the only one that warranted it.

4. Results

Model 17 from the extension of the statistical programme IBM SPSS 25, Process 4.1, was used to evaluate hypotheses H1 to H9. It is a moderated mediation model with one variable acting as a mediator and two variables acting as moderators. The entire model is performed at the significance level of α = 0.05.

Table 4. Association between corporate culture and working atmosphere.

Outcome Variable
Model Summary
R	R-sq	MSE	F (HC4)	df1	df2	p
0.7075	0.5006	0.6753	352.6481	1.000	439.000	0.000
Model
	Coeff	SE (HC4)	t	p	LLCI	ULCI
Constant	0.4781	0.0881	5.4255	0.000	0.3049	0.6513
Corporate Culture	0.7046	0.0375	18.7789	0.000	0.6308	0.7783

Source: own elaboration.

shows that R² = 0.5006, F (1, 439) = 352.6481, p < 0.05. Coeff values represent unstandardized regression coefficients (b). The association between corporate culture and the working atmosphere, with a coefficient value of 0.7046, which we will work with later. It is possible to state that SE = 0.0375, LLCI = 0.6308, ULCI = 0.7783, p < 0.05. The value of the standard deviation is 0.0375. The p-value of the test is zero and is less than the significance level of α = 0.05. From this, it can be concluded that we reject the null hypothesis and accept the alternative hypothesis H1. The value of the coefficient 0.7046 is positive. We have demonstrated that the level of corporate culture has a positive impact on the working atmosphere. It means that a better corporate culture leads to a higher level of psychological well-being among employees. Both the lower (LLCI) and upper (ULCI) confidence intervals do not contain zero; therefore, it is possible to say that the effect is statistically significant.

Based on the p-value results (

Table 5. Association between variables and psychological well-being.

Outcome Variable
Model Summary
R	R-sq	MSE	F (HC4)	df1	df2	p
0.8981	0.8067	0.3161	459.1341	8.000	432.000	0.000
Model
	Coeff	SE (HC4)	t	p	LLCI	ULCI
Constant	−0.2783	0.1339	−2.0787	0.0382	−0.5414	−0.0152
Corporate Culture	0.0390	0.0925	0.4216	0.6735	−0.1428	0.2208
Working Atmosphere	0.6355	0.1158	5.4870	0.0000	0.4079	0.8631
Pathological Relationships	0.4283	0.1746	2.4523	0.0146	0.0850	0.7715
Interaction 1	−0.0622	0.0581	−1.0720	0.2843	−0.1764	0.0519
Interaction 2	0.0486	0.0563	0.8637	0.3882	−0.0620	0.1592
Leadership	0.3725	0.1676	2.2227	0.0268	0.0431	0.7019
Interaction 3	0.0787	0.0644	1.2206	0.2229	−0.0480	0.2053
Interaction 4	−0.1459	0.0619	−2.3574	0.0188	−0.2675	−0.0243
Interaction 1: corporate culture and pathological relationships
Interaction 2: work atmosphere and pathological relationships
Interaction 3: corporate culture and leadership
Interaction 4: work atmosphere and leadership

Source: own elaboration.

) for the moderated mediation of the variables work atmosphere, pathological relationships, and leadership, it can be concluded that we reject the null hypothesis. We therefore accept the alternative hypotheses H3 and H7. There was a statistically significant positive association between psychological well-being and the occurrence of pathological relationships (b = 0.4283, p < 0.05), indicating a significant positive relationship. The data indicate that people who reported having disordered relationships had greater levels of psychological well-being, despite the initial hypothesis (H4) predicting a detrimental impact. Model summary states R² = 0.8067, F (8, 432) = 459.1341, p < 0.05. The level of working atmosphere has a positive impact on the level of psychological well-being of employees (b = 0.6355, SE = 0.1158, LLCI = 0.4079, ULCI = 0.8631, p < 0.05), while the value of the standard deviation is 0.1158 and the value of the coefficient is 0.6355 (positive value), and according to CI it is significant relationship. If the working atmosphere is higher, psychological well-being is also higher. Pathological relationships have a positive impact on the psychological well-being of employees (b = 0.4283, SE = 0.1746, LLCI = 0.0850, ULCI = 0.7715, p < 0.05). The value of the standard deviation is 0.1746. The coefficient is 0.4283. That means the fact if its occurrence of pathological relationships is higher, the psychological well-being of employees is better. The level of leadership has an impact on the level of psychological well-being of employees (b = 0.3725, SE = 0.1676, LLCI = 0.0431, ULCI = 0.7019, p < 0.05). It means that a higher level of leadership leads to better psychological well-being among employees. The coefficient is 0.3725 (positive relationship), and the standard deviation is 0.1676. The lower (LLCI) and upper (ULCI) confidence intervals do not contain zero; therefore, it is possible to say that the effect is statistically significant.

In the case of corporate culture (b = 0.0390, SE = 0.0925, LLCI = −0.1428, ULCI = 0.2208, p > 0.05, p = 0.6735) and interaction 1 (b = −0.0622, SE = −0.0622, LLCI = −0.1764, ULCI = 0.0519, p > 0.05, p = 0.2843), the significance level is higher than α = 0.05. Therefore, we accept the null hypothesis. It can be concluded that corporate culture has no effect on the psychological well-being of employees, and at the same time, pathological relationships do not moderate the relationship between corporate culture and the psychological well-being of employees. Interaction 2 tells us that there is no moderating relationship (b = 0.0486, SE = 0.0563, LLCI = −0.0620, ULCI = 0.1592, p < 0.05), because the p-value (0.3882) is higher than 0.05, between the occurrence of pathological relationships at the workplace and the variables work atmosphere and psychological well-being at the workplace.

Interaction 3 has the p-value greater than the significance level (b = 0.0787, SE = 0.0644, LLCI = −0.0460, ULCI = 0.2053, p > 0.05), so leadership does not have a moderating effect on the relationship between corporate culture and the psychological well-being of employees, because the p-value (0.2229) is higher than 0.05. The value of the standard deviation is 0.0644, and the value of the coefficient reached the level of 0.0787. Interaction 4 reveals a moderating relationship between the supervisor’s approach and the variables of work atmosphere and psychological well-being at the workplace (b = −0.1459, SE = 0.0619, LLCI = −0.2675, ULCI = −0.0243, p < 0.05). The p-value (0.0188) is lower than 0.05. The value of the standard deviation is 0.0619, and the value of the coefficient reached the level of −0.1459. The table also contains the R-squared value, which represents the coefficient of determination. The value of the coefficient is 0.8067, indicating that the model’s quality is 80.67%.

We still have two unanswered hypotheses, H6 and H9. To decide on the hypotheses mentioned, we must look at the confidence intervals (

Table 6. Conditional indirect effects.

Indices of Partial Moderated Mediation
	Index	BootSE	BootLLCI	BootULCI
Pathological relationships	0.0342	0.0347	−0.0344	0.1011
Leadership	−0.1028	0.0359	−0.1680	−0.0253

Source: own elaboration.

). The upper (BootULCI) and lower (BootLLCI) bootstrap intervals contain zero. Therefore, we accept the null hypothesis, and thus the occurrence of pathological relationships does not have a moderating effect on the mediating relationship of work atmosphere between corporate culture and employees’ psychological well-being (BootSE = 0.0347, BootLLCI = −0.0344, BootULCI = 0.1011). According to the findings, leadership considerably moderates the indirect association between psychological well-being and corporate culture through the work environment (BootSE = 0.0359, BootLLCI = −0.1680, BootULCI = −0.0253). In contrast to our original hypothesis, which expected an intensifying role of leadership, the effect is negative. Rather, the beneficial indirect association is weakened due to the moderating role of leadership. This counterintuitive result could also reflect measurement issues, as leadership was assessed with a single item, or suppression by correlated variables. Alternatively, strong leadership may reduce variance in the workplace atmosphere, making the moderation effect appear attenuated. As a result, H9 is only partially supported; it is statistically validated but goes against what theory would have expected. From the JD-R perspective, this pattern may reflect the role of leadership as a stabilising job resource: strong leadership is associated with more stable working atmosphere levels, which may attenuate rather than amplify the association between cultural resources and well-being.

Based on the above results, we created a new diagram for moderated mediation, which already includes coefficients and standard deviations. The so-called Path diagram is given in Figure 5.

5. Discussion

5.1. Theoretical Foundations and Methodological Framework

James and Brett (1984) reduced the additivity condition of mediation by incorporating a moderator into the model and attempted to demonstrate the presence of a moderated mediation model. According to Ng et al. (2024), mediation and moderation models are among the most employed statistical models in the social and behavioural sciences. They help to uncover the relationship’s underlying mechanisms and boundary conditions. Both models have undergone extensive testing over the last three decades. The authors identified an increasing trend in the usage of moderated mediation.

Furthermore, according to the Google Scholar Citation Index, significant methodological articles discussing the moderated mediation model have received 22,971 citations. The above provides a basis for the use of moderated mediation analysis in our research. The fact that the current conclusions are based on cross-sectional survey data should be emphasised. As a result, they show correlations rather than causes. Although the COR and JD-R frameworks provide causal logic, our design precludes us from asserting strong causality. To examine the temporal ordering of these interactions, more experimental or longitudinal studies are required.

5.2. The Impact of Corporate Culture and Work Atmosphere on Psychological Well-Being

We have confirmed that corporate culture has an impact on the working atmosphere. This outcome is consistent with the JD-R paradigm, which holds that business culture is associated with employees’ views of collaboration, support, and trust as a distal job resource. The work environment reflects these proximate resources and promotes psychological well-being. For example, L. F. Wang and Lin (2007) state that one of the most important factors influencing technological innovation is an innovative company culture that fosters new thinking and an ideal working atmosphere for corporate personnel.

Employee health and psychological well-being can be improved by encouraging them to participate in sports and engage in regular physical activity at work. Corporate culture has a tremendous impact on employees’ ability to achieve their goals. Those who are not properly motivated to achieve their goals are likely to have impaired psychological health (Shang, 2023). The relationship between corporate culture and well-being has been investigated, for example, by Haga et al. (2024), who contributed to the research issue by identifying practical implications for the positive effects of a strong corporate culture on employees and society. For example, Orsini et al. (2024) confirmed the impact of corporate culture on employee well-being, but they identified additional factors, including job dissonance, relationships with colleagues, achievement, professional development, relationships with superiors, status, workload, perks, feedback, workspace diversity, and pay. Many of these factors are an inseparable part of the working atmosphere, which supports our results.

From a COR standpoint, the flow of cultural resources across the workplace demonstrates how trust and organisational support function as cycles of resource gain that improve workers’ mental health.

5.3. Pathological Relationships and Their Complex Impact on Psychological Well-Being

According to Ribeiro et al. (2022), pathological relationships at work are associated with burnout as well as affective well-being. Since disordered relationships can be seen as social demands that deplete vital resources like emotional stability and social support, our own findings are consistent with the COR theory. Unless adequate compensatory resources are available, employees’ well-being declines when these resources are exhausted. To further examine the unexpected results, we conducted an additional subgroup analysis by gender. The main patterns of relationships remained stable—in both groups, a positive effect of corporate culture on the working atmosphere and, subsequently, the working atmosphere on psychological well-being was confirmed. Indirect effects (corporate culture → working atmosphere → psychological well-being) were significant and consistent for both men and women, which supports the robustness of the basic model. At the same time, interesting differences between genders emerged. For men, the direct effect of the occurrence of pathological relationships on well-being was not significant, while for women, a higher occurrence of pathological relationships was associated with higher well-being. Moreover, for women, the direct effect of corporate culture on psychological well-being became statistically significant at higher levels of leadership, suggesting a possible moderating role of leadership in gender-specific contexts. These differences suggest that the paradoxical positive effect of pathological relationships on well-being may not be universal, but rather may be linked with by gender and contextual factors, posing a challenge for further research.

A cross-sectional survey revealed that emotional rumination fully mediates the relationship between pathological relationships and employee well-being. In contrast, a daily diary study found that day-to-day fluctuations in pathological relationship events did not affect the ensuing morning mood. Thus, pathological relationships operate predominantly through affective ruminative thinking rather than a direct effect, particularly on persons with low neuroticism (Auweiler et al., 2023). The study by Koç and Keklik (2020) found that levels of pathological relationships, organisational commitment, and organisational cynicism strongly predicted psychological well-being among research workers. The research results clearly indicate that negative phenomena such as pathological relationships and conflicts in the corporate context significantly affect the psychological well-being of employees. However, the impact of these factors is not isolated or direct. The effects of these negative phenomena, such as pathological relationships, are mediated by multiple psychological mechanisms.

5.4. The Role of Leadership in Promoting Psychological Well-Being

The research results showed a positive association between leadership level and employees’ psychological well-being. This finding aligns with the JD-R model, as leadership is associated with key job resources such as feedback, recognition, and psychological safety, which are linked to lower demands and higher well-being. This trend suggests that, rather than directly magnifying cultural impacts, leadership functions as a proximal resource stabiliser inside the JD-R and COR frameworks, moderating daily changes in the work milieu and maintaining resource balance. The observed patterns of moderation may potentially be explained by other factors. Leadership’s apparent moderating influence in some pathways may have been diminished if it shared variation with other resource-related factors (such as organisational support, justice, or peer collaboration). In a similar vein, underutilised resources that are not included in the existing model could interact with management and the workplace culture to affect well-being. Employee autonomy, emotional intelligence, or resilience, for example, may make up for a lack of leadership resources. Future research should examine these possibilities since they imply that the interaction of job resources may be more intricate than our dual-moderated mediation model can represent.

The authors found that empowered leadership affected job engagement both directly and indirectly via psychological capital. Employee psychological capital completely moderated the connection between empowering leadership and psychological well-being, but only partially mediated the association between empowering leadership and job engagement (Park et al., 2017). Furthermore, psychological well-being moderated the impact of transformative leadership on Core self-evaluation. Furthermore, SSG positively mediated the direct effect of transformative leadership on PWB. Guanxi between supervisors and subordinates also favourably is related with the indirect effect of transformational leadership on core self-evaluation via psychological well-being (Ding et al., 2020).

As a multidimensional variable, the digital leadership-skill dimension has significant and positive effects on psychological well-being, whereas the digital leadership-attitude, competencies, and behaviour dimension has no significant effect (Dewi & Sjabadhymi, 2021). Added to the list of beneficial mental health impacts linked with transformative leadership and offer measures that firms might implement to increase employee well-being (Arnold et al., 2007). Abusive leadership has a considerable and detrimental impact on the psychological well-being of Saudi SME employees (Muthuswamy & Li, 2023). Els and Jacobs’ (2023) findings revealed that the proposed model was well-suited and demonstrated good linkages between authentic leadership and the relationships between emotional intelligence, cultural intelligence, and psychological well-being. Results of the mentioned research supported the positive impact of leadership on the psychological well-being of employees. Effective leadership is a protective factor against burnout, stress, and psychological difficulties.

Safety-specific transformational leadership enhances psychological well-being among healthcare professionals, with perceived risk of COVID-19 mediating this association (Irshad et al., 2021). The moderated hierarchical regression analysis results show that benevolent leadership has a strong beneficial link with employee psychological well-being. Furthermore, the positive association between benevolent leadership and well-being was stronger when psychological safety was high than when it was low (Erkutlu & Chafra, 2016). Mansouri et al. (2022) discovered that authentic leadership, meaningful work, and perceived organisational support all improved psychological well-being. The authors researched the impact of different types of leadership. Research reveals varying degrees of the impact of leadership on psychological well-being. The findings of our research support a general trend observed in the literature: leadership, regardless of its specific form, plays a crucial role in shaping employee psychological well-being.

Muhonen et al. (2013) found no direct relationship between leadership conduct (employee-centred and empowering leadership) and well-being in either the Swedish or Chinese samples. Furthermore, the study’s findings show that social climate mediates the association between leadership conduct and employee well-being; however, this appears to be culturally variable. The mediating effect is prevalent in cultures with a collective orientation and a high power distance. The authors examined the research issue in the same way we did, but they defined the relationships between the variables at a different level.

This trend suggests that, rather than directly magnifying cultural impacts, leadership functions as a proximal resource stabiliser inside the JD-R and COR frameworks, buffering daily changes in the work milieu and maintaining resource balance.

5.5. Interpretation of Rejected Hypotheses

Based on the results of the theoretical model of moderated mediation, six hypotheses were rejected: H2, H4–H6, and H8–H9. In the case of hypotheses H4 and H9, the dependence was confirmed, but with the opposite direction of the effect. It was found that the occurrence of pathological relationships in the workplace has a positive effect on the psychological well-being of employees, and the level of leadership has a moderating effect on the mediating relationship of work atmosphere between corporate culture and employees’ psychological well-being.

Hypothesis H2: The level of corporate culture has a positive impact on the level of psychological well-being of employees was rejected. This result aligns with the JD-R paradigm, which posits that employees’ perceptions of trust, support, and collaboration, considered distal job resources, are linked with company culture. These nearby resources are reflected in the workplace, which fosters psychological well-being. There may be several reasons for not confirming this hypothesis. One of them may be that corporate culture has an indirect effect on psychological well-being (the workplace atmosphere mediates this relationship). The relationship between culture and well-being is frequently mediated by the work environment and associated factors. Examples of how culture affects well-being include work–life balance (Stankevičienė et al., 2021; Agarwal et al., 2020), supportive supervision (Beauregard, 2011), and operational flexibility (Clark, 2001). Research indicates that although culture has a powerful direct impact, it also has a considerable indirect impact through mediators such as work–life balance, particularly in terms of improving happiness and reducing stress (Stankevičienė et al., 2021; Agarwal et al., 2020). Instead of vague claims about culture, employees are most impacted by tangible, everyday behaviours, particularly those of peers and leaders (Popa et al., 2023; Qin & Saufi, 2025). Another possibility is measurement limitations, as corporate culture was assessed with a single-item indicator, which may have reduced sensitivity. Although our total N was large, low variance on this measure could also have limited statistical power to detect small direct effects.

It is interesting to note that the findings contradicted our initial hypothesis in H4, which, based on the available information, predicted a negative correlation between employees’ psychological well-being and the presence of pathological workplace connections. According to COR theory, this paradoxical trend can represent cycles of resource gain, in which people subjected to more demanding situations (toxic encounters) may use coping mechanisms or resilience resources that temporarily improve their well-being. A statistically significant positive link (β = 0.4283, p < 0.05) was found in the analysis, suggesting that people who reported these relationships also reported feeling more content. This surprising discovery can be attributed to several possible causes. Another possibility is a suppressor effect, in which the direction of the coefficient may have changed due to the correlation between problematic associations and other variables in the model. Studies on regulated mediation have documented this pattern, in which overlapping conceptions can lead to suppression.

Initially, it is plausible that certain workers have acquired resistance or adaptive coping strategies that mitigate the adverse consequences of toxic relationships. The adaptable application of techniques (seeking social support, reassessing unpleasant events, and focusing on personal development) helps one regain self-worth and maintain control over behaviour and attitudes (Sarkar, 2025; Finstad et al., 2021; Parker et al., 2015).

Even in extremely stressful or toxic circumstances, greater resilience is linked to better coping, less psychological discomfort, and a decreased chance of negative consequences like burnout or suicidal thoughts (Leipold et al., 2019; Okechukwu et al., 2022; Murry et al., 2023). Problem-focused and meaning-focused coping strategies, which are associated with improved performance and well-being, are more frequently employed by resilient individuals (Parker et al., 2015). Over time, resilience can be strengthened via stressor exposure, self-reflection, and adaptive coping (Crane et al., 2019).

Conversely, in certain work cultures, interpersonal interactions that are confrontational or focused on pressure could be accepted and even linked to a performance-oriented setting.

Interpersonal conflict is more common in workplaces with higher job authority, particularly for men and younger employees. It suggests that situations that are competitive or hierarchical may normalise conflict while also increasing stress (Schieman & Reid, 2008).

Self-evaluation distortions or rationalisation systems, in which workers reframe bad circumstances as challenges rather than threats, could be another factor.

Employees can reframe negative occurrences as challenges, rather than threats, which helps them stay motivated and preserve their sense of self. This procedure is a component of larger self-assessment and rationalisation processes that promote mental health (Deffuant et al., 2024; Zabojnik, 2004). Contextual and individual modifiers of such surprising effects should be further explored in future research.

The H5 and H6 hypotheses were not verified. If the relationship between pathological relationships and the variables of interest is weak or the sample is small, the effect may not reach statistical significance. Some models show that if mediators influence each other’s effects, the indirect effect cannot be attributed to just one of them (Loh et al., 2022). The effect of pathological relationships may be specific to a certain population or context (e.g., age, culture, type of relationship), which can affect significance (Borairi et al., 2024). It is possible that pathological relationships in a given model do not actually have a significant moderating or indirect effect on the variables of interest, especially if other factors have a stronger influence (Vernimmen et al., 2025; Borairi et al., 2024). From the JD-R perspective, this pattern may indicate that negative social interactions at work act as long-term job demands rather than immediate moderators of well-being. Their effects may emerge over time as they gradually erode job resources, which could not be fully captured in the cross-sectional design of this study.

A straightforward binary variable of pathological relationships (presence vs. non-occurrence) was used to evaluate associations, which may have decreased the test’s sensitivity and variability. The COR hypothesis also explains this null finding: demands only become detrimental when resource loss surpasses specific thresholds. Due to the binary measurement, this threshold impact may not have been recorded in our data. The effect is heterogeneous, meaning that pathological relationships may only have an impact on certain people (such as those with lower resilience). The safeguards against adverse consequences are sensitive to cultural, contextual, and individual variations. What makes one individual or organisation resilient might not work for another (Ungar, 2013; Masten et al., 2021). Due to its inability to detect exposure gradients, this measuring option may have lower power. Additionally, threshold effects that may not be apparent in binary measurements are suggested by COR theory. Furthermore, because toxic relationships intersect with other psychosocial problems, suppressor dynamics cannot be completely ruled out.

One possible explanation for the rejection of hypothesis H8 is that workers may view leaders as independent individuals rather than as part of the organisation’s culture (Popa et al., 2023). This result challenges the JD-R assumption that leadership uniformly acts as a contextual job resource, suggesting that in some contexts, leadership is decoupled from culture and may therefore not strengthen the culture–well-being link. Multiple styles (transformational, authentic, and beneficent) have distinct consequences that are not captured by a single general indication.

For instance, the latter theory might have been disproven on the basis that the influence of culture can be stifled if informal leaders interfere in team dynamics (Raithel et al., 2021; Jamshed & Majeed, 2019). It is essential to recognise that cultural differences impact factors related to the workplace (Agarwal et al., 2020).

Overall, the pattern of results fits the dual logic of JD-R and COR: organisational culture and leadership act as sources of resource gain, while pathological relationships function as social demands that trigger or deplete those resources depending on context.

6. Conclusions

Based on the R-squared value, it is possible to state the quality of the model. We confirmed that the level of corporate culture has an impact on the working atmosphere; however, it does not have an impact on the psychological well-being of employees. The working atmosphere and leadership have an impact on the psychological well-being of employees. It is essential to note that the results in H4 contradicted our initial hypothesis, which predicted a negative relationship between employees’ psychological health and the presence of pathological workplace connections, given the available facts. The study’s results showed a statistically significant positive correlation, indicating that individuals who reported these ties also reported feeling more content. Numerous factors, including coping mechanisms, self-evaluation distortions, resistance to toxic relationships, and rationalisation systems, could account for this unexpected finding. We did not confirm the moderating association between leadership and the association between corporate culture and psychological well-being. Moreover, we refuted the moderating link of pathological relationships on the relationship between corporate culture and employees’ psychological well-being. The overall testing of the association between variables, pathological relationships, and leadership as moderators for the mediating relationship of work atmosphere between corporate culture and employees’ psychological well-being yielded the same result. This study is one of the first to test a dual-moderated mediation model in the workplace setting, which includes both leadership and pathological relationships. In terms of psychological well-being, the model shows significant predictive utility with an explained variance of 81%. Our results support the JD-R and COR frameworks, as well as the notion of transformational leadership, by demonstrating that with the right contextual or personal resources, perceived well-being can be preserved even in hazardous environments.

6.1. Theoretical Implications

This publication contributes to the research by examining the complex interplay between pathological relationships, company culture, leadership, and employee psychological well-being. It highlights the ongoing disagreement about the definitions of workplace bullying and mobbing, implying that greater clarification is required. We have covered these two different concepts under the term pathological relationships. The identified research gap is a lack of investigation into how organisational characteristics, such as corporate culture, work environment, and leadership, regulate the relation of pathological interactions (such as bullying) on employee well-being. This gap highlights the need for additional research into how leadership styles and organisational climate might reduce or amplify the harmful consequences of pathological relationships, as well as their overall relation on mental health across various organisational contexts.

By presenting a dual-moderated mediation model, the current study makes a significant theoretical contribution. It posits that the relationship between corporate culture and employee psychological well-being is mediated by the work environment and moderated by two different contextual factors: leadership quality and pathological workplace relationships. This integrated method captures how interpersonal and organisational dynamics can be associated with psychological results, building on the Job Demands-Resources (JD-R) and Conservation of Resources (COR) theories. To the best of our knowledge, neither the Slovak setting nor comparable sociocultural contexts have undergone empirical testing of this mediator/moderator combination.

6.2. Practical Implications

Data analysis is highly reliable, providing companies with genuine information to help them make informed decisions. The study demonstrates that company culture has a significant impact on the work climate, emphasising the importance of cultivating a positive corporate culture to enhance workplace conditions. The findings suggest that work climate, pathological connections, and leadership all have an impact on employees’ psychological well-being, indicating that organisations should prioritise these areas to enhance employee health and satisfaction. Leadership is crucial in striking a balance between corporate culture and employee well-being. Companies can use leadership development to increase staff performance. The study identified a moderated mediation model with high explanatory power (R² = 0.8067), offering firms a reliable framework to enhance employee well-being and improve the work environment. The study identifies major moderating effects, such as leadership’s impact on influencing workplace atmosphere, and provides actionable insights for focused management actions.

Based on the research results, the following steps can be taken to improve the work environment and enhance employees’ mental health. These suggestions fall into three main categories: Leadership Development, Organisational Culture Reform, and Workplace Climate Interventions and Employee Well-Being.

6.2.1. Leadership Development

To enhance the leadership of managers, companies can implement training programmes that focus on improving communication skills and recognising and addressing pathological relationships in the workplace. One option is to support transformational leadership, which is based on promoting open communication and fostering the psychological safety of employees.

6.2.2. Corporate Culture Reform

Corporate culture can be improved by introducing internal guidelines or a code of ethics to prevent toxic behaviour in the workplace, or by implementing programmes to strengthen positive relationships between employees. Companies can use mentoring or team-building activities.

6.2.3. Working Climate Interventions

Within the framework of the working atmosphere, it is possible to conduct assessments through anonymous surveys of satisfaction with the working atmosphere or to introduce or improve (if already in place) mechanisms for filing complaints and addressing corrections, or to make interventions more effective in identifying pathological relationships in the workplace. These measures will ensure the improvement of working conditions, as well as the psychological well-being of employees. Companies can provide employees with counselling services as part of psychological support or corporate wellness programmes for psychohygiene.

6.3. Limitations and Uniqueness of Research

The findings and study design have revealed several research limitations. These fall under the following categories: interpretive complexity, contextual and sampling limits, and methodological problems.

6.3.1. Methodological Limitations

The study’s methodology was limited to using single-item indicators rather than validated multi-item scales to measure several dimensions, including corporate culture, leadership, work climate, and psychological well-being.

This study’s primary drawback, though, is that single-item indicators were used to gauge psychological well-being, work atmosphere, leadership, and corporate culture. Single-item measures can be effective in reducing survey fatigue; however, they may not adequately represent the complexity of these constructs and may restrict the evaluation of construct validity and reliability. To provide more reliable and broadly applicable results, future research should use validated multi-item scales (such as Ryff’s Psychological Well-Being Scales, the Newman Work Environment Scale, and the Multifactor Leadership Questionnaire).

The operationalisation of pathological interactions is another methodological drawback. The dichotomisation of LWMS scores into presence and absence in this study may have decreased variability and facilitated the surprising positive correlation with psychological well-being. Future research should investigate treating pathological associations as a continuous variable to allow for a more sophisticated dose–response analysis and to avoid categorization bias.

The operationalisation of pathological relationships as a binary variable (presence vs. absence) as opposed to a continuous measure of frequency or intensity represents another constraint. Although this method may be less sensitive to minute variations in exposure levels, it is consistent with the study’s conceptual focus on the occurrence of hazardous interactions.

The operationalisation of pathological relationships may be one of the study’s limitations. Responses that indicated “never” were coded as absence of pathological relationships, while all higher frequency responses (e.g., “rarely,” “sometimes,” “often,” and “almost always”) were coded as presence. The variable was derived from the mean score of the LWMS factors, which were then dichotomised. The surprising positive correlation with psychological well-being may have been partially explained by the categorisation bias caused by this simplification. It is possible that binarization decreased the data’s variability, which would have made the estimate less accurate. A contradictory association, like this positive one, could have resulted from an exaggeration of the event (for example, marking “rarely” might not actually reflect a significant relation on psychological well-being).

The study primarily relies on self-reported data from surveys, which may introduce biases such as social desirability bias or recall bias. Respondents may have downplayed or misrepresented their experiences with pathological relationships and corporate culture, impacting the validity of the findings.

Another methodological limitation stems from the cross-sectional design of our survey. Although the hypothesised relationships are grounded in COR and JD-R theory, the present data only capture associations at a single point in time. As such, causal interpretations should be avoided. Future research should employ longitudinal or experimental designs to establish temporal precedence and causal mechanisms better.

6.3.2. Sample and Context Limitations

Despite efforts to ensure the percentage share of individual SK NACE categories of Slovak economic activities according to their real representation, the exact share was not achieved. However, we were unable to fulfil this factor, as over 1000 companies operating in the Slovak Republic were singled out.

The results are valid for the conditions of the Slovak Republic and countries with similar social, cultural, and economic development, such as those in the Visegrad 4 group. It may limit their applicability to regions with different cultural or organisational practices. However, it is questionable to what extent it is possible to unify the results for other countries with different developments. A longitudinal approach would help track changes over time and clarify causal pathways.

The research’s potential weaknesses include its lack of specificity regarding age cohorts. Different generations approach work in various ways, with distinct objectives and tendencies. In general, the younger generation Z seeks equitable compensation, employee benefits, good relationships, flexible work schedules, independence, and empathy.

Our conceptual decision to use the general term “pathological relationships” to encompass a range of unfavourable workplace occurrences (such as bullying, mobbing, harassment, and rudeness) is another drawback. This integrative technique may compromise theoretical precision by conflating constructs with different antecedents and outcomes, despite being consistent with the LWMS measurement and ensuring cross-cultural applicability.

6.3.3. Interpretational Complexity

Individual factors indeed have an independent relation on the psychological well-being of employees, in addition to corporate culture. The moderating relation of pathological relationships failed to change the direction of the relation of corporate culture on psychological well-being and changed the relation of the mediating relation of corporate culture through the working atmosphere on the psychological well-being of employees. This is an interesting finding, but it may be because the presence of pathological relationships in the workplace often is associated with the corporate atmosphere. This problem can be described as one of the possible limits of our research.

We consider the uniqueness of our research to be the demonstration of individual factors that impact the psychological well-being of employees, such as working atmosphere, pathological relationships, leadership, and the moderated effect of leadership on the mediating role of working atmosphere between corporate culture and employees’ psychological well-being. Many authors have addressed the issue of the psychological well-being of employees, but no one has examined the research problem from our point of view via moderated mediation analysis with the factors used. For a better understanding of pathological relationships, we used the scientific knowledge of the LWMS, consisting of five factors (criticism, ignoring, meaningless tasks, ridicule, and conflicts). Thanks to this, we were able to determine whether pathological relationships occur in the workplace based on the mentioned factors. It is an objective expression of the occurrence of pathological relationships, as the respondents are not directly aware that the questions refer to pathological relationships. Because of this, they can be more open and honest.

6.4. Future Research Directions

To enhance reliability and comparability across studies, future research should build upon these findings by utilising multi-item, psychometrically validated measures to assess important components. Furthermore, to detect non-linear effects and thresholds, pathological relationships should be described using latent variable approaches or treated as continuous variables. This method may clarify whether less frequent or severe negative interactions have a different association with well-being than more frequent or severe ones.

Future research directions could focus on examining the moderating effect of pathological attachments on the mediating effect of the work atmosphere between corporate culture and psychological well-being. While there is a well-established link between pathological relationships (e.g., bullying, mobbing) and psychological well-being, the role of leadership and the workplace environment as moderators of this relationship, particularly in terms of corporate culture, has not been thoroughly investigated. This gap highlights the need for additional research into how leadership styles and workplace atmosphere might either reduce or amplify the harmful impact of pathological relationships on well-being. By filling these gaps, future research may provide more comprehensive insights into how corporate culture, leadership, and pathological relationships interact to relation employee well-being, as well as how these elements can be managed to mitigate undesirable consequences, such as pathological relationships.

The relationship between national culture, organisational culture, and psychosocial risks (such as bullying) is understudied, particularly in terms of leadership’s involvement in influencing these environments. Extending the research to several cultural and organisational contexts would help validate the findings and provide evidence on whether the models proposed in the study can be universally applied or need to be adapted under specific conditions.

Future studies should employ longitudinal designs to capture temporal changes and better understand the causal links between company culture, leadership, and mental health. Incorporating approaches beyond self-report surveys, such as observational research, peer reviews, and physiological measures of well-being, may provide a more comprehensive picture of the phenomenon.

Investigating how digital leadership and workplace technologies affect the dynamics of corporate culture, unhealthy relationships, and employee well-being would help address current organisational difficulties. These steps will help us better understand and provide practical information for organisations trying to create healthier and more productive work environments.