Organisational Support as a Mediator between Work–Life Balance and Job Performance in Universities

Introduction

In the contemporary era, the rapid advancement of digital technology has become a crucial factor influencing operational practices and personnel conduct across various sectors, including higher education (Tarafdar et al., 2019). Information and Communication Technology (ICT) has experienced significant development, with Cloud Computing providing greater flexibility in data storage and retrieval from any location (Armbrust et al., 2010). The availability of high-speed internet has facilitated enhanced communication efficiency and continuous information exchange (Cisco, 2020; Perrine et al., 2023). Online collaboration platforms such as Microsoft Teams, Zoom, and Google Workspace are widely used in higher education to support remote work and interdepartmental collaboration. These technological developments promote operational efficiency, adaptability, and transform work practices among university staff (Tarafdar et al., 2019). University personnel demonstrate increased flexibility in task execution through digital channels beyond traditional working hours (Nguyen et al., 2020; Chidambaranathan & Culathur, 2022), while Information Technology capabilities augment efficiency, reduce workload and costs, and foster interaction among staff, students, and departments (Molino et al., 2020; Nguyen et al., 2020; Pavithra et al., 2023).

However, the expectation of constant responsiveness or an always-on mentality blurs the boundaries between professional and personal life, thereby contributing to stress, fatigue, and technostress (Tarafdar et al., 2019; Canassa & Baldin, 2022). As a result, careful management of technology use is essential to maintain work effectiveness and quality of life. Work-Life Balance (WLB) significantly influences employee productivity and well-being (Staniec et al., 2023). University staff face challenges in balancing workload with familial responsibilities, particularly within dual-earner or extended family contexts (Srinivasan & Sulur Nachimuthu, 2021). Organizational policies and coping strategies, such as stress management, mindfulness, and emotional intelligence, aim to improve WLB, although the implementation of such policies may lag behind organizational needs

Work-Life Balance (WLB), which is intricately linked to productivity, job satisfaction, and organizational commitment (Johari et al., 2018; Wilhelmy et al., 2022; Staniec et al., 2023), is influenced by employees' perceptions of working conditions, autonomy, supervision, and organizational culture. Traditional statistical techniques may inadequately represent the complexity inherent within personnel groups. The combination of Machine Learning (ML) with conventional path analysis provides deeper insights into nonlinear relationships affecting WLB, particularly within the framework of digital transformation. Beyond coping strategies, organizational support is essential in cultivating work-life balance (WLB) and enhancing employee performance. Job performance comprises task performance—directly linked to primary responsibilities—and contextual performance, which includes voluntary behaviors that benefit the organization, such as assisting colleagues and demonstrating initiative (Borman & Motowidlo, 1997; Macrì et al., 2023). Organizational support—encompassing clear communication regarding WLB, flexible working hours, and recognition—serves to augment both aspects of performance. Family support also indirectly enhances performance by increasing work engagement, illustrating that social support promotes proactive behavior and commitment (Mubayrik et al., 2022; Pham, 2024). Such support alleviates technostress, fosters psychological safety, and motivates employees to exceed their role expectations.

Within academic institutions, supportive environments foster effective staff performance and promote collaboration, thereby enhancing organizational efficiency and employee well-being. Leadership and organizational dynamics further influence performance outcomes. For example, transformational leadership has been shown to affect both task-related and voluntary behaviors (Skeie & Klock, 2023; Ha & Hang, 2024), providing a framework for investigating similar mechanisms within the academic domain. Optimized work-life balance (WLB) allows employees to perceive increased support, which positively impacts both task performance and contextual performance. This study examines the relationship between WLB and the performance of university personnel, considering organizational support as a mediating factor that enhances well-being and encourages sustainable performance in the digital era.

Literature Review

Digital Technology and Evolving Work Patterns

The integration of digital technology profoundly influences contemporary society, impacting daily life through innovations such as cloud computing, smartphones, and the internet. These technologies facilitate collaboration from any location with internet access, supporting remote work and telecommuting (Bughin et al., 2018; Son & Lee, 2024). Consequently, distributed teams and virtual workspaces have transformed traditional workplace boundaries and geographic limitations (Manyika et al., 2016). Digital transformation, characterized by the integration of digital technologies into organizational activities, reshapes operations and the delivery of value to stakeholders. In higher education, this enables improved productivity and enhances the student experience, prompting universities to reevaluate strategies to remain competitive in the digital era. Digital transformation also redefines workforce roles and necessitates the acquisition of new skill sets. As organizations increasingly adopt digital tools and automation, the distinctions between functions become less clear, requiring both technical proficiency and interpersonal skills. Nonetheless, excessive dependence on or incorrect application of technology can lead to technostress. Kumar (2024) delineates five categories of technostress: Techno-Overload, arising from an abundance of e-mails, tasks, and notifications; Techno-Invasion, where the boundaries between professional and personal life are compromised; Techno-Complexity, involving challenges in mastering new technological tools; Techno-Insecurity, reflecting concerns about displacement by others or automation like AI; and Techno-Uncertainty, characterized by unease amidst rapid technological advancements. These issues are especially common in educational institutions, where digital technology raises expectations for performance and service (Arabia, 2023; Bourlakis et al., 2023).

Work-Life Balance – WLB

The adoption of digital work models has enhanced agility and efficiency; however, it also introduces challenges to employees' work-life balance (WLB). Persistent connectivity, unrestricted access to information, and continuous communication often lead to overlaps between professional and personal time, as observed by Farivar et al. (2023). Organizational support and flexible work arrangements are vital for fostering WLB, including suitable working hours and remote work options, as highlighted by Haider (2024). Such initiatives help reduce work-related stress, increase employee commitment, and ensure the long-term retention of skilled personnel. Integrating employee well-being into organizational policies ensures support not only for work responsibilities but also for mental, emotional, and physical health. Work-Family Conflict is pivotal in understanding how conflicting occupational and domestic demands can undermine well-being and performance. Organizations addressing Work-Family Conflict via flexible arrangements, supportive culture, and workload management enhance WLB and overall employee welfare. Modern work is no longer confined to traditional hours or office locations, requiring adaptation to remote and digital operations. When combined with supportive policies, technological tools improve efficiency, motivation, and performance (Farivar et al., 2023). Universities utilize information systems and management software to streamline functions, reduce workload, enable multi-location work, and improve task agility.

Organizational Support

Organizational support (OS) encompasses a commitment to employee well-being, including emotional, instrumental, and WLB-related support (García & Jaramillo, 2023). OS enhances employee–organization relationships, thereby fostering commitment and retention. Emotional support addresses psychological and emotional needs to reduce stress and increase engagement; instrumental support supplies resources and training to improve task efficiency; and WLB support alleviates stress and elevates job satisfaction. OS reinforces the relationship between WLB and positive outcomes. WLB influences job performance and satisfaction, effects that are intensified when employees perceive high levels of OS (Faisal et al., 2022; Medina-Garrido et al., 2023; Jegede, 2024). OS alleviates stress, enhances performance capacity, and fosters organizational commitment, particularly in environments characterized by high WLB. Consequently, OS serves as a vital moderating factor, enhancing the benefits of WLB for performance, satisfaction, and employee retention.

Performance

Organizational success depends not only on task execution but also on behaviors supporting a positive work environment. Job performance comprises task performance—core duties—and contextual performance, involving behaviors that enhance the organizational climate (Borman & Motowidlo, 1997; Nguyen et al., 2022; Sugimori et al., 2022). Examples include assisting colleagues, demonstrating initiative, and contributing to a constructive environment. Balancing work and personal life amidst technological demands is critical; failure to do so can impair performance. This study examines the relationship between WLB and job performance, with OS as a mediating factor. Maintaining a favorable work–life equilibrium can inform recommendations for university administrators to enhance personnel well-being and work efficiency, enabling sustainable adaptation in rapidly changing environments.

Materials and Methods

To facilitate a comprehensive multi-dimensional analysis, the researchers undertook processes including statistical path analysis and machine learning, as detailed below.

The sample population for this study comprised operational staff from various faculties at Mahasarakham University in Thailand. According to Hair et al. (2010), the recommended sample size for Structural Equation Modeling should be between ten and twenty times the number of observed variables. Since this study includes a total of eight observed variables, the appropriate sample size should range from eighty to one hundred sixty. Kline (2015) suggests that a sample size below two hundred may result in sampling errors. Using Yamane's formula to estimate the population, the calculated sample size is 242. Consequently, the researcher selected a sample size of two hundred forty-two units, which is deemed suitable for path analysis based on the methodology of Hair et al. (2010), to ensure an adequate sample and reliable results extrapolation. The researcher employed probability sampling, specifically multi-stage cluster sampling. Path analysis was used to evaluate both direct and indirect effects. The results were assessed using various fit indices (Table 1).

Table 1. Fit Index.

The examination of the quality inspection data for research instruments includes the calculation of Cronbach's alpha coefficient, denoted by the symbol α, as well as the analysis of moisture content, factor loadings, and the evaluation of extracted variance, represented by the symbol AVE. Additionally, the assessment of the composite reliability of the variables, indicated by the symbol CR, is conducted. The results of this analysis are documented in Table 2.

Table 2. Reliability and Convergent Validity of Work–Life Balance, Organizational Support, and Job Performance Constructs.

According to Table 2, the factor loadings derived from the confirmatory factor analysis exceeded 0.40. The Cronbach's alpha coefficient () for each variable (n = 242) exceeded 0.70. The Average Variance Extracted (AVE) was above 0.5, and the Composite Reliability (CR) of the variables was greater than 0.70. These findings suggest that each variable is appropriate as a component in accordance with the criteria established by Hair and colleagues (2012).

Machine-Learning Clustering of Employee Profiles

We conducted secondary data clustering to profile employees based on WLB and OS. Likert-type items prefixed WBL. and OS. served as features; the overall job performance index (JP) was reserved for external validation. Missing values were imputed using an Iterative Imputer, extreme values were winsorized at the 1st–99th percentiles, near-constant items (SD<0.10) were removed, and features were robustly scaled (median/IQR). A six-component PCA solution was extracted and used as the input space. We evaluated k-means, Gaussian mixture models, and k-medoids for K = 3–6 using the average silhouette score, the Gap statistic, and the BIC (GMM only). Model retention balanced quantitative fit with parsimony and interpretability. Stability was examined using 100 bootstrap resamples (80% subsamples), summarized by the median adjusted Rand index (ARI). Profile interpretation relied on cluster-wise standardized means; UMAP visualizations were used for 2-D inspection. External validity was assessed using the Kruskal–Wallis test on overall job performance (α = .05, two-tailed).

Results and Discussion

The analysis of the demographic characteristics of a sample comprising 242 individuals revealed that the majority were female (63%) and that the age range of 25 to 35 years accounted for 38%. A significant proportion held a bachelor's degree in education (82%) and possessed over seven years of work experience (53%). The median income ranged from 000 to 30,000 Baht, accounting for 63%, and the predominant occupational role was in general administrative positions, at 34%. WLB showed a direct positive effect on OS, whereas its direct effect on JP was not significant (DE=0.18, p=.148); OS, in turn, had a significant direct effect on JP (β=0.37, p<.001). However, the outcomes of the hypothesis testing indicated that WLB does not exert a statistically significant direct effect on JP (β₁ = 0.18, t = 1.4, p > 0.05) (Tables 3 and 4).

Table 3. Correlation Analysis and Discriminant Validity

Significance levels: *p < 0.05, **p < 0.01, ***p < 0.001.

Table 4. Model Fit Indices

Conversely, WLB exhibits a direct positive impact on OS. The hypothesis testing confirmed that WLB has a direct positive effect on OS (β = 0.782, t = 11, p < 0.05). Furthermore, OS demonstrates a direct positive influence on JP. The results of the hypothesis tests validate that OS has a direct and positive effect on JP (β = 0.37, t = 9.6, p < .001). OS functions as a mediating variable between WLB and JP. The hypothesis testing revealed that WLB does not have a direct effect on JP (DE = 0.18, p = 0.148). The perception of organizational support has a direct effect on job performance (DE = 0.37, p < 001), and the perception of balance between work and life positively influences the perception of organizational support (DE = 0.78, p = 003). The indirect effect of perceived organizational support on job performance is statistically significant (IDE = 0.28, p = 0.004), indicating that work-life balance influences job performance via organizational support. Ultimately, OS functions as a complete mediator between WLB and JP (Table 5).

Table 5. Hypothesis Testing and Mediation Analysis

Quantitatively, average silhouette values ranged from .284 to .307 across K=3–6, with only marginal improvements beyond K=4; the Gap statistic increased monotonically, and the elbow curve showed no distinct knee. Balancing quantitative fit with parsimony and interpretability, a four-cluster solution was retained. The solution exhibited high stability (median ARI = 0.956). A significant Kruskal–Wallis omnibus test for external validity supported overall job performance, H(3) = 37.49, p = 3.63 × 10⁻⁸ (Table 6). Profile inspection of standardized item means revealed four interpretable segments: a Thriving cluster (high WLB/OS), an At-risk cluster (low WLB/OS), a Constrained cluster (mildly below average overall), and a Supported-but-Stretched cluster (near-average overall with targeted OS strengths). UMAP projections showed compact, well-separated clusters, consistent with the stability evidence (Table 6).

Table 6. External validity for overall job performance across four clusters (Kruskal–Wallis)

The values reported in Table 6 are the Kruskal–Wallis H statistic with df = 3 and two-tailed α = 0.05; p is shown in scientific notation (here, 3.63 × 10⁻⁸). Effect size is epsilon-squared (ε²) computed as (H − K + 1)/(N − K), N = 242.  

Figure 1 displays cluster-wise means after winsorization (1st–99th percentiles), removal of near-constant items, and robust scaling (median/IQR); values are standardized (z = 0 at the sample mean). Positive (negative) cells indicate above-average (below-average) responses on each WLB/OS item. The pattern delineates four segments: Thriving, Supported but Stretched, Constrained, and At Risk.

The findings of this study confirmed the hypothesis in a manner that exceeded the researchers' initial expectations. Initially, the researchers hypothesized that OS would function as a mediating variable, exerting both direct and indirect effects on JP among personnel at the university, particularly in a context where work and personal life frequently intersect within a digital social system. However, the results of the structural analysis depicted a markedly different scenario, illustrating that organizational support functions as a full mediator, implying that the balance between WLB does not exert a direct effect on job performance but influences it exclusively through the perception of OS.

This discovery has substantial implications for understanding human resource management and organizational behavior. It emphasizes that cultivating a balance between employees' professional and personal lives alone is insufficient to enhance performance outcomes unless an institutional mechanism exists to facilitate translating this perception into motivation and commitment to fulfill duties. In other words, even if personnel demonstrate effective work-life balance management, without the organization fostering a truly supportive environment, this balance will not translate into improved performance results.

This finding is consistent with the Perceived OS theory articulated by Eisenberger et al. (2002), which posits that when employees perceive that their organization values and cares about their well-being, they respond with commitment, effort, and loyalty to foster positive organizational outcomes. Furthermore, this aligns with the research conducted by Lee et al. (2022), which identified POS as a significant psychological mechanism mediating the relationship between job stressors and work performance.

An in-depth interpretation further indicates that WLB is characterized solely as a latent potential that requires stimulation or enhancement through organizational support systems, including flexible policies, resource support, open communication channels between supervisors and staff, and a corporate culture that values human dignity. Accordingly, WLB should not be viewed as a destination but rather as a starting point that must be actively advanced through structural and symbolic support provided by the organization.

In theory, this discovery reinforces the Full Mediation Model, suggesting that the causal relationship between WLB and JP is entirely mediated, with OS serving as the pivotal mechanism that triggers changes in personnel behavior and attitudes. This not only elucidates the operational mechanism of WLB within a global context but also significantly contributes to academic understanding of the specific circumstances faced by universities in Thai society, where staff encounter challenges arising from societal expectations and organizational structural limitations. Beyond the structural findings, our clustering analysis provides complementary, practice-oriented evidence. A four-cluster solution offered the best balance of parsimony and interpretability and proved highly stable (median ARI = 0.956). The profiles—Thriving (high WLB/OS), Supported-but-Stretched (near-average with targeted OS strengths), Constrained (mildly below average), and At-risk (low WLB/OS)—map onto distinct combinations of resources and constraints. Importantly, between-cluster differences in overall job performance were statistically significant (H = 37.49, p = 3.63 × 10⁻⁸), reinforcing the proposition that perceived organizational support operates as the effective conduit through which WLB is translated into performance. These segmentation results suggest the need for differentiated managerial levers. For Thriving staff, maintaining autonomy and recognition may protect gains without additional cost. Supported-but-Stretched personnel appear to benefit from specific supervisory or informational supports; targeted coaching and workload calibration may convert localized strengths into generalized productivity. Constrained employees require low-intensity, broad-reach interventions (e.g., streamlined processes, transparent scheduling) to lift baseline constraints. The At-risk segment calls for a combined package—credible supervisory support, access to resources, and reconfiguration of role demands—before traditional WLB initiatives can exert measurable effects. Methodologically, the convergence between the structural model and the unsupervised profiles strengthens construct validity: OS emerges not only as a mediator in the path model but also as a principal dimension that differentiates naturally occurring employee groups. At the same time, we acknowledge the limitations typical of secondary, cross-sectional designs, including self-report measures, single-context data, and moderate absolute levels of separation characteristic of Likert-based behavioral clustering (Ahmad et al., 2023). Future work should employ longitudinal tracking to examine transitions across clusters (e.g., At-risk → Supported-but-Stretched → Thriving) as OS initiatives are rolled out, and incorporate multi-source job performance to mitigate common-method bias.

Conclusion

This study aims to examine the relationship between Work-Life Balance (WLB), Organizational Support (OS), and Job Performance (JP) among 242 university personnel. The sample primarily comprised females (63%), aged 25–35 years (38%), holding a bachelor's degree (82%), and having over seven years of professional experience (53%), with general administrative staff being the most common position (34%). Hypothesis testing indicated that WLB did not have a statistically significant direct effect on JP. Nevertheless, WLB positively influenced OS, which in turn significantly affected JP. OS functioned as a full mediator between WLB and JP, demonstrating that achieving WLB alone is insufficient to enhance performance without organizational support. These findings are consistent with the Perceived Organizational Support theory (Eisenberger et al., 2002) and with studies by Caesens and Stinglhamber (2014), indicating that OS functions as a psychological mechanism linking work pressures to performance outcomes. Work-Life Balance (WLB) is a latent potential that necessitates support through policies, resources, and organizational culture to be translated into measurable performance. Organizational Support (OS) enhances psychological safety, diminishes technostress, and motivates employees to surpass role expectations, thereby fostering effective task and contextual performance. The research identified four employee profiles based on WLB/OS characteristics, which reflect significant differences in performance. Administrators are advised to implement tiered OS strategies: maintaining conditions that promote thriving staff, providing targeted support for supported-but-stretched employees, eliminating systemic obstacles for constrained personnel, and delivering integrated support for at-risk employees. Monitoring shifts between these clusters can serve as an indicator of HR program performance. Quasi-experimental or stepped-wedge interventions may evaluate whether improvements in OS encourage personnel to transition toward higher-performing segments, ultimately converting WLB practices into measurable organizational outcomes.

In summary, OS serves as a vital mediator between WLB and JP. The study offers both theoretical insights and practical guidance for Human Resources management in higher education, emphasizing sustainable support systems that transform balanced work-life conditions into enhanced performance, engagement, and overall organizational effectiveness.

Limitations and Dimensions for Future Research

Limitations: This research primarily focuses on university contexts, which may not accurately reflect the realities of other organizational types. Therefore, future investigations should aim to evaluate and adapt this model across diverse contexts, such as private enterprises, government agencies, and multinational corporations, to discern how organizational support systems vary across settings and how these variations affect the WLB–OS–JP mechanism at different levels. This study employed a quantitative methodology, which enables the identification of statistically significant relationships; however, it lacks depth concerning the experiential dimensions of personnel. Consequently, subsequent research should consider adopting a mixed-methods approach, integrating qualitative techniques such as in-depth interviews or focus group discussions, to elucidate how personnel perceive organizational support and identify factors that influence the translation of these perceptions into actual workplace behaviors.

Future Research Directions: Although this study provides empirical evidence on organizational support (OS) as a mediator between WLB and JP, several theoretical and methodological aspects merit further exploration. The cross-sectional design limits causal inference; a longitudinal study could confirm whether WLB initially affects perceived OS, which in turn impacts JP, and reveal the progression of this relationship over time. While this research focuses on universities, future studies should adapt the model to diverse organizational contexts, including private firms, government agencies, and multinational corporations, to examine variations in OS systems and their influence on the WLB–OS–JP mechanism. A mixed-methods approach is recommended: quantitative analysis can validate statistical relationships, whereas qualitative methods—such as interviews or focus groups—can explore employees' experiences, perceptions of OS, and factors shaping how these perceptions translate into actual work performance.

Acknowledgments: We would like to express our sincere gratitude to Mahasarakham University and the Program in Business Administration and Digital Innovation, Mahasarakham Business School, Mahasarakham University, Thailand for their continuous support and facilitation throughout this research, particularly through the Mahasarakham University Research Grant, which made this study and its publication possible. Furthermore, we sincerely thank the respondents who participated in this study, whose cooperation, willingness to share their experiences, and active engagement were essential to the successful completion of this research.

Conflict of Interest: None

Financial Support: This research project was financially supported by Mahasarakham University.

Ethics Statement: This study was approved by the Research Ethics Committee of Mahasarakham University in accordance with institutional regulations.