Digital stress, employee performance, and mental well-being: examining the moderated mediation role of social media use

1 Introduction

Digital technologies are becoming a part of workplaces and pose opportunities and challenges for employees’ well-being and performance (Valtonen and Holopainen, 2025). Digital technology use may positively influence the efficiency, communication and productivity of employees; however, excessive use may negatively affect employees’ mental health (Supriyadi et al., 2025). In addition, the effect of digital technologies on mental well-being is becoming an important social problem now a days. Sometimes its overuse/misuse may cause such disorders as loneliness, anxiety and depression (Scott et al., 2017). Digital stress refers to stress triggered by a failure of an individual to cope with digital technology development and adjustment to changes in digital environment. Digital stress is one of the most important factors concerning the mental well-being of the employees in digital environment. Digital stress is thus the inability to cope with demands and expectations imposed by technology which can also result in stressful experiences and negatively affect mental well-being and employee performance (Brod, 1984; Tarafdar et al., 2007; Sellberg and Susi, 2014; Li et al., 2025). Individuals who are suffering from digital stress are highly likely to suffer from additional problems like headache, fatigue, and sleep difficulties. Digital stress negatively correlates with well-being, health, and work capacity (Reinecke et al., 2017).

Previous work shows that digital stress continues to grow as an employee's issue because work-related stress produces dissatisfaction among the employees along with burnout symptoms and performance declines (Riedl et al., 2023). The ‘Digital Stressors Scale's creation clearly indicates how firms must evaluate and treat stressors that arise as a result of workplace digital technology implementations (Fischer et al., 2021). Digital tools use by organizations requires understanding and mitigation of digital work-related stress due to their ongoing broad adoption.

The digitalisation of the workplace has led to greater exposure to digital requirements among employees in today's working life, thereby increasing digital stress. Given that digital stress can impact employee mental health and job performance, it is important that organisations consider the implications for their employees. Knowing how digital stress affects mental well-being and employee performance, workaholism can give them a competitive edge (Dospinescu and Dospinescu, 2020).

The resolution of this problem demands immediate attention because it supports staff welfare and enables both productivity stability and functional workplace atmosphere maintenance (Marsh et al., 2022).

In today's world, social media use (SMU) has been seen to create stress and help people to deal with the stress at the same time (Haug et al., 2024). Social media platforms as well as their design elements generate stress levels for users. The combination of continuous notifications along with the need to maintain idealized online personas and encounter unfriendly online content produces psychological stress for users (Oyinbo et al., 2024). Users on social media platforms spread their accomplishments and lifestyles publicly while peers end up making unfavorable social comparisons with each other (Derbaix et al., 2025). People who engage in such comparison behavior typically develop negative feelings about their self-worth through this process

Keles et al. (2023). suggests that when young people engage in unfavorable social comparison activities, they experience increased anxiety and depressed symptoms. The consistent exposure to negative content such as cyberbullying harassment along with the offensive discourse results in detrimental impacts on mental well-being (Park et al., 2024). Individuals in modern society make social media use platforms as important tools to relieve themselves of stress and enhance psychological efficiency. Social media platforms provide users with features to develop coping strategies through socioemotional support and information seeking and mental disengagement (Chaudhry et al., 2024). Also, social media platforms combine the benefits of connecting with friends and family and joining online communities so users receive emotional support along with the sense of belongingness. During the pandemic, people found comfort through SMU by connecting to supportive virtual communities that minimized their social isolation (Arslan et al., 2021). The ability to obtain both up-to-date and appropriate information stands vital in stress management. The previous studies show that people use social media as their primary source to obtain crisis information which enables them to control their levels of uncertainty and decrease stress (Abbas et al., 2021; Hatamleh, 2024, Nusrat et al., 2023).

Employee mental well-being impacts happiness, productivity, and resilience to stress and obstacles at work (Singh et al., 2021). Mental well-being in the workplace relates to an employee's ability to manage job demands, maintain emotional equilibrium, and perform well in a professional situation. But the digital age has blurred the line between personal and professional lives. Employees are frequently expected to be available after regular office hours, which leads to work-life problems and mental tiredness (Wang et al., 2021). Mental well-being can be understood as the state of psychological well-being of an individual's ability to cope with the daily life stress (Mphande, 2023). SDG-3 and SDG-8 prioritize employees’ overall mental well-being at work (Mariappanadar, 2024). These organizational goals are backed by digital stress management as it generates more productive work environments that enable the maintenance of calmness among employess and improves their work performances. Those organisations that design digital stress mitigation strategies would support the maintenance of their labour force and increase their economic resilience (World Health Organization, 2024). The employee performance term refers to an individual work achievement made due to the necessary contribution of effort associated with the aim of meaningful work, an involved attitude, and colleagues or boss assistance (Vuong and Nguyen, 2022). Employees who have all these qualities, skills, commitment, attitudes, and knowledge are considered assets for the organization (Nayem et al., 2024).

The fast-paced adoption of the digital technologies in the working environment has impacted communication, collaboration, and performance within the workplace setting (Jaiswal at al., 2024). Use of the social media tools is always perceived as a reason for the digital stress caused due to over-connectivity and overflow of information (Huusko and Sivunen, 2026). The dual SMU and its contribution in making the work environment better is limited in the existing literature. In this context, the current study proposed the potential role of the SMU for reducing the levels of digital stress and promoting well-being and performance of employees through JD-R Theory (Ayyagari et al., 2011). Nevertheless, the possibility of employing SMU as a tool for work to counteract the negative influence of digital stress on workplace performance has been overlooked so far.

The current study helps fill the research gap via demonstrating these relationships empirically through Structural Equation Modeling (SEM) and collected data among employees of digital workplaces (Ayyagari et al., 2011). The possibility of SMU as a way of work in order to counter the negative impacts of digital stress on the productivity of the workplace is understudied. Additionally, the research addresses Sustainable Development Goals (SDGs) in particular, SDG 3 and SDG 8, provides the ideas on how the workplace can be made efficient digitally for the benefit of the employees’ health and productivity (Mariappanadar, 2024).

With the rise of digital platforms, marketing strategies have evolved towards more interactive, consumer-oriented and technologically-driven methods. The advancements are expanding the knowledge of digital marketing ecosystems and the need to look at the role and impact of creative, unconventional marketing strategies for today's consumer behaviour (Gündüzyeli, 2025).

Despite the rise in the literature that explores digital stress, social media use, and employee performance, few gaps can be found. First, there is no doubt that the existing literature mainly concentrates on the negative impact caused by SMU and its role in creating digital stress. But the positive impact of SMU in coping with digital stress is rarely examined in the studies. Previously, digital stress, psychological well-being, and employee performance have been studied as isolated phenomena. The theoretical exploration of mental well-being as a mediator between digital stress and employee performance is at an early stage of infancy. Through the integration of all these factors into one framework, this study provides an insight on the effects of digital stress on employees as well as the potential SMU as a stress coping technique. Finally, most digital stress studies were conducted in non-office settings, among students or the general population, rather than in workplace settings. Such interactions in the workplace are rarely studied empirically.

While the studies on digital stress, SMU, mental well-being, and employee performance exist, there is limited research available on the interrelationships of all these constructs (Bodhi, 2025; Khetawat and Steele, 2023; Murari et al., 2024). The existing studies mainly focus on the negative impacts of digital stress and SMU; however, very few researchers pay any attention towards the importance of SMU for work-related reasons (Ma and Zhang, 2025; Treem and Leonardi, 2013). Moreover, limited research has examined SMU as a moderating variable and mental well-being as a mediating mechanism within an integrated framework, particularly in the context of the banking industry in developing economies. In order to fill the gaps highlighted above, the current research explores the relationship between digital stress, SMU, mental well-being, and employee performance among banking professionals in Uttarakhand, India.

The job demands-resources (JD-R) theory is undertaken to explain how resources and demands at work affect mental well-being. This study offers new perspectives on how SMU can be a significant resource at workplace for managing the mental well-being and performance of the employees. The current study aims to measure the effect of the SMU on reducing the digital stress and improving the employee performance.

Thus, the research questions are formulated-

RQ1: Does Digital stress affect mental well-being and employee performance in an organisation?

RQ2: Does SMU moderate the association of digital stress and mental well-being

RQ3: Does SMU moderate the association of digital stress and employee performance

2 Theoretical contribution

2.1 Job demands- resources (JD-R) model

The JD-R paradigm balances job demands (stressors) with job resources (support mechanisms) to explain how working conditions affect employees’ performance and well-being. Digital stress acts as a job demand, leading to burnout, fatigue, and cognitive overload (Shamsi et al., 2021). Nevertheless, there are job resources including use of social media which offer coping strategies that enhance the well-being of employees and stimulate creativity. Employees who socialize, network, and find support in social media use can easily alleviate digital stress to enable them to shift the cognitive attention to creative thought. These are some of the resources, being rated among the top and include social support, decision autonomy, and improved feedback, which help in the utilization of adaptive responses to high job demands. Considering the study, depression and task stress have a high causal connection with other social processes. According to this concept, job stress is enhanced by the increased working demands which are workload, time and cognitive resources. In addition, JD-R model assists in explaining how job stress facilitates the development of anxiety disorders. The model provides an explanation for how resources and demands at work affect psychological well-being. The paradigm is therefore useful in designing interventions to improve and maintain the psychological resources with an aim of reducing anxiety at the workplace.

2.2 Conservation of resources (COR) & social capital theory

From the theoretical perspective, human beings desire to acquire and protect resources such as time, energy, and mental strength (Hobfoll, 2017). In cases of high digital stress, employees also become drained of cognitive resources that impact adversely on their creativity. In order to sustain high levels of creativity, social media use, on the other hand, helps employees replenish these psychological resources through stress relief, collaborative learning, and emotional support. It aligns with studies revealing that people with large social networks tend to be resilient to workplace stress as well as produce more creative ideas. Numerous ideas that can be considered when researching the constructive impacts of SMU on mental well-being. The above-discussed theories represent an in-depth understanding of how the use of SMU may help employees overcome certain barriers. The Social capital theory (SCT), which expounds how social ties are helpful to individuals and groups. Close social networks also have the potential of granting more access to resources and opportunities than individuals with less and weaker social ties (Bourdieu, 1985).

The JD-R is the central theory in the current research, whereas the Conservation of Resources (COR) theory & SCT are employed as the secondary theories in order to examine employee reactions to the resource conservation in digital work environments. In this context, digital stress is defined as digital job demands which result from permanent technological exposure, information overload and growing demands for digital work. Mental well-being is defined as employees’ psychological resource state, which either preserves or exhausts their resources when dealing with digital demands. SMU is identified as a digital and social resource that strengthens employee ability to cope with digital demands. The end result of this demand-resource process is employee performance. The proposed framework provides insights into the interplay between organizational resources and individual digital capabilities and their impact on mental well-being and performance.

The research offers a novel perspective to the existing literature through the extension of the SCT in the realm of highly digitalized organizations (Kasim et al., 2022). SCT argues that social capital, which is the social networks and relationships, are valuable resources that provide access to information, emotional support, trust, and collaborative opportunities. Few studies have explored how SCT can explain negative consequences of digital stress at work in organizations. In addition to SCT, this research expands the notion of SMU as one type of digital social capital which offers information and socio-emotional capital to the employees. By applying SCT, this study will contribute by offering empirical support to the potential of using digital social capital obtained from SMU in the modern workplace.

2.3 Hypothesis development

2.3.1 Digital stress and mental well-being

The term “digital stress” describes the psychological strain that individuals experience from using digital devices excessively at work. Constant connectivity, information overload, and an inability to disconnect from work are the causes of it (Marsh et al., 2022). In earlier research, the phenomenon of digital stress was termed as “technostress” (Tarafdar et al., 2007). This situation occurs when an individual or an organization finds difficulty in adapting to the use of new technology leads to technostress. Due to the growing use of digital devices, stress is no longer only related to one's job or one's home life, whether it occurs at work or at home. HRM should gradually consider employees’ lives outside of work's traditional temporal and physical limits, i.e., their personal lives, while understanding the work in this digital age. Increased psychological stress, more intense negative emotions, and lower organizational commitment are all related to increased workload stress. Lower desired email loads are associated with lower workload stress, while higher desired email loads are associated with higher workload stress (Stich et al., 2019). Communication overload, fear of missing out, concern about earning acceptance, and availability stress are four potential causes of digital stress. Digital stress may be a major factor in regulating the association between unfavourable psychosocial consequences and digital media intake (Marciano et al., 2021). With the help of information and communication technologies, the employees are enabled to work at any time and from anywhere, and thus now experiencing technological stress (Bondanini et al., 2025). The fast rate of technological development, as well as COVID-19 lockdown restrictions, transformed the way of work and preconditioned the appearance of the so-called digital platform work (Arslan et al., 2022).

Effective personal functioning, fulfilling social relationships, and happy emotions are all components of mental well-being, which is a complex phenomenon (Guinto et al., 2025) However, technostress is a great risk for this perfect psychological balance as it rapidly undermines the individual's mental health by creating negative emotions and dissatisfaction, thus decreasing efficiency and a feeling of social connectedness (Wang et al., 2022).

H1: Digital stress negatively affects employee’s mental well-being

2.3.2 Digital stress and employee performance

The past years have seen the widespread use of digital tools, remote work, and ICT. The use of these technologies does help people work and live more smoothly, yet it sometimes leads to digital stress which refers to the psychological issues caused by too much or difficult use of electronic devices. Digital stress affects employees’ performance by influencing their productivity, how satisfied they are with work and their ability to be innovative. Techno-overload, techno-invasion, techno-complexity, techno-insecurity, and techno-uncertainty are the primary causes of stress that people experience when they are unable to handle information technology (Tarafdar et al., 2007). The JD-R theory from organizational behavior proposes that digital stressors at work can negatively affect people's well-being unless there are strong job resources such as help or freedom for them. Performance of employees is seen in their work output, as well as its quality, efficiency and effectiveness (Demerouti et al., 2001). Digital stress makes it harder for employees to perform well by leading to burnout, tired emotions and thinking too much. Yet, how well people use technology, their independence and the help they have can change this relationship. As businesses continue to go digital, they have to carefully introduce changes in ways that don't overburden their teams and keep their performance and well-being in check.

H2: Digital stress negatively affects employee’s performance

2.3.3 Mental well-being and employee performance

Organizations need their employees to perform well and mental well-being is becoming known as a significant factor in their performance. To be mentally well, a person must look after their emotional, psychological and social health because it affects their thinking, feeling and behavior at work. When workplaces get more demanding and complex, the mental state of employees wears down which affects their energy to perform, their decision skills and their output. People with mental well-being can deal with commonly experienced stresses, be productive in their job and contribute positively to their local communities (Ryff, 1989) It involves: Having positive emotions and a satisfactory outlook on life Healthy psychological state: being independent, gaining self-confidence and respecting one's self fulfillment from belonging to a group and taking part in society. SDG 3 focuses on two essential objectives involving mental well-being promotion while decreasing workplace conditions that create psychological health consequences for employees. Digital transformation brings employees three stress factors: technostress combined with digital overload while requiring continuous connectivity which results in cognitive fatigue and emotional exhaustion (Riedl et al., 2023). The employees’ mental well-being has a significant impact on their capacity to think creatively and solve problems (Jaiswal et al., 2024). When employees undergo extreme level of workplace stress, they tend to have a decreased cognitive flexibility, lack of intrinsic motivation, and lack of engagement in creative work (Wang et al., 2022).

Gajdzik et al. (2022) indicates that positive mental state also contributes to openness to new experiences, risk-taking and generation of ideas, and, therefore, is a substantial predictor of workplace creativity. When organizations invest in workplace well-being programs on social media-based collaboration tools employees perform more creatively. Employees need to be in good health so as to enhance good work relations and job satisfaction. The mental well-being is as important as the physical one is the factor of employee well-being. Encompassed within mental well-being are happy feelings, proper management of stress, and positive outlook (Malik et al., 2024). Mental well-being is interconnected to the performance of employees in digital workplaces. The positive mental state helps the employees to be more productive, creative, and motivated, whereas poor MW may lead to absenteeism, low efficiency, and high turnover rates (Selimović et al., 2021). Based on the above, the hypotheses are as follows:

H3: Mental well-being positively affects employee’s performance

H4: Mental well-being mediates the relationship of digital stress and employee performance

2.3.4 Moderating role of social media use (SMU)

In this research study, SMU is conceptualized as the platforms for work-related communication, knowledge sharing, professional networking, collaboration, and seeking or providing social support within the workplace. SMU is defined as employees’ intentional and task-related as a means of communication, professional networking, information sharing, knowledge transfer, and socio-emotional support in work-related settings. This conceptualization is grounded in SCT, which posits that social interactions facilitated through digital platforms can enhance information exchange, strengthen professional relationships, and foster employee well-being and performance. Accordingly, the construct captures intentional and productive engagement through social media use rather than passive browsing, entertainment-oriented use, or non-work-related activities. In accordance with the SCT, SMU can be described as a social information resource which provides the employees with valuable information and social resources via online interactions. The construct specifically refers to active and supportive social media interactions that help employees to be effective and healthy psychologically.

According to recent research, SMU affects employees’ mental well-being rather than their performance directly (Karatepe et al., 2021). The psychological resilience of employees, who are involved in professional social media networking, exchange of knowledge online, and emotional support interactions is higher, and such a result positively impacts their creative potential (Malik et al., 2024). The association which is supported by the COR Theory, which states that employee's psychological resources can be replenished by SMU, leading to a more creative thinking process (Hobfoll, 2017). In this manner, the SMU is an indirect cause of employee performance via its impact on well-being. Professional uses of social media are extensively spread in promoting collaboration, alleviating the work-related stress, and knowledge-sharing (Yue et al., 2024). SMU has served two significant purposes such as emotional support and professional network by employees. Workplace social media groups enable the employees to share the experiences they have, get assistance, and informally learn, which lessens the workplace anxiety (Pitafi et al., 2025). Excessive SMU can lead to work interruptions, social comparison stress, and digital exhaustion, it can also help employees manage stress (Haug et al., 2024). SMU by employees has a good effect on their productivity, and consequently a positive effect on employees’ mental well-being. Notably, being a successful employee completely mediates the relationship between employees’ SMU and their well-being (Zhu et al., 2024).

Employee mental well-being is affected positively by social networking. Blogs like LinkedIn also offer career growth, networking, and learning, which raises self-esteem and job satisfaction (Murari et al., 2024). Studies show that organizations that implement formal digital interaction standards combined with work-life balance plans to reduce staff stress levels and create better productivity performances fulfill the requirements of SDG 8 about decent working environments (Gálvez et al., 2020). Digital stress arises when employees experience technological overload, constant connectivity, and digital work pressure, leading to anxiety and fatigue. However, employees frequently use social media to reduce professional stress by seeking social support and informal learning (Sasidharan, 2022). Despite the detrimental effects of DS on productivity, some people utilize social media for emotional support, stress relief, and engagement at work (Supriyadi et al., 2025). The following hypothesis is put out in light of the discussion as follows:

H5a: SMU moderates the relationship between digital stress and mental well-being

H5b: SMU moderates the relationship between digital stress and employee performance

H6: SMU moderates the indirect relationship of digital stress and employee performance through mental well-being

Figure 1 depicts the research model.

3 Research methodology

This study has employed SEM to investigate the interaction of variables; DS, EP and MW. Both online and offline methods were used to collect the data. The survey questionnaire was divided into two sections. The first section described the constructs followed by the details of participants. The second section was prepared on the Likert scale statements for each construct. SmartPLS 4 was used to investigate both direct and indirect path hypotheses in order to perform a comprehensive evaluation of the designated research model (Hair et al., 2021).

The constructs were taken from previous studies: DS (Reinecke et al., 2017); SMU (Sun et al., 2023), the SMU in this study mainly focus on supportive and purposeful social media engagement; MW (Ng Fat et al., 2017); and EP (Yener et al., 2021), (Tarafdar et al., 2007; Ayyagari et al., 2011). The rating of items was done on a 1–5 Likert scale with clear survey instructions provided to all participants and their confidentiality was also assured. DS was reverse coded so that higher scores indicated lower digital stress.

3.1 Sample and data collection procedure

Data were collected using non-probability purposive sampling technique from the banking professionals working in both private and public sector banks of Uttarakhand, India. Banking employees were chosen as they use digital technologies, online communication platforms, and technology-enabled service delivery systems extensively. The data was gathered for nearly three months and the format of the data collection tool was structured questionnaire in both online and offline forms. Survey links were sent via professional networks, institutional contacts and managers of various branches, and printed questionnaires were sent and filled out by employees at selected bank branches. All participants were voluntary and confidentiality and anonymity were guaranteed. 450 questionnaires were distributed and 412 questionnaires were returned which were 91.56% of the total distributed. Data screening led to the removal of 26 questionnaires because of high levels of missing data, straight-lining and incomplete responses. Thus, 386 valid responses were considered for the final analysis. The data set was checked for missing data as well as duplicate data, outliers, and inconsistencies in responses before analysis. To ensure data quality, cases with missing data in excess of 40%, and cases with consistent answers were removed. The final dataset met the assumption for PLS-SEM analysis. The state was selected to represent the population density, literacy level, and use of mobile in the region relying on the current statistics (eUttaranchal.com, 2023). Majority of the participants (41.4) were men with women taking 58.6. According to participant age, the most important category was 31–35 years (33.3%), followed by 36–40 years (31.3%), 41–45 years (15.2%), 26–30 years (11.1%), 21–25 years (7.1%), 45 + years (1.0%), and 50–55 years. In terms of education level, the majority of respondents (82.8%) held a post graduate degree, followed by a graduate degree (13.1%) and a doctorate (4.0%). The scales with their items are shown in Appendix.

4 Results

The results from the application of SMART PLS 4.0 are presented in the following sections. The measurement model results are summarised as follows-

4.1 Measurement model results

Table 1 displays the reliability and validity statistics for the latent constructs assessed using PLS-SEM. Cronbach's alpha, average variance extracted (AVE), and composite reliability are the measures used. Each construct's Cronbach alpha value ranges from 0.768 to 0.923, suggesting adequate internal consistency dependability; values greater than 0.70 are regarded as excellent (Sarstedt et al., 2021). This validity will be attained when all items significantly load onto their designated latent variables (Hair et al., 2012).

Table 1. Composite reliability and convergent validity.

DS, digital stress; EP, employee performance; MW, mental well-being; SMU,social media use

The standardized outer loadings were considered to measure the dependability of the indicator (Hair et al., 2021). All indicators showed satisfactory indicator reliability, surpassing the suggested level of 0.70. As a result, the final measurement model retained all measurement items.

Since data for all constructs were collected from the same respondents using a single survey instrument, common method bias (CMB) was assessed. First, Harman's single-factor test was performed, and the results indicated that the first unrotated factor accounted for less than 50% of the total variance, suggesting that CMB is not a significant concern. Additionally, following Kock (2015), a full collinearity assessment was conducted. The variance inflation factor (VIF) values for all latent constructs were found to be below the threshold value of 5, further confirming the absence of substantial CMB in the dataset.

All measurement items and standardized outer loadings for all retained measures are listed in Appendix. All items of the questionnaires (DS1–DS3, SMU1–SMU3, MWB1–MWB3, and EP1–EP9) had an outer loading value greater than the recommended 0.70, so all were kept in the final measurement model. As a result, no items were removed during the measurement model evaluation. Indicator retention followed the recommendations (Hair et al., 2021).

All the loadings of the outer model as shown in Table 2 were strong on their constructs with values ranging from 0.766 to 0.898 that are higher than the recommended value of 0.70, so indicator reliability is assumed and all indicators are retained in the final PLS-SEM model.

Table 2. Outer loadings.

Table 3 depicts the outer VIF values, all values are below 5.0, indicating no problematic indicator-level multicollinearity.

Table 3. Outer VIF values.

4.2 Discriminant validity

Discriminant validity results are displayed in Table 4, where the most often employed criterion in PLS-SEM is the Heterotrait-Monotrait (HTMT) correlation ratio. A construct's discriminant validity establishes if it is distinct from the other elements of the model. Since all construct pairs in this table have HTMT values below the conservative criterion of 0.85, each construct is empirically unique (Henseler et al., 2015). This signifies the robustness of the measurement model and attainment of the discriminant validity.

Table 4. Discriminant validity (HTMT ratio).

Table 5 shows the findings of the discriminant validity test using the Fornell and Larcker (1981) criteria. Every latent construct in the model is ensured to be different from every other construct because to discriminant validity. Each construct is experimentally different from the others because the diagonal values are greater than the corresponding off-diagonal correlations.

Table 5. Discriminant validity (Fornell-Larker criterion).

The Fornell-Larcker criterion was mostly satisfied because the square root of the AVE for most constructs was higher than the inter-construct correlations.

4.3 Structural model results

The bootstrapping results support the proposed PLS-SEM model. All measurement indicators are significant as well as the main structural paths, and the model accounts for a large variance in both MW and EP. The findings show that employees’ performance and mental well-being are significantly impacted by digital stress. EP and DS play a major mediating impact on MW. Additionally, SMU plays a significant moderating role, especially when it comes to the relationship between DS and MW. The bootstrapping results support the proposed mediation model and validate the statistical significance of the measurement and structural models, with all outer loadings being statistically significant at p < 0.001 and the relationship between MW and EP and between DS and EP being statistically significant. The R² values are used to determine the amount of variance accounted for by the models; the variance accounted for by the model is 57.5% in the case of MW and 67.7% in the case of EP. In addition, the obtained interaction effects are statistically significant indicating the presence of mediation and moderation effects in the proposed model. The structural model is shown in Figure 2 and Path coefficients in Table 6.

Table 6. Path coefficients.

DS, digital stress; EP, employee performance; MW, mental well-being; SMU, social media use

Table 6 presents the path coefficients obtained. The results of the structural model show that DS has a significant effect on EP (β = 0.180, t = 4.153, p < 0.001). Since the DS construct was reverse coded, the positive coefficient indicates that lower levels of DS contribute positively to EP. Similarly, DS exerts a significant positive effect on MW (β = 0.492, t = 13.922, p < 0.001). Given the reverse coding of DS, this finding implies that lower digital stress significantly improves employees’ mental well-being. The results further indicate that MW has a significant positive impact on EP (β = 0.536, t = 12.014, p < 0.001). The interaction effect between SMU and DS on MW was found to be positive and statistically significant (β = 0.375, t = 9.716, p < 0.001). However, the interaction effect between SMU and DS on EP was statistically insignificant (β = 0.054, t = 1.645, p = 0.100). Finally, the interaction effect between SMU and MW on EP was also found to be insignificant (β = 0.033, t = 0.933, p = 0.351).

As depicted in Table 7, Conditional effects were examined at low (- 1 SD), mean and high (+ 1 SD) levels of SMU. The results reveal that as SMU increases, the association between DS and MW grows stronger. The MW scores were positively and significantly related with DS at low levels of SM [β = 0.176, 95% CI (0.085, 0.270)]. The effect was significantly larger, however, at the mean level of SMU [β = 0.492, 95% CI (0.420, 0.559)]. Furthermore, at high levels of SMU (+1 SD), the relationship became strongest [β = 0.808, 95% CI (0.718, 0.887)]. The results suggest that the positive effect of higher DS on employee's MW is significantly mediated by the moderator (SMU). Likewise, the direct effect of DS on EP increased from level of SMU. The effect was significant but relatively weak at low levels of SMU [β = 0.125, 95% CI (0.050, 0.203)], moderate at mean levels [β = 0.180, 95% CI (0.094, 0.265)], and strongest at high levels of SMU [β = 0.234, 95% CI (0.106, 0.367)]. The interaction effect was not significant in the structural model but the conditional effects indicate that there is a positive DS reduction-performance relationship for high social media engagement levels. The effect of MW on EP was significant and positive at all levels of SMU. Specifically, the relationship was significant at low SMU [β = 0.503, 95% CI (0.393, 0.607)], mean SMU [β = 0.536, 95% CI (0.445, 0.619)], and high SMU [β = 0.569, 95% CI (0.453, 0.678)]. The relatively stable coefficients indicate that MW is a robust predictor of MW regardless of the level of SMU.

Table 7. Conditional effects.

The R² values obtained showed in Table 8, that the proposed model accounted for 67.7% of the variance in EP variable (R² = 0.677; Adjusted R² = 0.672) and 57.5% in MW variable (R² = 0.575; Adjusted R² = 0.571). Such values suggest a high level of predictive and explanatory power of the model, which explains employee outcomes in workplaces that rely heavily on digital technology.

Table 8. R square.

The f² results showed in Table 9, indicates that the direct effect of DS on MW is large, and the direct effect of MW on EP is also large, which shows that the mediating effect of MW is very important, and the direct effect of DS and SMU on EP is small, the interaction effect between SMU and DS has a small effect on the mediation of MW to EP, and the interaction effect between SMU and MW has a negligible effect on the mediation of DS to EP.

Table 9. Effect size.

The predictive capability of the proposed model was assessed using the PLSpredict procedure in SmartPLS. The results in Table 10 reveal that all Q²predict values for the indicators of MW and EP are greater than zero, ranging from 0.290 to 0.417. This indicates that the model possesses predictive relevance for the endogenous constructs. Furthermore, a comparison of prediction errors revealed that the PLS-SEM model generated lower RMSE values than the benchmark linear model (LM) for all indicators. Specifically, the PLS-SEM RMSE values were consistently lower than the LM RMSE values across all MWB and EP indicators, suggesting that the proposed model exhibits high predictive power. Therefore, the findings provide robust evidence that the model has substantial out-of-sample predictive capability in explaining employees’ mental well-being and performance in digitally intensive work environments.

Table 10. Predictive relevance of the structural model.

5 Discussion

Digital transformation within the organisation has significantly changed the manner in which how the employees engage with the technology, handle the work demands, and remain mentally and professionally efficient. This study analyzes the impact of DS on EP and MW, and a significant moderating factor, SMU. The results are valuable to the research on DS within modern working settings, and they are compatible, expanding, and challenging the current literature, respectively.

The results show that DS has a significant effect on MW (β = 0.492, t = 13.922, p < 0.001). Higher scores indicate lower levels of perceived DS as the items were reverse-coded. Thus, the findings indicate that a lower DS is associated with higher MW among those working in digitally intensive jobs. Similarly, EP is greatly impacted by DS. (β = 0.180, t = 4.153, p < 0.001). However, the positive correlation is unexpected for the DS construct, as it is reverse-coded, indicating that reduced DS is associated with enhanced EP. When the digital needs are seen to be beyond what one can cope with it will cause distress, frustration and poor mental state. This framework is supported by the results of the research indicating that the employees who view DS as unmanageable are more psychologically affected. Therefore, H1 is supported.

A large number of cognitive, emotional, and motivational factors affect employee performance. Employees who undergo high rates of DS have a higher chance of developing some concentration difficulties, mental exhaustion and lack of motivation. All these are contributing to poor performance (Ragu-Nathan et al., 2008; Tarafdar et al., 2015). The frequent disruptions and multitasking requirements in digital environments, as well as constant monitoring and tracking can disrupt the ability to perform specific duties and make decisions, which can be stressful. Thus, H2 is supported, and with the aim of minimising DS, employees will have the opportunity to work more effectively and efficiently.

The JD-R theory provides a comprehensive model on which this finding can be substantiated. This theory suggests that strain and burnout follow job demands exceeding the available resources (including technical support, autonomy, or digital competence), ultimately deteriorating job performance (Demerouti and Bakker, 2001). The findings indicate that the lack of resources to handle the digital needs results in worse EP, which confirms H2. This is also in line with the study that indicates that employees affected by DS tend to complain of decreased efficiency, missed deadlines, and less innovativeness (Ayyagari et al., 2011). Also, the same study concluded that technostress disrupts work engagement and has a negative impact on the quality of work performance, particularly in knowledge-intensive work (Day et al., 2012). It is on the basis of the correlation between MW and EP which is strong and statistically relevant to the current studies where psychological health correlates with improved job performance. Employees with strong MW would be better motivated, engaged, and focused hence are more likely to perform better in the tasks (Wright and Cropanzano, 2000).

MW is one of the factors that lead to positive affect, less mental interference, long-lasting energy levels, which are essential during challenging or demanding work. On the other hand, employees with DS and mental burden are prone to cognitive fatigue, low self-efficacy as well as absenteeism (Hakanen et al., 2006). The results also indicate that MW has a positive effect on EP (β = 0.536, t = 12.014, p < 0.001). Therefore, H3 is supported. This finding indicates that employees that have strong MW are more likely to demonstrate better job performance, productivity, and effectiveness. These impacts can minimize the chances of producing homogenous and quality work and elevate the chances of mistakes and lack of involvement therefore contributing to H3.

For mediation analysis, both the direct path from DS to MW as well as the direct path from MW to EP are statistically significant, suggesting MW as a mediating variable between reduced DS and EP. As the bootstrapped indirect effect is statistically significant, the hypothesis H4 is accepted, which also suggests MW is a mediator between DS and EP. Since both the direct effect (DS → EP) is significant, and the indirect effect (DS → MW → EP) is significant, the relationship represents partial mediation.

The moderation analysis showed SMU to be a significant moderator between the relationship of DS and MW (β = 0.316, t = 11.613, p < 0.001), which supports H5a. The positive interaction coefficient indicates that the positive relationship between SMU and MW outcomes of employees is reinforced with increased SMU. The moderating effect of SMU on the relationship between DS and EP, however, was not significant (β = 0.054, t = 1.645, p = 0.100). Therefore, the direct effect of DS on EP, with H5b showing that SMU did not have a significant influence.

5.1 Moderated mediation analysis

To test for conditional indirect effect as theorized in H6, a moderated mediation model is analysed using the bootstrapping method in SmartPLS with 5,000 resamples. The results reveals that the relationship between the indirect effect of DS on EP with MW differed among levels of SMU. Specifically, the conditional indirect effect was smaller at low levels of SMU (β = 0.089), moderate at mean levels of SMU (β = 0.264), and larger at high levels of SMU (β = 0.460). These conditional indirect effects were all statistically significant as the bootstrapped confidence intervals did not overlap with zero. Moreover, the interaction effect of SMU on the relationship between DS and MW (β = 0.375, t = 9.716, p < 0.001) suggests that reduced DS has a greater positive impact on MW when combined with SMU. That is, the higher the employees’ SMU, the greater the psychological benefits they gain from less DS, which in turn leads to better job performance. Findings support the moderated mediation mechanism, implying that the indirect pathway between DS and EP via MW is dependent on the level of SMU. So, the positive outcome of lower digital strain on employee productivity, due to better mental health, is amplified if employees are more active in meaningful, positive SMU.

These results indicate that those who are more active in their SMU experience more psychological benefits from the DS, which then relate to job performance. The results therefore support H6.

The moderating effect of SMU on DS and MW is shown in Figure 3. Results show the positive correlation between DS and MW strengthens as the amount of SMU increases. Specifically, the negative relationship between reduced DS and MW is strongest for those with high SMU and weakest for those with low SMU. These findings suggest that the negative relationship of reduced DS with employees’ MW is strengthened by the presence of SMU thus supporting H5a.

The moderating effect of SMU between DS and EP is shown in Figure 4. The interaction effect was statistically insignificant (β = 0.054, p = 0.100) although the plot shows that the more that employees scored on SMU, the more that they scored on EP at different levels of DS. Thus, SMU does not significantly moderate the direct link between DS and EP; hence H5b is rejected.

The interaction effect between SMU X DS and EP is provided in Figure 5. Higher levels of MW are associated with better EP at every level of SMU, as suggested by the positive slopes. But the almost parallel interaction lines indicate that the interaction between MW and EP is relatively constant with different levels of SMU. This finding aligns with the insignificant interaction effect (β = 0.033; p = 0.351), suggesting that the interaction between MW and EP is not significant in the presence of SMU.

6 Implications

6.1 Theoretical implications

The research contributes to the current literature. First, this study extends JD-R Theory by re-contextualizing digital stress as a contemporary, technology-based work demand and introducing the SMU as an unconventional but effective work resource. Contrary to the conventional notion of DS being associated with negative outcomes, the results indicate that DS can be managed if employees are given resources to support them in their work, including through SMU in a purposeful and supportive way. This complementary perspective contributes to the theoretical frameworks that have so far not considered the complex nature of digital tools, characterizing them as being positive or negative. The interaction between requirements and assets in our studies suggests a more dynamic relationship and the challenge to academics to re-evaluate the effect of the digital space on human behavior in business.

Second, the study demonstrates how employess proactively use the social media platforms to conserve and replenish psychological resources such as emotional support, cognitive clarity, and belonging. This study employs SCT for assessing the potential of digital social capital created through SMU to be a key resource for meeting the digital demands in the workplace and supporting employee well-being in today's working environments. The results not only contradict the prevailing perception of social media as a distraction but also suggest that purposeful and supportive SMU can serve as a coping mechanism that helps sustain employees’ mental well-being in digitally driven workplaces.

Furthermore, this study enhances the theoretical arguments about the influence of technology on stress by empirically validating the mediating role of MW between DS and EP which has not been well studied in the literature of organizational behavior. Although the existing body of knowledge has been more inclined towards perceiving DS, MW, and EP as independent entities, our unified model is more integrated, demonstrating the interactions between these aspects in complex, yet meaningful, manners in the context of digital spaces. This has a consequence in updating and reinterpreting the traditional stress-performance models in the digitally evolving contexts. Moreover, the study also adds conceptual novelty by placing the SMU not simply as an instrument of communication but a psychological tool, which becomes part of the digital workplace. This arrangement offers fresh grounds of theorizing technology as a dual agent a demand and a resource as the perceptions and uses of employees of technology vary. The findings motivate the researchers to explore further how the digital technologies interact with human psychology, leading to new theoretical frameworks of work-related stress, coping, and resilience in future studies. This study does not merely build upon the existing theories but it refines and broadens the existing theories to reflect more accurately the realities of digital work environment today. It paves the way to additional investigation into creating a more contextual and balanced model of the effect of digitalization on human capital, mental well-being, and organization performance.

Overall, study highlights SMU as a strategic asset of modern workplaces, not only as the means of communication, but as the instrument of psychology. SMU indirectly maintains mental well-being, and decreases the adverse effects of DS. The supportive role of SMU has tangible benefits that can be exploited by organizations even though the results of performance improvement will not be achieved directly by mental well-being. This is reinforced by the existing literature and generates practical implications of the policies of human resource and technology.

6.2 Managerial implications

The study has some managerial implications. As digital technology and platforms continue to reshape the limits of work, managers have to be proactive in crafting policies and interventions that balance technological innovation with employee well-being and performance results. This observation invites managers to reconsider DS not only as a risk, but also as a possible performance driver, as long as employees are equipped with the proper tools and resources. It stresses that it is necessary to foster a work culture which considers digital adaptability and resilience as strategic abilities, instead of liabilities to be avoided. The study presents evidence that SMU can also be a beneficial organizational resource. Managers should develop intentional and well-designed strategies that allow SMU including internal enterprise social networks, peer mentoring communities, and knowledge-sharing applications that are connected to work. Such programs should encourage professional networking, sharing of information, and emotional support, not recreation or entertainment through SMU. Structured and work-related social media platforms can be used to reinforce employee connectedness and psychological support. The results of this study do not, however, reflect the acceptance of unlimited or excessive interaction on SMU, but rather of meaningful and supportive interaction in the workplace. They can be psychological support devices, mitigating isolation and increasing social connectedness, particularly within remote or hybrid work environments. Instead of banning SMU altogether, organizations should have guidelines that allow purposeful, supportive and work-related interactions while discouraging an excessive amount of non-work interactions. The mediating role of MW between DS and EP emphasizes the need for employers to attend to employee MW as a central element of performance management. Managers have to move beyond standard wellness programs and incorporate mental well-being initiatives into digital work policy. This might involve access to mental well-being tools, digital detox intervals, mindfulness sessions, and workload management procedures that minimize cognitive overload. the research indicates that not all digital interventions result in productivity. The study upholds the coordination of workplace practices with larger global objectives, such as the SDG 3 and SDG 8. Organizations that invest in digital stress reduction and social support systems not only enhance individual performance but also enhance long-term sustainability, employee retention, and innovation capacity. Finally, the research empowers managers to turn digital distress into a strength by embracing an integrated approach that weaves technology use, mental well-being, and performance-focussed HR practices together. Organizations that learn to do so will be able to better compete in the digital era while enjoying a healthy, engaged, and resilient workforce. However, the protective effect of SMU should be interpreted cautiously. Therefore, it is important that organizations have clear policies regarding digital communication and SMU, which sets boundaries between constructive and informative use and passive use, or excessive and addictive use. Given the results of this study, SMU is recommended in professional settings, but cannot be extrapolated to general SMU. HR policies should, therefore, promote reasonable, intentional and professionally appropriate use and reduce overuse and non-work use during work time.

From an HR practice perspective, the findings offer a number of actionable recommendations. HR teams can use DS assessments as part of employee well-being surveys to pinpoint employees in high levels of digital strain and develop specific action plans. For example, banks can implement a digital well-being initiative that incorporates digital fasts, guidelines on technology use, stress reduction seminars, or periodic mental well-being assessments. HR teams can also start an internal social media use or an enterprise collaboration tool to help employees support each other, share knowledge and feel connected, especially when working remotely or as a hybrid team. In addition, digital resilience and tech adaption training can be included in onboarding and ongoing learning programs to increase and build employee confidence in their use of digital technology, while reducing stress caused by technology. All these efforts can contribute to improved employee mental well-being, improved organizational resilience and better sustainable performance of employees.

6.3 Conclusions, limitations and future directions

The study has certain limitations, Firstly, the study is cross-sectional in nature, measuring the associations among DS, SMU, MW, and EP at one point in time. Although this method gives us a snapshot of employee's attitudes, it restricts us from making causal statements. Longitudinal or experimental designs could be taken in future studies to explore how these associations change across time, particularly as digital technologies and organizational practices continue to evolve. Secondly, the research was based on self-reported information, which is open to biases like social desirability bias. While measures were taken to guarantee anonymity and minimize response bias, future research might usefully include multi-source information for instance, supplementing employee questionnaires with managerial ratings or objective performance data to increase the robustness and validity of the results. Thirdly, the sample is geographically confined to bank professionals in the banking industry of Uttarakhand, India. Although the setting offers useful insights into DS in a high-stress industry, the results might not be entirely generalizable to other industries, cultures, or countries. In the present study only purposeful and supportive SMU in the context of work is of interest. Future studies might be able to distinguish between passive browsing, entertainment use, professional networking, and supportive communication and investigate the potentially distinct benefits of each to the employee MW and EP.

Future studies should seek to replicate this study in various cultural and occupational environments to test the universality or context-specificity of the model. Fourthly, although the research centers on SMU as a coping tool, it does not distinguish between the frequency, quality, or types of social media interaction. Various platforms (e.g., LinkedIn vs. Instagram) or patterns of use (passive browsing vs. active engagement) might differentially influence stress and performance. Future research would do well to investigate the grain of digital behaviour, testing how various social media affordances work together with personality characteristics or occupational demands to generate unique effects. Fifth, MW was conceptualized as a unidimensional construct, although it is really multidimensional and includes emotional, psychological, and social well-being. Disaggregation of MW and specification of which dimensions are most impacted by digital stress and strongest at mediating performance outcomes should be the focus of future research. Lastly, although this study draws from the JD-R and COR theories, future research could integrate other theoretical frameworks such as self-determination theory or social exchange theory to better understand motivation, trust, and reciprocity in digital work environments. This would be able to bring a richer theoretical understanding of how employees engage with technology, manage stress, and derive meaning from digital interactions.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Ratih Pratiwi, Universitas Wahid Hasyim, Indonesia

Reviewed by: Octavian Dospinescu, Alexandru Ioan Cuza University, Romania

Tongqiang Dong, Qufu Normal University, China