The association between work-life integration, job satisfaction, job stress, and innovative work behavior

Maintaining work-life balance has become a pressing issue that people are increasingly concerned about and striving to achieve amid the rapidly changing socio-economic landscape. Driven by globalization, industrialization, modernization, and rising expectations for a better quality of life, employees devote more time and effort to work to earn higher incomes and advance their careers. However, this can lead to conflicts between work and family responsibilities (Liu et al., 2018; Kumar et al., 2023). According to Allen and Finkelstein (2014), individuals are always constrained by time and physical resources. If people spend too much time at work, the time available for personal and family life will be significantly reduced. It can increase stress, declining health, growing emotional distance among family members, and even family conflicts (Cho et al., 2018; Liu et al., 2018). Therefore, work-life integration (WLI) has emerged as an essential trend to maintain a balanced lifestyle. According to Kumar and Sarkar (2021), WLI refers to the harmonization of work and non-work roles reasonably and coherently. Individuals strive to attain satisfaction in both their professional and personal lives by integrating work and private life functions. Employees were more satisfied with their jobs and performed better when they had time for their families and met social needs while still fulfilling work duties (Bataineh, 2019).

However, WLI also presents several challenges. At the same time, family responsibilities, including child or elder care and interpersonal conflicts (between spouses or between parents and children), can cause mental and physical stress (Toffoletti and Starr, 2016; Kumar et al., 2023). Job stress (JST) tends to increase when work-life balance is not maintained (Tran, 2023), which reduces creativity and overall work performance (Bloom and Van Reenen, 2006; Wu et al., 2018), can negatively affect innovative work behavior (IWB). Besides, Lyu and Fan (2022) asserted that individuals experiencing work-family conflict (WFC) tend to have lower motivation to engage in innovative work activities. Accordingly, many recent studies have focused on analyzing how WLI influences JST and IWB. However, very few have explored the direct negative relationship between WLI and IWB among employees. Therefore, this research contributes both practically and theoretically. First, the identification of a direct negative association between WLI and IWB theoretically demonstrates that the connection between work-life dynamics and innovation extends beyond emotional strain. Second, by introducing job satisfaction (JSA) into a sequential mediation model, it provides explicit insights into the psychological continuum connecting adverse work-life conditions to degraded innovative behaviors. Through this, the study expands the generalizability of the theory regarding the relationship between WLI, JST, JSA, and IWB. To conduct these tests, the PLS-SEM and ANN approaches were adopted.

According to ten Brummelhuis and Bakker (2012), the W-HR model interprets the interaction between work and home domains based on the exchange of personal resources at time, energy, and mood. Resources gained in one domain can enhance functioning in the other, while resource loss in one domain can lead to strain in both. The model highlights how work–home enrichment or interference depends on this resource flow.

The JD-R model (Bakker and Demerouti, 2007) explains how job characteristics impact employee well-being and performance. It distinguishes between job demands, such as workload and time pressure, which increase stress and burnout, and job resources, such as support and autonomy, which promote motivation and engagement. The model outlines two key processes: the health impairment process, which links high job demands to strain, and the motivational process, in which job resources enhance performance.

Affective Events Theory, developed by Weiss and Cropanzano (1996), discusses the theory of the structure, causes, and consequences of emotional experiences in the workplace. AET proposes that workplace events trigger emotional reactions, which then influence attitudes (like JSA) and behaviors. It emphasizes that emotions are central to understanding how people experience their work and respond in the short and long term.

Work-life balance refers to an ideal, conflict-free state in which the emotional, behavioral, and time demands of work and personal roles are balanced simultaneously (Quinn and Hartland-Grant, 2026). In contrast, WLI represents a dynamic and transactional approach to harmonizing work-life needs, which is particularly crucial as boundaries increasingly blur in modern work paradigms (Morris and Madsen, 2007; Gao et al., 2025). Rather than strictly separating roles, WLI effectively synchronizes an individual's time and energy to fulfill professional, family, and social commitments (Kumar and Anooja, 2022). When blurring these boundaries, work-life conflict emerges as an overarching state of incompatible responsibilities between work demands and personal or family obligations (Skomorovsky et al., 2025). At the core of this state is work-life interference, a bidirectional inter-role conflict in which incompatible demands from work and personal life collide (Lourel et al., 2009; Handelzalts et al., 2024). This interference occurs when employees fail to balance these competing demands, as fulfilling obligations in one domain depletes the limited resources required for the other (Yan et al., 2022; Skomorovsky et al., 2025).

According to Kumar and Sarkar (2021), WLI describes how responsibilities are adjusted in the work and non-work domains, with technology enabling people to complete activities concurrently. It involves work-family interference, family-work interference, work-life enrichment, and work-life strain.

Kumar and Sarkar (2021) developed a WLI measure based on the work-family interference scale research of Allen and Finkelstein (2014). Interference refers to conflicts that flow in a particular direction (i.e. work to family, family to work), whereas work-family conflict is an umbrella term encompassing work-family conflict in general (Allen and Finkelstein, 2014).

Performing many roles at the same time can lead to interference, disputes, and overload in both the personal and professional spheres. For instance, bringing office work home might cause problems in personal life and vice versa (Galinsky and Kim, 2000; Desrochers et al., 2005). These conflicts and disruptions add up to work-family role strain (Duxbury et al., 1994). Additionally, when someone works from home regularly, their net time for family, friends, self, or society decreases and begins to interfere with their personal life. Similarly, excessive personal work interfering with office time diverts attention from work and causes stress due to decreased productivity (Kumar and Anooja, 2022).

Interference has the benefit of enhancing work-life enrichment (Julien et al., 2017). People feel satisfied when they can handle the responsibilities of both parties and have flexible alternatives for both domains. To fulfill their desires, millennials are eager to work in an environment that not only balances work and life but also integrates them harmoniously (Kumar and Sarkar, 2021).

A common definition of JST is a sense of difficulty, frustration, anguish, or tension related to one's employment (Grossi et al., 1996). According to the National Institute for Occupational Safety and Health (1999), JST is the negative aspect of physical and emotional reactions that arise when role (job) criteria do not align with employees' needs, resources, and skills.

The initial iteration of the JST scale was called the Occupational Stress Indicator (Williams and Cooper, 1997). It was created to assess the essential elements of work and the stress process in a broad range of businesses. Based on this, Wu et al. (2018) developed a construction workers' job stress scale with six elements. It is used to evaluate JST in relation to productivity, work-life balance, and remote working conditions (Wolor et al., 2023). It has also been used to investigate the relationship between occupational stress, work engagement, and perceived support from coworkers (Azila-Gbettor et al., 2022). The term job itself describes the stress that comes with a job because of things like a lot of work, little time, too much responsibility, and hazardous working circumstances. Family-work conflict (FC) is the term used to describe JST brought on by long-term employment or the strain of a poor income that prevents communication with the family. Career development (CD) is defined as occupational stress caused by insecure work. Organizational style (OS) is conceptualized as the stress that certain unprotected rights or illogical systems cause at work. Interpersonal relationships (IR(S)) are the source of stress at work that results from tense relationships with superiors or coworkers, or a leader's lack of understanding and support. Role management (RM) refers to the stress that workers endure at work due to competing job demands or lack of clarity on their responsibilities.

According to Rosdaniati and Muafi (2021), positive feelings regarding one's work that arise from an assessment of its attributes are known as JSA. It is a collection of an employee's sentiments regarding the positive and negative aspects of their work.

Many variables have been linked to JSA, including the system of rewards at work, the perceived level of supervision, the working environment, and personal characteristics like seniority, age, marital status, and years of experience that influence whether an individual is happy or unhappy in their position (Mosadeghrad et al., 2008). This is actually how much people enjoy or detest their profession (Spector, 1997). Employee satisfaction is a metric that discusses how employees feel about their jobs and the workplace in general. It gauges how they approach their work and how much the job is meeting their needs.

The process involving creating activities, methods, strategies for management, and organizational models that differ from the past is known as IWB (McGuirk et al., 2015). According to Birdi et al. (2014), IWB is the ability to generate fresh, truly unique ideas, investigate work outcomes that still have room for improvement, and incorporate these ideas into daily tasks. Additionally, IWB is a conscious attempt to create organizational procedures as well as to introduce and apply fresh concepts to each job, organization, or group tasks so that they can be carried out in accordance with each section's duty (Momeni et al., 2014).

De Jong and Den Hartog (2010) distinguished four dimensions of IWB and labeled them as idea exploration, idea generation, idea championing, and idea implementation. Idea exploration is the process of identifying an issue or an opportunity. Looking for methods to enhance existing goods, services, or procedures, or attempting to think about them in different ways, are examples of idea exploration. Idea generation means generating ideas for new goods, services, or procedures; expanding into new markets; enhancing existing workflows; or, more generally, finding answers to issues identified. Idea championing is crucial after an idea is generated, as it often does not align with existing practices within an organization. It involves individuals in informal roles pushing creative ideas beyond obstacles, finding support, building coalitions, and expressing enthusiasm, confidence, persistence, and involvement of the right people. Idea implementation also entails integrating innovations into routine behaviors and work processes, such as creating new work processes or products, testing them, and making necessary adjustments.

Based on that, Abstein et al. (2014) developed the IWB scale with four dimensions: problem recognition (PR), idea generation (IG), idea championing (IC), and idea realization (IR).

According to Kumar and Sarkar (2021), the benefits of WLI include remote work and the flexibility to handle important concerns. However, it also has drawbacks, including greater pressure from work, family obligations, and physical and mental stress. Besides, Lourel et al. (2009) believed that perceived stress and JSA are correlated with both work-to-home conflicts and home-to-work conflicts or interference separately. This outcome is in line with the findings of Kinnunen et al. (2006) regarding the detrimental effects of work-home interactions on stress. JSA was adversely correlated with negative work-to-home and home-to-work interference. Accordingly, hypotheses are predicted:

Since there seems to be a considerable correlation between job stress and satisfaction, researchers have recently concentrated on this relationship (Ismail et al., 2015). High levels of JST are associated with low levels of JSA, and occupational stressors are predictive of job unhappiness (Fairbrother and Warn, 2003). In terms of the relationship between stress and satisfaction, high levels of stress are typically associated with poor levels of JSA (Adamopoulos and Syrou, 2022). Therefore, the research hypothesizes:

Amabile and Kramer (2011) claimed that contented lecturers typically display higher levels of intrinsic motivation, which is frequently reflected in the creation of original concepts and cutting-edge teaching strategies. According to Hartner-Tiefenthaler et al. (2023), lecturers with higher work satisfaction are more willing to collaborate, share knowledge, and try out novel strategies, reflecting a significant relationship between work satisfaction and creative work behaviors. Moreover, intrinsic motivation for innovation may wane in a setting where extrinsic imbalances, like pay and recognition, dominate satisfaction, according to Tjoa and Arief (2022), which would result in fewer innovative ideas. Additionally, Dara and Saparuddin (2025) determined that for university lecturers, both the quantitative analysis and the qualitative findings demonstrated the positive relationship between IWB and work satisfaction. Thus, the hypothesis proposed is:

Workers experiencing conflict may feel that they lack the resources needed for creative tasks, thus avoiding innovative work practices to prevent further resource loss (Wang et al., 2022). WFC has also been linked to stress and reduced creativity (Luo, 2016), as well as greater sensitivity to negative environmental cues such as conflict, oppression, and restraint, which lowers participation in innovation (Lyu and Fan, 2022). In addition, employees may focus on resolving conflicts and contradictions, which reduces their work output and makes innovative behavior more difficult. As a result, WFC is negatively associated with knowledge workers' innovative behavior (Song et al., 2023):

According to Kumar and Sarkar (2021), to meet work demands, one may work longer hours. However, studies have shown that extended work hours can negatively effect on an employee's physical and mental well-being, reduce earning potential, hinder career advancement, and increase personal conflicts and neglect of work. Putting in more time and effort at work could raise income levels, but social life would suffer as a result. When one finds it difficult to balance both sides and requests more time, effort, and energy in another area, conflicts become more intense.

According to Hariri et al. (2024), a poor work-life balance is strongly related to higher levels of stress, burnout, and health problems. Besides, work-life conflict may obstruct and interfere with creative behavior (Kinnunen et al., 2010). Therefore, the research hypothesizes:

Stress can be linked to other people's actions at work, and if an employee does not have enough support from colleagues in their business endeavors, it frequently results in job discontent and less creative behavior (Hrnjic et al., 2018). Additionally, Anjum and Zhao (2022) found that stress at work can significantly affect healthcare workers' health and ability to innovate in the workplace. Therefore, hypotheses are predicted:

Accordingly, the research model is displayed in Figure 1:

The measurement scales in this study were adapted from previous research. A five-point Likert scale ranging from 1 “Strongly disagree” to 5 “Strongly agree” was used. The research employed a two-stage Hierarchical Component Model (HCM). In the first stage, Lower-Order Components (LOCs), the measurement was conducted using a reflective model for indicators and components. In the second stage, Higher-Order Components (HOCs) involved the latent variables, which included both Reflective Latent Variables (RLV) and Formative Latent Variables (FLV).

The latent variable WLI was developed based on the study by Kumar and Sarkar (2021). The latent variable IWB was adopted from the studies of Jong and Hartog (2010) and Abstein et al. (2014), reflecting employee behaviors at different stages of the innovation process in the workplace. The latent variable JST used the scale from Wu et al. (2018). The latent variable JSA was measured using the original scale developed by Hackman and Oldham (1975).

This study employs the PLS-SEM approach (SmartPLS software version 4) to examine the relationships among variables in the proposed research model. However, PLS-SEM might not completely capture the sophistication of human behavior (Wan et al., 2022). Thus, the study will incorporate ANN. This is to ensure the predictive accuracy of the findings, especially regarding the detection of non-linear associations (Lim et al., 2022). ANN technique helps to measure both linear and nonlinear associations simultaneously during the second-stage analysis, which can facilitate more in-depth insights and lead to improved decision-making (Sharma et al., 2024).

Sohaib et al. (2019) proposed that using both methods is an effective complementarity in the model.

This study surveyed employees from a variety of industries, sectors, and organizational types, with participants holding different job positions and responsibilities. Convenience sampling was applied, and data were collected through online (Google Forms) and paper-based questionnaires. The online survey was distributed through communities of working individuals on social media platforms, while paper questionnaires were administered directly to accessible employees. This approach enabled the study to reach respondents from diverse occupational backgrounds. The research uses the sample size calculation rule developed by Hair et al. (2021).

Following the survey and data screening process, 329 valid responses were retained, accounting for 80.44% of the total sample, while the remaining 19.56% were excluded due to missing data or identical responses across all observed variables, indicating invalid responses to reverse-coded items. The sample is considered appropriate overall. Firstly, women are both engaged in income-generating labor and primarily responsible for family care; thus, they face significant pressure in integrating work and family. On average, women spend more than twice as many hours as men doing housework and caring for their families (ILO, 2021; Tran, 2025). Secondly, respondents who have been working for five years or less and belong to Generation Z are under pressure regarding work-life imbalance. It partly contributes to the increasing burnout among Generation Z workers in Vietnam (Nguyen and Vu, 2025).

The demographic characteristics of the sample are summarized in Table 1:

To mitigate CMB, we used the Full Collinearity Test. In Supplementary Table A1, the Variance Inflation Factor (VIF) values calculated among the latent variables in the model are all below 3.3, indicating that the model is not affected by CMB (Hair et al., 2021).

As presented in Figure 2 and Supplementary Table A2, items of the LOCs demonstrated satisfactory reliability, with Cronbach's Alpha and Composite Reliability values surpassing the 0.7 benchmark (Hair et al., 2021). However, PR with Cronbach's alpha coefficient (α = 0.557) in the range of 0.5–0.7 is still acceptable in exploratory research (Nunnally, 1978).

Regarding convergent validity, the model met the expected criteria-each indicator had an outer loading greater than 0.7 on its corresponding latent construct, and the Average Variance Extracted (AVE) values for all constructs were above 0.5 (Hair et al., 2021). The study then evaluated the discriminant validity of the LOCs using two key approaches: The Fornell-Larcker criterion and the Heterotrait-Monotrait (HTMT) ratio. As shown in Supplementary Table A3, the square roots of the AVE values for all constructs exceeded their respective inter-construct correlations, satisfying the Fornell-Larcker criterion for discriminant validity (Fornell and Larcker, 1981). Furthermore, in Supplementary Table A4, all HTMT values remained below the recommended cutoff of 0.85, providing additional evidence to support the distinctiveness of the constructs (Hair et al., 2021).

For assessing RLV, Hair et al. (2021) recommended a minimum outer loading (λ) ≥ 0.70 for confirmatory research. But for exploratory research, an outer loading in the range of 0.4–0.7 is still acceptable to preserve the original latent indicators (Hair et al., 2021). Therefore, FWI (λ = 0.680) is retained.

Based on the evaluation thresholds of Hair et al. (2021), in Table 2, all outer loadings, Cronbach's alpha, average variance extracted (AVE), and the Fornell–Larcker criterion were accepted (Table 3). Besides, the HTMT coefficient between WLI and JSA concepts was 0.772, which is below 0.90. Therefore, these concepts were measured to have good discriminant properties and satisfy the discriminant validity criterion (Hair et al., 2021).

In Table 2, all VIFs <5 indicate that there is no high multicollinearity, ensuring that each component contributes separately, without excessive overlap. Therefore, the VIF index in the outer model satisfies the evaluation threshold of Hair et al. (2021).

Although the outer weights of IC, IR, and OS are not statistically significant (p > 0.05), their outer loadings exceed 0.50 (Hair et al., 2021), thereby confirming a strong absolute relationship with the FLV. The outer weight of IR(S) is not statistically significant, but its outer loading exceeds 0.4 (p < 0.001). However, IR(S) relates to workplace relationships, such as colleagues and leaders, which are important factors affecting the work pressure of an employee (Wu et al., 2018). Moreover, JST is an FLV. After considering the meaningfulness of IR(S) in contributing to this scale, the study did not remove it from the model. Together with acceptable VIFs and a solid theoretical rationale, therefore, IC, IR(S), IR, and OS were retained in the measurement model for FLV.

Following the evaluation of the measurement model, the study investigated potential multicollinearity among the latent variables by analyzing their VIF values. As indicated in Table 4, all exogenous variables in the inner model exhibit VIF below 3.3, indicating no serious multicollinearity among the predictors and ensuring stable estimation of the path coefficients (Hair et al., 2021).

Table 5 presents the values of the Coefficient of Determination (R²) and Predictive Relevance (Q²) for the endogenous constructs. The model explains 28.6% of the variance in IWB (R² = 0.286), indicating moderate explanatory power, with acceptable predictive relevance (Q² = 0.194). For JSA, the explained variance is substantial (R² = 0.565) and accompanied by strong predictive relevance (Q² = 0.458). Similarly, JST exhibits high explanatory power (R² = 0.616) and considerable predictive capability (Q² = 0.604). Overall, the model shows strong explanatory and predictive performance for JSA and JST, and moderate performance for IWB. (Henseler et al., 2016).

As shown in Table 5, the empirical findings provide support for all hypotheses. WLI exhibited a strong positive association with JST (β = 0.785), indicating that higher levels of interference between work and life are strongly associated with elevated JST levels. Conversely, WLI was negatively associated with JSA, suggesting that greater interference corresponds to JSA. Furthermore, JST was negatively associated with JSA. This finding underscores the negative relationship between elevated stress tolerance and employee satisfaction. In turn, JSA shows a positive association with IWB. In addition, WLI exhibits a small but significant direct correlation with IWB. Lastly, emphasizing that WLI indirectly undermines innovation through both satisfaction and stress pathways.

The mediation hypotheses were analyzed via bootstrapping methods using 5,000 resamplings in PLS-SEM (Hair et al., 2021). It was found that all indirect paths were statistically significant and distinct from zero since the 97.5 bias-corrected confidence intervals did not include zero. This lends support to H6a and H6b. Moreover, the Serial Indirect Path was also shown to be statistically significant (β = −0.151; p < 0.001), thus supporting H7. Overall, the study showed that JSA and JST indeed mediated the relationship between WLI and IWB (Preacher and Hayes, 2008; Hayes, 2018).

The robustness analysis demonstrates strong methodological convergence between the PLS-SEM and ANN approaches, as indicated by consistent statistical significance and identical importance orders for all proposed path relationships. A close correspondence between the PLS-SEM path coefficients and the ANN sensitivity measures supports the robustness and stability of the obtained results, indicating that they are not driven by functional form assumptions (Supplementary Table A5 and Supplementary Table A6).

First, the strong positive association between WLI and JST indicates that managing multiple roles concurrently may create considerable pressure for employees in daily work and life. Over time, this strain can reduce well-being (Kumar et al., 2023) and make it more difficult to sustain innovative behavior at work.

Second, in terms of partial mediation, the results indicate that WLI has a directly negative association with IWB, alongside its indirect paths. Unlike previous studies demonstrating the positive and negative aspects of WLI (Greenhaus and Allen, 2011; Kumar et al., 2023), this study identifies a direct negative linkage with IWB through the empirical findings. Excessive boundary blurring forces employees to switch roles and multitask repeatedly. This weakens the concentration needed to recognize complex problems and realize ideas, directly hindering IWB even before serious emotional stress appears.

Third, regarding the sequential mediation model, the results demonstrate a multi-stage continuum of psychological deterioration (Hayes, 2018). Through the W-HR, JD-R, and AET frameworks, excessive work-life interference acts as an environmental demand that depletes resources, significantly associated with higher JST. Stress acts as a negative emotional event, eroding JSA (Terry et al., 1993; Jaafar et al., 2021), which links to diminished IWB (Wu et al., 2024). Lacking psychological safety, dissatisfied employees adopt a risk-averse stance, withdrawing from generating new ideas.

Crucially, contextualizing WLI within cultural and gender norms explains the depth of this stress. While Western paradigms prioritize flexibility (Bryant, 2024), Eastern cultures typically normalize long work hours (Gao et al., 2025) and traditional gender roles (Berger Ploszaj et al., 2025), where mothers are culturally expected to prioritize childcare (Gao et al., 2025). Consequently, when blurring boundaries, it contradicts the ideal concept of work-life balance (Wessels and Lachman, 2020). Women, favoring integration for family proximity (Desrochers and Sargent, 2004), often face conflicting demands and increased workloads (Rau and Triemer, 2004), which are linked to severe occupational stress (Berger Ploszaj et al., 2025). Although the level of stress from remote work varies globally (Klarsfeld et al., 2024), it is particularly severe for women in traditional societies.

From a practical standpoint, organizations can consider management policies tailored to local cultural norms. Recognizing that traditional gender roles strongly influence priorities between work and family (Berger Ploszaj et al., 2025). It is supported by the fact that working parents constitute a highly committed segment of the workforce when properly supported (Gao et al., 2025). Furthermore, the decline in innovative behavior occurs sequentially through increased stress and decreased satisfaction. Hence, organizations should adopt a result-oriented culture focusing on work quality and outcomes while enhancing employee self-efficacy to manage flexible work arrangements effectively (Kumar et al., 2023).

This study makes several contributions to the academic literature and organizational practice. First, by establishing a direct negative association between WLI and IWB, it explores the “dark side” of integration. The disruption of bidirectional work–family role integration (FWI and WFI) within highly blended boundaries induces severe cognitive fragmentation and stress. Second, the sequential mediation framework positions JSA as a critical affective state necessary for the later stages of innovation. While cognitive bandwidth is required to generate new ideas, our findings demonstrate that positive job attitudes are mandatory for the idea execution phases. Employees whose satisfaction is eroded by stress lack the intrinsic enthusiasm to contribute new ideas to the organization. Therefore, they proactively withdraw from idea championing and do not dedicate the necessary effort to idea realization. Practically, the research translates its complex psychological findings into Eastern contexts where collectivism creates blurred boundaries. To sustain the complete innovation pipeline, a dual approach is required, minimizing work-life interference and proactively maintaining employees' affective satisfaction.

This study has several limitations that should be acknowledged. First, the convenience sampling and sample size may limit the generalizability of the findings. Second, the cross-sectional data does not provide a sufficient basis for making causal interpretations. Moreover, WLI is influenced by various contextual factors, such as the nature of the job, family circumstances, cultural environment, socio-economic factors, and individual resilience. These factors could shape how employees perceive stress arising from the integration of work and life domains. Based on that, researchers can explore factors that impact WLI on both positive and negative sides. Furthermore, a longitudinal design and experimental or quasi-experimental designs would provide the most rigorous testing for causality. Further research could also investigate potential mediating mechanisms, such as cognitive distraction and multitasking pressures from work-family demands. Expanding research in these directions would deepen the understanding of the dynamic relationship between work-life, stress, and innovation.

The supplementary material for this article can be found online