The effectiveness of well-being-oriented human resource management in the context of telework

Abstract

Organizations can support employee well-being through a wide variety of HRM practices, which can be summarized as well-being-oriented HRM (WBHRM). It is unclear, however, whether these practices can also foster the well-being of teleworkers. Indeed, due to reduced visibility and salience, it is likely that WBHRM is less or even not effective in the telework context. To prove this assumption, we analyze how strong specific domains of WBHRM are related to different dimensions of employee well-being (including happiness, health, and relational well-being) in a non-telework and telework context. In the telework context, we also analyze the moderating role of telework intensity. Our analyses are based on data from 1,980 German employees at two consecutive points in time. Our findings reveal that the overall relationship between WBHRM and employee well-being in terms of happiness (engagement, job satisfaction) and health (strain) is not weakened in the context of telework. Nevertheless, we found differences in the importance of specific domains to foster happiness and health well-being. In terms of relational well-being, we found that WBHRM is positively related to social isolation in the telework and non-telework context (which is counterintuitive). We did not find support for any moderating effect through telework intensity.

Keywords:

• Telework
• HRM
• well-being
• job satisfaction
• engagement
• strain
• social isolation

Introduction

Telework, also known as telecommuting or remote work (Gajendran & Harrison, 2007), is on the increase, at least since the COVID-19 pandemic has forced a large number of employees to work from home (Eurofound, 2020). It can have manifold effects on employees, some of which are positive, such as greater autonomy and a better work-life balance (Eurofound, 2020; Gajendran & Harrison, 2007; Sardeshmukh et al., 2012). However, telework can also have negative effects, such as increasing information and communication technology (ICT) demands and social isolation (Becker, Belkin et al., 2022; Day et al., 2012; Wong et al., 2022). The question of how to ensure the well-being of spatially dispersed workers is therefore of growing interest (e.g. Becker, Belkin et al., 2022; Song & Gao, 2020; Straus et al., 2022; Wang et al., 2021).

Organizations can support employee well-being by applying human resource management (HRM) practices that specifically target employee well-being. Following Guest (2017), such a well-being-oriented HRM (WBHRM) approach includes five domains, namely investing in employees, providing engaging work, a positive social and physical environment, voice, and organizational support. However, while this approach seems promising in a regular work context, it is questionable whether it is equally effective in a telework context. Research on HRM processes highlights that the effectiveness of HRM practices depends on their visibility and salience, with lower visibility and salience being associated with a lower influence of HRM practices (Bowen & Ostroff, 2004; Garg et al., 2021). Teleworkers face a lack of organizational symbols and face-to-face interactions with other organizational members (Ashforth, 2020; Wiesenfeld et al., 2001), which might reduce the visibility and salience of WBHRM and thus its effectiveness.

Research comparing the effectiveness of the same HRM strategy (e.g. WBHRM) in the telework and non-telework context is, to our best knowledge, not yet available. Typically, studies on telework focus on the experiences of teleworkers and the question of how organizations can adapt their HRM strategies when employees telework (e.g. Günther et al., 2022; Wang et al., 2021). However, organizations may not have the capacity to adapt their HRM strategies, relying on common measures without knowing whether they are equally effective in different contexts. This could lead to inefficiencies if established HRM strategies are less or even not effective in the telework context. Against this background, we analyze the following research question: Is the effectiveness of WBHRM compromised in a telework context compared to a non-telework context? In particular, we analyze how strong specific domains of Guest’s (2017) WBHRM strategy are related to different dimensions of employee well-being (including happiness, health, and relational well-being) in telework and non-telework contexts. Based on the assumption that the relationship between WBHRM and employee well-being becomes weaker as the intensity of telework (i.e. the number of days per week an employee teleworks) increases, we further analyze the moderating role of telework intensity.

Our study makes several contributions. First, we contribute to the telework literature by providing further empirical evidence on how HRM can support employee well-being as one of the key challenges of telework. While the vast majority of research focuses only on selected aspects of HRM (Becker, Belkin et al., 2022; Mihalache & Mihalache, 2022; Straus et al., 2022; Wang et al., 2021), our reference to Guest’s WBHRM approach provides a more holistic perspective on the influence of HRM in the telework context. Second, by comparing the relationship between WBHRM and the well-being of teleworkers and non-teleworkers, we provide insights into the contextual boundaries of HRM. The influence of HRM is supposed to be highly contextspecific (Jackson et al., 2014). In particular, context should affect how employees perceive HRM practices (van Beurden et al., 2021). Guest’s framework is conceptual in nature and does not consider the influence of specific contexts. With our research, we make a step in improving this situation. Third, we further contribute to the telework literature by considering possible moderating effects of telework intensity on the relationship between HRM and employee well-being, which has largely been neglected (Becker, Belkin et al., 2022; Mihalache & Mihalache, 2022; Straus et al., 2022; Wang et al., 2021). Fourth, by considering the multidimensionality of WBHRM and employee well-being, we provide a nuanced understanding that accounts for potential differences in specific aspects of employee well-being. Previous studies have usually focused on certain aspects of employee well-being (Becker, Belkin et al., 2022; Mihalache & Mihalache, 2022; Russo et al., 2020; Song & Gao, 2020; Straus et al., 2022). However, this might be a too narrow understanding since HRM can have differentiated and even trade-off effects on different dimensions of employee well-being (Guerci et al., 2022).

Theory and hypotheses

Telework and its effects on teleworkers’ well-being

Employee well-being can be described as the overall quality of an employee’s experience and functioning at work (Warr, 1987). It has been studied and measured in different ways. One of the most common conceptualizations was proposed by Grant et al. (2007), who identified three dimensions of employee well-being. The first dimension concerns happiness, which can be divided into hedonic and eudaimonic well-being. Hedonic well-being is concerned with a subjective feeling of happiness such as a positive affect towards the job. Eudaimonic well-being describes an employee’s feelings of fulfillment and the realization of valued human potential. The second dimension of employee well-being is health and encompasses the freedom from physical (e.g. backache or headaches) and mental illnesses (e.g. depression or anxiety). Third, relational well-being describes the perceived quality of employees’ social interactions such as levels of social support, perceived trust, or fairness of treatment.

Teleworking can have manifold effects on employee well-being, some of which are positive such as higher job satisfaction (e.g. Gajendran & Harrison, 2007; Karácsony, 2021; Syrek et al., 2022) or better work-life balance (Eurofound, 2020; Gajendran & Harrison, 2007; Karácsony, 2021). However, teleworking can also pose a threat to the different dimensions of employee well-being. Regarding happiness well-being, Sardeshmukh et al. (2012) and Straus et al. (2022) found telework negatively related to job engagement. Furthermore, while previous studies have often found a positive relationship between telework and job satisfaction, other studies have found that not to be evident in all cases (Golden, 2006; Möhring et al., 2021). For example, Golden (2006) found a curvilinear relationship between telework and job satisfaction, such that lower levels of teleworking are positively related to job satisfaction, but higher levels of teleworking are negatively related to job satisfaction.

Relational well-being can also be compromised by telework since teleworkers and their colleagues work spatially separated from each other, which is associated with reduced communication and a decrease in regular face-to-face interactions with peers and managers (van Zoonen & Sivunen, 2022; Wong et al., 2022). Therefore, it is no surprise that social isolation has been identified as a key challenge of teleworking (e.g. Becker, Belkin et al., 2022; Carillo et al., 2021; van Zoonen & Sivunen, 2022; Wong et al., 2022).

Finally, telework can also be a threat to employees’ health (e.g. Carillo et al., 2021; Heiden et al., 2021; Song & Gao, 2020; Wöhrmann & Ebner, 2021). Blurring boundaries between work and non-work hours as well as the availability of technology to continue working during non-work hours can lead to work-family conflicts (Molino et al., 2020; van der Lippe & Lippényi, 2020) and a lack of recuperation, as employees may fail to mentally detach and switch-off from work (Charalampous et al., 2019). Moreover, teleworkers also need to face different ICT demands which are associated with an increase in teleworkers’ strain and (techno-)stress (Day et al., 2012; Molino et al., 2020; Suh & Lee, 2017).

Well-being-oriented HRM

Common HRM strategies such as high-performance work systems (HPWS) (Appelbaum et al., 2000; Huselid, 1995) or high-commitment HRM (Walton, 1985) aim to increase employee performance. While these HRM strategies may also affect employee well-being, this is not their primary goal, and the well-being effects of these strategies are controversial (see the discussion on mutual gains vs. conflicting outcomes; e.g. Peccei & van de Voorde, 2019). Given that employee well-being is a key moral obligation of organizations and a potential means to increase performance, Guest (2017) proposed the concept of WBHRM as an alternative analytical framework to understand the impact of HRM on employee well-being. Drawing mostly on the JD-R model (Bakker & Demerouti, 2007) and the core conditions for high quality of working life (Walton, 1974), he proposed five sets of HRM practices that should foster employee well-being. The first domain concerns investments in employees to enhance their competencies, create a feeling of security, and aid the development of self-efficacy. This can, for example, be achieved through training and development, mentoring, and career support. The second domain, providing engaging work, describes the need to create opportunities for control, skill use, and variety at work. Therefore, the core HRM practice for promoting engaging work is job design. The third domain, positive social and physical environment, includes prioritizing employee health and safety, providing opportunities for social interaction at work, avoiding bullying and harassment, promoting equal opportunities and diversity, and ensuring fair rewards and employment security. The fourth domain describes the need to give employees a voice. Relevant HRM practices include extensive two-way communication, attitude surveys, and relevant forms of collective representation. The final domain describes the importance of organizational support to ensure employee well-being. The relevant HRM practices reflect widely recognized antecedents of well-being, such as participative and supportive management, involvement climate and practices, flexible and family-friendly work arrangements, or developmental performance management.

Following the argumentation of Guest (2017), we expect WBHRM to positively affect employee well-being.

Hypothesis 1: WBHRM is positively related to employee happiness well-being.

Hypothesis 2: WBHRM is positively related to employee health well-being.

Hypothesis 3: WBHRM is positively related to employee relational well-being.

Visibility and salience of HRM in the context of telework

Over the last decade, researchers have paid increasing attention to employees’ perception of HRM (Beijer et al., 2021; Hewett et al., 2018; van Beurden et al., 2021; Wang et al., 2020). Most studies in this research field are based on the seminal work of Bowen & Ostroff (2004), who highlight that ‘HRM practices communicate messages constantly and in unintended ways, and messages can be understood idiosyncratically, whereby two employees interpret the same practices differently’ (p. 206). The unequal perception of HRM can refer to different components, namely the what, the how, and the why of an HRM practice (Wang et al., 2020). The what considers the content of an HRM practice through which the organization sends a message to the employees. The why considers discrepancies about how employees judge the underlying motivation of the organization to implement the respective HRM practice, also referred to as HR attribution (Hewett et al., 2018; Hewett, 2021; Nishii et al., 2008). The how considers that the same HRM practice may lead to different outcomes depending on how the practice is framed and received by employees.

Bowen and Ostroff (2004) further argue that HRM needs to send unambiguous messages in order to be effective. Thereby, they introduce the concept of HRM systems strength, with a strong situation being defined as a situation in which the employees share a common understanding of an organization’s HRM practices and the behaviors that are expected and rewarded. Bowen and Ostroff (2004) identified different factors (i.e. distinctiveness, consistency, and consensus) that influence the strength of an HRM system. Among those factors, the visibility and salience of HRM practices can be considered particularly relevant to the question of whether HRM practices might be perceived differently in a telework vs. a non-telework context. Visibility and salience are basic prerequisites for employees’ perception and interpretation of HRM. However, individuals are selective consumers of stimuli (Markus & Zajonc, 1985) due to the limitations of the human mind to attend to and process large amounts of information (Fiske & Taylor, 1991). This implies that individuals’ attitudes and behaviors are influenced by the information that is more prominent in the cognitive field (Schneider, 2000). To teleworkers, who are physically separated from their colleagues and work outside of their organization’s premises, their organization as a whole is less prominent in their cognitive field (Ashforth, 2020; Sardeshmukh et al., 2012; Wiesenfeld et al., 2001), and thus also their organization’s HRM practices are less salient to them. Since the low salience of an HRM practice leads to a less sensitive influence of this measure (Garg et al., 2021), teleworkers should be less affected by HRM practices in their well-being than non-teleworking employees. Therefore, we predict that the relationship between WBHRM and employee well-being is weaker in a telework context compared to a non-telework context.

Hypothesis 4: The relationship between WBHRM and employee happiness well-being is weakened in the context of telework compared to a non-telework context.

Hypothesis 5: The relationship between WBHRM and employee health well-being is weakened in the context of telework compared to a non-telework context.

Hypothesis 6: The relationship between WBHRM and employee relational well-being is weakened in the context of telework compared to a non-telework context.

As described above, the visibility and salience of an HRM practice are important factors in how employees perceive this HRM practice (Bowen & Ostroff, 2004; Garg et al., 2021), and both factors might be reduced in the telework context due to the lack of organizational trappings and face-to-face interactions with other organizational members (Ashforth, 2020; Sardeshmukh et al., 2012; Wiesenfeld et al., 2001). Building on these assumptions, we further assume that the relationship between WBHRM and employee well-being becomes weaker with higher telework intensity, since the visibility and the salience of the WBHRM system is compromised the more employees work from home. Therefore, we predict a moderating effect of telework intensity on the relationship between WBHRM and employee well-being.

Hypothesis 7: A higher extent of telework intensity weakens the relationship between WBHRM and teleworker happiness well-being.

Hypothesis 8: A higher extent of telework intensity weakens the relationship between WBHRM and teleworker health well-being.

Hypothesis 9: A higher extent of telework intensity weakens the relationship between WBHRM and teleworker relational well-being.

Furthermore, we expect that the relative importance of WBHRM domains may differ in the context of telework compared to traditional office work. For example, supportive and participative management, flexible and family-friendly work arrangements, provision of health care, autonomy, and social contacts have been identified as key resources for teleworkers (Franken et al., 2021; Günther et al., 2022; Russo et al., 2020; Straus et al., 2022). Therefore, the WBHRM domains of organizational support, a positive social and physical environment, and engaging work, which include the abovementioned measures, might be more important in the telework context than in the non-telework context. However, empirical research that compares the importance of HRM practices in the two contexts and therewith supports such assumptions is, to the best of our knowledge, not yet available. Therefore, we refrain from formulating further hypotheses at this point and approach the investigation of the importance of specific WBHRM domains in an exploratory manner.

Method

Sample and procedure

We collected our data in Germany using an online questionnaire which was distributed by a professional market research institute. Data was collected at two different points in time. T1 was conducted between the end of March and mid-April 2021, and T2 was conducted from mid-June up to July 2021. Our sample results in a total of N = 1980 employees completing both surveys, of which 50.6% were male and 49.4% were female. The participants were working in a wide range of branches (e.g. 15.6% in public administration; 10.3% in metal and electrical industry; 8.7% in banking and insurance; 6.1% in logistics, transport, and traffic; 8.4% in IT, telecommunication). They had a mean age of 45.21 years (SD = 12.94).

For a better comparison of non-teleworkers and different types of teleworkers, we aimed for a sample in which non-teleworkers, rare users (1 day a week or less), medium users (2–4 days a week) and almost only users (5 days a week or more) where equally distributed. With regard to non-teleworkers, we only included responses from employees whose jobs, according to their own assessment, could theoretically be done from home. Therewith, we wanted to make sure that we were not just comparing different types of employees (i.e. teleworkers with employees whose jobs cannot be done from home). The final sample included N = 300 non-teleworkers, N = 181 rare users, N = 287 medium users with two days, N = 325 medium users with three days, N = 256 medium users with four days, and N = 631 almost only users.

To reduce the risk for common method bias (Podsakoff et al., 2013) while analyzing direct and indirect relationships between WBHRM practices and employee well-being, WBHRM practices were assessed at T1 and well-being related outcomes two months later at T2. Further, we used a marker variable analysis to test for common method bias. The questionnaire did not include a specific marker variable, therefore we used the smallest observed correlation among all the substantive variables as a proxy, as suggested by Lindell and Whitney (2001). The lowest correlations turned out to be below r = 0.001, thus common method bias should not affect our results.

Measures

To assess WBHRM, we followed Guest’s (2017) theoretical framework. We considered investing in employees to include comprehensive selection, extensive investments in training and mentoring, and career support. For providing engaging work we referred to jobs designed to provide autonomy and challenge, skill utilization, and the provision of information and feedback. We considered a positive social and physical environment to include the priority of health and safety, equal opportunities and diversity management, zero tolerance for bullying and harassment, required and optional social interaction, fair collective rewards and high basic pay, as well as employment security and employability. Voice consists of extensive two-way communication, employee surveys, and a collective representation. Organizational support includes participative and supportive management, involvement climate and practices, flexible and family-friendly work arrangements, and developmental performance management. Where possible, existing items were adopted from the literature (Jensen et al., 2013; Mostafa et al., 2015; Zhang et al., 2018). The complete list of items can be found in Appendix Table A1. We conceptualized WBHRM as a formative-formative higher-order construct, following Jiang et al. (2012) and Hauff (2021). Thereby, individual HRM practices formatively operationalize the domains of WBHRM (e.g. voice), which in turn formatively operationalize the overall construct.

Table 1. Relationship between WBHRM and employee well-being in the context of telework and non-telework.

Construct	Group	Happiness well-being			Psychological strain as indicator for health well-being			Social isolation as indicator for relational well-being
		Path β	p-value	95% CI	Path β	p-value	95% CI	Path β	p-value	95% CI
WBHRM	Non-telework	0.60	.000	[0.506; 0.661]	−0.40	.000	[−0.479; −0.287]	0.29	.000	[0.171; 0.377]
	Telework	0.60	.000	[0.536; 0.647]	−0.37	.000	[−0.432; −0.283]	0.23	.000	[0.140; 0.298]
	Differences	0.00	.487		0.04	.281		0.08	.105

Multi-group analysis based on a sample of N = 300 non-teleworkers and N = 631 high intensity teleworkers (5 or more days teleworking per week). Control variables were age, gender, change in telework intensity (T1 to T2) and prior experience with telework.

Happiness well-being was measured as a reflective second-order construct consisting of job satisfaction and work engagement. To assess job satisfaction, we used a single-item measure adopted from the German version of the Copenhagen Psychosocial Questionnaire. The reliability and validity of this measure were shown by Nübling et al. (2006). To measure work engagement, we followed Schaufeli et al. (2017), using the ultra-short measure for work engagement (UWES-3). Sample items were: ‘Within the past four weeks, I have been full of exuberant energy at my work’ or ‘Within the past four weeks, I have been completely absorbed in my work’. Cronbach’s Alpha was α = 0.90.

Health well-being was measured as perceived psychological strain, which we assessed using the irritation scale by Mohr et al. (2005). Due to parsimony, the scale was shortened to four items. Sample items were: ‘I have difficulty relaxing after work’ and ‘I get irritated easily, although I don’t want this to happen’. Cronbach’s Alpha was α = 0.87.

Relational well-being was assessed by the indicator social isolation, which we measured using a professional isolation scale by Golden et al. (2008). Due to parsimony, the scale was shortened to three items. Sample items were ‘I miss face-to-face contact with my co-workers’ and ‘I feel isolated’. Cronbach’s Alpha was α = 0.85.

To assess telework intensity, we asked participants about the extent to which they were teleworking on average over the past four weeks, using a six-point scale (from not at all; up to 1 day; 2 days; 3 days; 4 days; 5 or more days per week).

WBHRM and social isolation were rated on a five-point Likert scale from 1 = ‘not at all true’ to 5 = ‘completely true’. Work engagement and strain were rated on a five-point Likert scale from 1 = ‘never’ to 5 = ‘almost always’ and job satisfaction was rated on a five-point Likert scale from 1 = ‘very dissatisfied’ to 5 = ‘very satisfied’.

Furthermore, we considered different control variables that were mentioned as relevant to our dependent variables in previous studies (Donati et al., 2021; Gajendran & Harrison, 2007; Heiden et al., 2021; Park & Cho, 2022), or where we suspected a relationship in our research area. Namely, we considered age, gender (0 = male, 1 = female), marital status (married, not married, living separately, divorced or widowed, single), care responsibilities (0 = no responsibilities, 1 = care responsibilities), education level (1 = no educational qualification, 2 = secondary school diploma, 3 = middle school diploma, 4 = high school diploma, 5 = vocational training, 6 = university degree, 7 = doctorate), branch (0 = public sector, 1 = private sector), employment status (0 = temporary employment, 1 = permanent employment), leadership responsibilities (0 = no leadership responsibilities, 1 = leadership responsibilities), prior experience with telework and changes in telework intensity between T1 and T2 (0 = changes, 1 = no changes).

Analysis

To allow a nuanced understanding of the differentiated effects of WBHRM on employee’s well-being, while simultaneously avoiding a bias that might arise from unknown effects between the well-being dimensions (Günther et al., 2022), we calculate three different models, one for each dimension of employee well-being (Figure 1).

Figure 1. Conceptual Models.

To test our hypotheses, we conducted two different analyses: a multi-group analysis to highlight the differences between teleworkers and non-teleworkers and a moderation analysis to illustrate the influence of telework intensity. In the multi-group analysis, we considered participants who teleworked either zero or five days per week, resulting in a sample size of N = 931. Hybrid workers were excluded from the multi-group analysis to make valid statements about the influence of different contexts avoiding unintended influences and spillovers (e.g. social isolation in the context of telework vs. in the context of traditional office work). In the moderation analysis, we considered participants who telework between one and five days per week, resulting in a sample size of N = 1680.

Following Nielsen and Raswant’s (2018) recommendations, we first estimated initial models that included all of the control variables mentioned above. In the next step, we removed all control variables without a significant effect on our dependent variables. The remaining control variables were added to the final models (Figure 1). A correlation matrix for the multi-group analysis can be found in Appendix Table A3 and for the moderation analysis in Appendix Table A4.

To test our final models, we chose partial-least-squares structural equation modeling (PLS-SEM) employing the SmartPLS 3 software (Ringle, 2015). PLS-SEM was particularly useful since we conceptualized WBHRM as a formative-formative higher-order construct (Ringle et al., 2020). Following the recommendations of Becker, Cheah et al. (2022), we chose the two-stage approach, utilizing latent variable scores of the latent predictors to estimate the higher-order component in our model. In all steps of our analysis, we used the following settings: path weighting scheme, 300 iterations, stop-criterion 0.0000001, and replaced missing values by the mean value. By applying the bootstrapping procedure, we determined the significance with the following settings: 5000 subsamples, as many observations per subsample as in the original sample, and the no sign change option.

Results

Measurement model

To ensure the validity and reliability of our reflective measurement models, we followed the recommendations of Hair and colleagues (2017) and evaluated the indicator reliability, internal consistency, and convergent and discriminant validity. Strain had one item slightly below the threshold of 0.708 in the multi-group analysis. Nevertheless, AVE= 0.76 and CR= 0.92 were satisfactory (Hair et al., 2019). In the moderation analysis, the same item of strain was slightly below the threshold and again, AVE = 0.73 and CR = 0.91 were satisfactory. All items of engagement had significant loadings above the threshold of 0.708, both in the multi-group and moderation analysis. AVE and CR were both satisfactory in the multi-group analysis (AVE = 0.74; CR = 0.92) and the moderation analysis (AVE = 0.83; CR = 0.94). All items of social isolation had significant loadings above the threshold of 0.708, both in the multi-group and moderation analysis. AVE and CR were both satisfactory in the multi-group analysis (AVE = 0.79; CR = 0.92) and the moderation analysis (AVE = 0.76; CR = 0.90). We did not encounter an issue with discriminant validity, as all items contain a value in the heterotrait-monotrait (HTMT) criterion below 0.85. This applies to both the multi-group analysis and the moderation analysis.

We evaluated our lower and higher-order formative models based on multicollinearity and their relevance for the formative construct. None of the variance inflation factor (VIF) scores indicated an issue of multicollinearity, since none of the values exceeds 5 (Hair et al., 2017; Hair et al., 2019).

In the multi-group analysis, the weights of the indicators were predominantly significant. To determine the relevance of the indicators whose weights were not significant, their outer loadings were examined. Three lower-order indicators fell below the threshold of 0.5. They were retained in the model owing to the significance of their loadings. None of the higher-order indicators had loadings below the threshold of 0.5 and all of them were significant. An overview of the weights and loadings of the indicators is given in Appendix Table A1.

In the moderation analysis, the weights of the indicators were predominantly significant. Once again to determine the overall relevance of the indicators whose weights were not significant, their outer loadings were examined. The loadings of four lower-order indicators slightly fell below the threshold of 0.5. They retain in the model owing to the significance of their loadings. None of the higher-order indicators had loadings below the threshold of 0.5 and all of them were significant. An overview of all weights and loadings of the formative measurement models in the moderation analysis is given in Appendix Table A2. Overall, these results confirm the satisfactory quality of our measurement models.

Table 2. Moderating role of telework intensity in the relationship between WBHRM and employee well-being.

Construct	Happiness well-being			Psychological strain as indicator for health well-being			Social isolation as indicator for relational well-being
	Path β	p-value	95% CI	Path β	p-value	95% CI	Path β	p-value	95% CI
WBHRM	0.55	.000	[0.508; 0.586]	−0.34	.000	[−0.388; −0.292]	0.19	.000	[0.135; 0.236]
Telework intensity	−0.05	.008	[−0.091; −0.013]	0.00	.876	[−0.051; 0.041]	0.02	.331	[−0.024; 0.073]
Moderating effect of telework intensity	0.02	.464	[−0.026; 0.057]	−0.01	.733	[−0.058; 0.04]	0.02	.486	[−0.035; 0.075]

Moderation analysis based on a sample of N = 1777 teleworkers with a telework intensity between 1 and 5 days per week. Control variables were age, gender, marital status, care responsibilities, change in telework intensity (T1 to T2) and prior experience with telework.

Structural model

The control models in the multi-group analysis can explain 4.4% variance in happiness well-being, 3.1% variance in health well-being, and 3.6% variance in relational well-being. The full models on the other hand can explain 36.1% variance in happiness well-being, 16.7% variance in health well-being, and 10.3% variance in relational well-being. The control models in the moderation analysis can explain 4.2% variance in happiness well-being, 3.2% variance in health well-being, and 2.5% variance in relational well-being. The full models can explain 31.4% variance in happiness well-being, 14.7% variance in health well-being, and 5.3% variance in relational well-being. Thus, all values of the full models significantly exceeded the values of the control model.

Relationships between WBHRM and employee well-being

The results in Table 1 show that WBHRM is positively related to happiness well-being, both for teleworkers (β = 0.60; p < .001) and non-teleworkers (β = 0.60; p < .001), supporting Hypothesis 1. Furthermore, the results in Table 1 show, that WBHRM is negatively related to psychological strain, both for teleworkers (β = −0.40; p < .001) and non-teleworkers (β = −0.37; p < .001), which in turn represents a positive relationship between WBHRM and health-related well-being, supporting Hypothesis 2. For relational well-being (Table 1) we found a positive relationship between WBHRM and social isolation for teleworkers (β = 0.23; p < .001) and non-teleworkers (β = 0.29; p < .001). Both results contradict Hypothesis 3.

The relationships of WBHRM and employees’ well-being do not show significant differences between teleworkers and non-teleworkers. This was found in case of happiness well-being (β = 0.00; p = .487), psychological strain (β = 0.04; p = .281) and social isolation (β = 0.08; p = .105). Hypotheses 4, 5, and 6 are thus rejected.

Furthermore, we hypothesized that telework intensity moderates the relationship between WBHRM and employees’ well-being. Table 2 shows non-significant moderating effects for happiness (β = 0.02; p = .464), health (β = −0.01; p = .733) and relational well-being (β = 0.02; p = .486) contradicting Hypotheses 7, 8 and 9.

The importance of specific domains

Another aim of this study was to gain insights into which domains of WBHRM are particularly important to foster the well-being of teleworkers and non-teleworkers. This relative importance is represented by the weights of the domains, as they reflect the value that specific domains contribute to promoting employee well-being through WBHRM as a whole. The full results can be found in Table 3.

Table 3. The importance of specific domains in the relationships between WBHRM and different dimensions of teleworkers and non-teleworkers well-being.

Construct	Group	Happiness well-being			Psychological strain as indicator for health well-being			Social isolation as indicator for relational well-being
		Path β	p-value	95% CI	Path β	p-value	95% CI	Path β	p-value	95% CI
Investing in employees	Non-telework	0.04	.716	[−0.176; 0.258]	−0.37	.031	[−0.325; 0.410]	0.21	.259	[−0.163; 0.572]
	Telework	0.11	.154	[−0.041; 0.263]	−0.08	.527	[−0.526; 0.129]	0.63	.002	[0.237; 1.037]
	Difference	0.07	.306		0.29	.087		0.42	.069
Providing engaging work	Non-telework	0.65	.000	[0.371; 0.948]	0.29	.112	[−0.716; −0.034]	−0.40	.005	[-0.675; −0.117]
	Telework	0.34	.000	[0.152; 0.534]	0.30	.102	[−0.339; 0.170]	−0.21	.424	[−0.730; 0.309]
	Difference	0.31	.960		0.01	.484		0.19	.257
Positive social and physical environment	Non-telework	0.02	.890	[−0.309; 0.341]	0.74	.000	[0.051; 0.718]	0.39	.007	[0.099; 0.648]
	Telework	0.50	.000	[0.293; 0.697]	0.82	.000	[−0.195; 0.504]	0.30	.155	[−0.124; 0.710]
	Difference	0.47	.007		0.08	.359		0.09	.628
Voice	Non-telework	0.33	.018	[0.070; 0.605]	0.03	.858	[0.274; 0.339]	0.53	.000	[0.243; 0.808]
	Telework	−0.24	.015	[-0.434; −0.048]	−0.19	.254	[0.241; 0.293]	0.20	.414	[−0.258; 0.687]
	Difference	0.56	.999		0.23	.813		0.33	.877
Organizational support	Non-telework	0.04	.780	[−0.230; 0.292]	0.39	.025	[0.166; 0.315]	0.18	.294	[−0.138; 0.532]
	Telework	0.36	.000	[0.184; 0.552]	0.16	.381	[0.202; 0.291]	0.26	.255	[−0.219; 0.672]
	Difference	0.32	.024		0.23	.822		0.08	.387

Multi-group analysis based on a sample of N = 300 non-teleworkers and N = 631 high intensity teleworkers (5 or more day’s teleworking per week). Control variables were age, gender, change in telework intensity (T1 to T2) and prior experience with telework.

In terms of happiness well-being, our results show significant differences in the importance of specific domains for the well-being of teleworkers and non-teleworkers. The domain providing engaging work in particular (β = 0.65; p < .001) as well as the domain voice (β = 0.33; p = .018) are important to foster happiness well-being in the non-telework context, while the remaining domains are not significant. In the context of telework, providing engaging work (β = 0.34; p < .001), positive social and physical environment (β = 0.50; p < .001) and organizational support (β = 0.36; p < .001) are significant and therefore important to increase happiness well-being. Contrary to the non-telework context, voice is negatively related to happiness well-being in the telework context (β = −0.24; p = .015).

Further, our results show significant differences in the importance of specific domains to foster health well-being between teleworkers and non-teleworkers. The domain positive social and physical environment is particularly important to foster the health well-being of teleworkers (β = 0.82; p < .001) and non-teleworkers (β = 0.74; p < .001). Non-teleworkers’ health well-being can further be promoted by organizational support (β = 0.39; p = .025). Investing in employees, however, is negatively related to the health well-being of non-teleworkers (β = −0.37; p = .031), while the relationship is non-significant in a telework context (β = −0.09; p = .527). The remaining domains are non-significant for teleworkers and non-teleworkers.

In the case of relational well-being, our results show that investing in employees is the only driver for the relationship between WBHRM and social isolation in the teleworking context (β = 0.63; p = .002), while the remaining domains are non-significant. In a non-telework context, a positive social and physical environment (β = 0.39; p = .007) and voice (β = 0.53; p < .001) are responsible for an increase in social isolation, whereas engaging work (β = −0.40; p = .005) with a negative weight, counteracts this relation.

In summary, WBHRM is positively related to happiness and health well-being, while at the same time, it increases social isolation. These results were found for both teleworkers and non-teleworkers. Furthermore, telework intensity does not moderate the relationship between WBHRM and employees’ well-being. Differences between teleworkers and non-teleworkers were found only in the strength of the different weights that cause the relationship between the system and employee well-being.

Discussion

Implications for theory and research

By analyzing the relationship between WBHRM and employee well-being in the context of telework and non-telework, our study makes several contributions. First, we contribute to the telework literature by providing further empirical evidence on how HRM can support teleworkers’ well-being. While the vast majority of prior research focuses only on selected aspects of HRM (Mihalache & Mihalache, 2022; Straus et al., 2022; Wang et al., 2021), we provide a holistic perspective on HRM in the telework context. Thus, we are able to show that, although not all domains of Guest’s (2017) WBHRM approach contribute with a significant weight, a wide variety of different HRM practices can be applied to promote teleworkers’ happiness and health well-being. For example, in line with Straus et al. (2022) and Becker, Belkin et al. (2022), we found that autonomy, as part of providing engaging work, is positively related to teleworkers’ well-being. Beyond this, we can show that jobs that are not only designed to provide autonomy, but an appropriate level of challenge and skill utilization as well as the provision of information and feedback, summarized in the domain of providing engaging work, can foster teleworkers’ happiness well-being. Further, consistent with existing research (Russo et al., 2020; Straus et al., 2022), we found high levels of job security and the promotion of teamwork, both as part of a positive social and physical environment, to be positively related to well-being. Beyond this, we found a variety of HRM practices such as high basic pay, diversity management, zero tolerance for bullying and harassment, prioritizing the health and safety of employees, and overload protection, which together form the domain of a positive social and physical environment, to promote happiness and health well-being of teleworkers. In line with prior research (Chong et al., 2020; Mihalache & Mihalache, 2022), we also found that supportive and participative management and flexible and family-friendly working arrangements, which form the domain of organizational support, can foster teleworkers’ happiness well-being.

Second, we provide insights into the contextual boundaries of HRM (Jackson et al., 2014). In this regard, our results show that the overall relationship between WBHRM and employee well-being in terms of happiness and health is not weakened in the context of telework compared to non-telework. In other words, our results show that WBHRM has the potential to promote the happiness and health well-being of teleworkers and non-teleworkers to a comparable extent. However, our results also show that certain WBHRM domains vary in their importance to promote employee well-being depending on whether employees are teleworking or not. In terms of happiness well-being, providing engaging work and giving employees a voice are the key drivers of non-teleworkers’ engagement and job satisfaction. To promote teleworkers’ happiness well-being, in addition to engaging work, organizational support and a positive social and physical environment were important. Previous research has considered several HRM practices in these domains as important resources for teleworkers (Franken et al., 2021; Russo et al., 2020; Straus et al., 2022). Thus, it is not surprising that these domains are more important to foster teleworkers’ happiness well-being. The differences in the relative importance of these WBHRM domains could also provide a possible explanation for our unsupported hypotheses about the weakened effectiveness of WBHRM in the telework context. Garg et al. (2021) point out that the salience of an HRM practice is also influenced by the extent to which it addresses the needs of employees. Since domains such as organizational support and a positive social and physical environment correspond to the specific challenges in the telework context (e.g., social isolation or work-family conflict), WBHRM remains salient and visible to teleworkers, despite the greater distance from the organization.

Our results further show that giving employees a voice can be negatively related to teleworkers’ happiness well-being, while it can be positively related in the non-telework context. Building on Guest (2017), we expected that voice is positively related to happiness well-being regardless of a telework or non-telework context. A possible explanation of the negative influence may be that expressions of voice can always be associated with costs. Indeed, to speak up means to perform additional tasks besides the regular work (Bolino & Turnley, 2005). Therefore, it might result in role overload, job stress, or work-family conflicts (Bolino & Turnley, 2005). In addition, teleworkers can only rely on mediated communication (e.g. via video conferencing), which makes it more difficult for them to participate compared to a non-telework context, for example, because they have problems with interpreting non-verbal communication cues (Taylor, 2011). Previous research found negative relationships between voice and employee well-being if employees were exposed to stressful working conditions. For example, Röllmann et al. (2021) found voice negatively related to employee well-being when job insecurity was high. Zacher et al. (2019) found an increase in proactive behavior promoting negative mood when organizational support was low. Applied to our results, the change in sign of the relationship between voice and happiness well-being when comparing teleworkers and non-teleworkers may be due to higher demands on employees in the telework context (e.g. through social isolation, technostress, work-family conflicts, and issues due to computer-mediated communication).

In the case of health well-being, our results show that a positive social and physical environment is the only driver to reduce the psychological strain of teleworkers. To reduce non-teleworkers psychological strain, organizational support is also important. Organizational support consists of supportive and participative management, an involvement climate and practices as well as family-friendly work arrangements. Teleworkers usually have an improved work-life balance anyway (Eurofound, 2020), so even more flexibility may not have a significant impact on their perceived psychological strain. Further, participative management and an involvement climate and practices may lead to additional tasks besides the regular job tasks, which in turn can lead to job stress (Bolino & Turnley, 2005), outweighing the positive effects of supportive management or flexible work arrangements. Investing in employees, however, was positively related to non-teleworkers psychological strain, while the relationship was non-significant in a teleworking context. A possible explanation could be that training and development as well as career management might cause additional demands in a regular working context (Bolino & Turnley, 2005; O’Brien et al., 2018). In contrast, teleworkers who had to face a variety of new job demands (e.g. ICT demands or work-family conflicts) might perceive such investments as a possibility to gain additional resources which may outweigh the potential detrimental effects of additional tasks.

Regarding relational well-being, our results revealed that WBHRM is related to stronger feelings of social isolation and thus negatively related to relational well-being in the context of telework and non-telework. In the case of telework, the domain of investing in employees is the only driver for this relationship since the remaining domains are not significant. One explanation could be that training and development measures might lead to even less contact between teleworkers and their colleagues for example due to off-the-job training. Anyway, the full model explains just 9.4% of the variation in teleworkers’ social isolation, which indicates that traditional HRM practices as included in Guest’s (2017) WBHRM approach only marginally explain variance in the relational well-being of teleworkers. Therefore, traditional HRM measures should be complemented by those specifically designed to reduce teleworkers’ social isolation, e.g. by promoting informal communication (Wang et al., 2021). In the case of non-telework, we found the domains of a positive social and physical environment and voice responsible for the increase in social isolation, while the domain of engaging work counteracts this effect. The latter is in line with prior research (Wang et al., 2021) and the effect of voice is comprehensible even if unexpected, since speaking up is not always appreciated by others and may even lead to damaged relationships (Milliken et al., 2003). However, we cannot explain the effect of a positive social and physical environment that reinforces social isolation. Further research is needed on this.

Third, we further contribute to the telework literature, by taking possible moderation effects of telework intensity between WBHRM and employee well-being into account, which has largely been neglected in research (Becker, Belkin et al., 2022; Mihalache & Mihalache, 2022; Straus et al., 2022; Wang et al., 2021). Our results show that telework intensity did not moderate any of the relationships between WBHRM and happiness, health, or relational well-being. In other words, the effectiveness of WBHRM practices is independent of the extent of telework in the context of hybrid work. Parallel to the results of the multi-group comparisons, this suggests that it is not the effectiveness of the WBHRM per se that changes, but only the domains that are responsible for the relationship between the WBHRM system and employee well-being.

Fourth, by considering different dimensions of employee well-being, our results provide a nuanced understanding that underlines the necessity of considering the multidimensionality of this construct, which has been largely neglected in research (Becker, Belkin et al., 2022; Mihalache & Mihalache, 2022; Straus et al., 2022). Indeed, in the telework context, our analysis revealed a trade-off effect of WBHRM, since it promotes teleworkers’ happiness and health well-being while simultaneously reducing relational well-being.

In summary, we provide further empirical evidence on how HRM can support teleworker well-being and show that WBHRM is not limited to a non-teleworker context. Nevertheless, it should be noted that depending on the context, different domains (and therefore different HRM measures) are responsible for the relationships between WBHRM and employee well-being. However, our results also show that WBHRM increases feelings of social isolation, whereby further research is needed to clarify the origin of this relation. To foster relational well-being, traditional HRM measures should therefore be complemented by those specifically designed to reduce employees’ social isolation, regardless of a teleworker or non-teleworker context. Further, we are able to show that the effectiveness of WBHRM practices is independent of the extent of teleworking in the context of hybrid work. Following the results of the multi-group comparisons, this could be due to the circumstance that the effectiveness of the WBHRM system does not change, but only the domains that are responsible for the relationships between the WBHRM system and employee well-being. Finally, we have also been able to show that a differentiated view of the construct of employee well-being is necessary. This is shown for example by the trade-off effect of WBHRM, as it promotes teleworkers’ happiness and health well-being while simultaneously reducing relational well-being.

Implications for organizational and managerial practice

Overall, our findings show that organizations are well advised to invest in WBHRM to foster the happiness and health well-being of teleworkers. Regarding happiness well-being, our findings demonstrate, however, that WBHRM domains vary in their influence depending on whether employees are teleworking or not. Thus, organizations who want to promote the happiness well-being of both teleworkers and non-teleworkers should consider adapting their strategy depending on the target group. For non-teleworking employees, they should focus on providing engaging work, for example through jobs designed to provide autonomy through an increased scope for decision-making, an appropriate challenge, and the opportunity for employees to use a variety of skills and abilities through job enrichment as well as giving their employees a voice, for example through fostering extensive two-way communication and collective representation such as a works council. For teleworkers, on the other hand, they should focus on providing engaging work, but also on creating a positive social and physical environment, for example, by providing high job security, promoting teamwork, and equal opportunities through functioning diversity management. Furthermore, they should also focus on providing organizational support for example through participative and supportive management or flexible and family-friendly work arrangements such as flexible working hours. In addition, organizations should be aware that giving teleworkers a voice may negatively affect their happiness well-being. Existing literature supports the assumption that the negative relation between voice and happiness well-being might be due to the greater demands employees face in the telework context. Therefore, they might have fewer, and possibly insufficient resources to speak up since that means performing additional tasks besides the regular work (Bolino & Turnley, 2005). Thus, reducing these demands might also prevent potential negative effects of voice. Furthermore, we found that in a telework context, WBHRM can lead to an increase in perceived social isolation and thus can reduce relational well-being. However, organizations should see this circumstance more as a reason to extend traditional WBHRM practices with those that are particularly designed to reduce teleworkers’ social isolation, for example by promoting informal communication (Wang et al., 2021).

Limitations and further research

Although we shed some light on the effectiveness of WBHRM in the context of telework and non-telework, considering the multidimensionality of WBHRM and employee well-being, our findings have limitations, which offer the potential for further research. First, even though we have incorporated a time lag of about three months between the collection of our independent and dependent variables, our study design does not allow for causal interpretation. Therefore, the derived implications of our study should be treated with caution. Second, our data was collected between March and July 2021, which bears the potential risk of having a bias due to the ongoing global pandemic situation. The vaccination campaign started in Germany at the beginning of 2021. Around March, the third wave of infection was broken, and the number of new infections decreased. At the time of our survey, the outbreak of the pandemic was about a year ago (Thurau & Bosen, 2021). Thus, employers and employees were rather familiar with the situation. Nevertheless, transferring the results to a non-pandemic situation should be done with caution. Further studies conducted after the end of the pandemic could make an important contribution to the telework literature.

Besides addressing these limitations, there are other potential avenues for future research. Our study reveals that telework bears specific challenges for teleworkers’ well-being, such as social isolation, which may not be overcome by traditional HRM practices alone. Therefore, additional studies could examine how HRM can best address challenges like social isolation. Future research could also examine if and how specific HRM practices or domains interact when influencing teleworkers’ well-being. Additional research could also explore whether the relationship between HRM and teleworkers’ well-being is dependent on further context variables such as branch, job type, or individual differences such as personality or experience with telework.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The datasets presented in this article are not readily available because of privacy restrictions.

Additional information

Funding

This work was supported by dtec.bw – Digitalization and Technology Research Center of the Bundeswehr.

Appendix

Table A1. Formative measurement models – multi-group analysis.

Construct and indicators (My organization…)	Well-being dimension	Weights/loadings	p-values	95% CI weights
Investing in employees
…uses a rigorous selection process to select new recruits for a position similar to mine.	Happiness	0.14/0.63	.003/.000	[0.049; 0.228]
	Health	0.19/0.67	.010/.000	[0.052; 0.330]
	Relational	0.63/0.90	.000/.000	[0.366; 0.899]
…invests considerable time and money in my training and development.	Happiness	0.39/0.87	.000/.000	[0.286; 0.485]
	Health	0.40/0.88	.000/.000	[0.270; 0.543]
	Relational	0.24/0.78	.084/.000	[−0.043; 0.490]
… actively promotes my career (e.g. through career planning, mentoring).	Happiness	0.61/0.94	.000/.000	[0.520; 0.703]
	Health	0.56/0.92	.000/.000	[0.409; 0.670]
	Relational	0.32/0.77	.036/.000	[−0.035; 0.558]
Providing engaging work
…allows me to make a lot of job decisions on my own.	Happiness	0.30/0.78	.000/.000	[0.221; 0.393]
	Health	0.34/0.79	.000/.000	[0.184; 0.511]
	Relational	0.01/0.56	.826/.000	[-0.125; 0.126]
…ensures that my job includes an appropriate level of challenge and variety to keep my work interesting.	Happiness	0.19/0.86	.000/.000	[0.095; 0.290]
	Health	0.08/0.83	.401/.000	[−0.117; 0.234]
	Relational	0.16/0.77	.042/.000	[−0.019; 0.306]
…keeps me informed about business issues and about how well it’s doing.	Happiness	0.15/0.74	.000/.000	[0.065; 0.232]
	Health	0.10/0.69	.229/.000	[−0.063; 0.235]
	Relational	0.32/0.80	.000/.000	[0.210; 0.441]
…provides meaningful feedback to me, regarding my performance at least once a year.	Happiness	0.23/0.73	.000/.000	[0.125; 0.325]
	Health	0.17/0.69	.046/.000	[−0.021; 0.308]
	Relational	0.59/0.91	.000/.000	[0.469; 0.796]
…ensures that I can use my skills and abilities appropriately.	Happiness	0.36/0.89	.000/.000	[0.263; 0.463]
	Health	0.53/0.93	.000/.000	[0.356; 0.753]
	Relational	0.09/0.73	.361/.000	[−0.139; 0.247]
Positive social and physical environment
…does everything possible to prevent accidents and health hazards in the workplace.	Happiness	0.07/0.66	.092/.000	[−0.012; 0.144]
	Health	0.13/0.69	.249/.000	[−0.087; 0.355]
	Relational	0.12/0.69	.123/.000	[0.001; 0.321]
…ensures that I am not overloaded by my work.	Happiness	0.23/0.78	.000/.000	[0.150; 0.316]
	Health	0.36/0.79	.016/.000	[0.140; 0.675]
	Relational	0.04/0.70	.731/.000	[−0.234; 0.181]
…offers practices and incentives to make me take more care of my health (e.g. sports courses, providing fruit).	Happiness	0.08/0.63	.027/.000	[0.012; 0.153]
	Health	−0.18/0.45	.065/.000	[−0.413; −0.018]
	Relational	0.32/0.77	.004/.000	[0.179; 0.584]
…ensures that all employees have the same opportunities and that no one is disadvantaged.	Happiness	0.15/0.80	.002/.000	[0.059; 0.247]
	Health	0.11/0.76	.373/.000	[−0.154; 0.316]
	Relational	0.25/0.84	.001/.000	[0.106; 0.409]
…does not tolerate any forms of bullying or discrimination at work.	Happiness	0.06/0.67	.168/.000	[−0.029; 0.148]
	Health	0.10/0.70	.415/.000	[−0.178; 0.307]
	Relational	0.11/0.71	.080/.000	[0.002; 0.275]
…strongly encourages teamwork.	Happiness	0.34/0.84	.000/.000	[0.259; 0.425]
	Health	0.39/0.84	.000/.000	[0.221; 0.610]
	Relational	0.23/0.80	.000/.000	[0.099; 0.349]
…provides an average pay level (including bonuses) to our employees that is higher than that of our competitors.	Happiness	0.27/0.70	.000/.000	[0.213; 0.345]
	Health	0.04/0.52	.697/.000	[−0.203; 0.228]
	Relational	0.20/0.64	.003/.000	[0.034; 0.311]
…ensures that I can keep my job in the long term.	Happiness	0.12/0.67	.002/.000	[0.043; 0.201]
	Health	0.27/0.75	.007/.000	[0.108; 0.505]
	Relational	0.06/0.64	.342/.000	[−0.070; 0.174]
Organizational support
…supports me when I have a problem.	Happiness	0.39/0.80	.000/.000	[0.299; 0.481]
	Health	0.36/0.81	.000/.000	[0.175; 0.520]
	Relational	0.77/0.97	.000/.000	[0.598; 0.975]
…often asked employees to participate in work-related decisions.	Happiness	0.62/0.89	.000/.000	[0.539; 0.697]
	Health	0.50/0.84	.000/.000	[0.339; 0.639]
	Relational	0.21/0.66	.179/.000	[−0.133; 0.478]
…enables me to combine my work with my private life very well (e.g. through flexible working hours, telework).	Happiness	0.19/0.70	.000/.000	[0.118; 0.269]
	Health	0.36/0.80	.000/.000	[0.237; 0.555]
	Relational	0.16/0.71	.093/.000	[−0.059; 0.333]
Voice
…encourages an open and comprehensive exchange of information between me and my supervisor.	Happiness	0.71/0.94	.000/.000	[0.631; 0.784]
	Health	0.82/0.97	.000/.000	[0.704; 0.944]
	Relational	0.13/0.62	.387/.000	[−0.193; 0.391]
…has formal grievance procedures to take care of employee complaints or appeals.	Happiness	0.28/0.71	.000/.000	[0.209; 0.360]
	Health	0.20/0.64	.003/.000	[0.053; 0.314]
	Relational	0.46/0.85	.000/.000	[0.326; 0.648]
…conducts regular employee surveys that give all team members a voice.	Happiness	0.17/0.73	.000/.000	[0.088; 0.249]
	Health	0.09/0.68	.126/.000	[−0.038; 0.199]
	Relational	0.29/0.78	.001/.000	[0.113; 0.464]
…has a works council or similar body that actively represents my interests.	Happiness	0.02/0.42	.629/.000	[−0.048; 0.081]
	Health	0.05/0.41	.387/.000	[−0.049; 0.159]
	Relational	0.39/0.77	.004/.000	[0.150; 0.681]
Well-being oriented HRM
Investing in employees	Happiness	0.08/0.74	.200/.000	[−0.041; 0.214]
	Health	−0.15/0.51	.165/.000	[−0.346; 0.06]
	Relational	0.49/0.79	.001/.000	[0.195; 0.772]
Providing engaging work	Happiness	0.45/0.94	.000/.000	[0.292; 0.609]
	Health	0.3/0.81	.022/.000	[0.052; 0.566]
	Relational	−0.39/0.24	.006/.058	[−0.655; −0.104]
Positive social and physical environment	Happiness	0.35/0.92	.000/.000	[0.171; 0.525]
	Health	0.77/0.95	.000/.000	[0.544; 0.991]
	Relational	0.36/0.66	.013/.000	[0.072; 0.63]
Organizational support	Happiness	0.25/0.88	.002/.000	[0.094; 0.42]
	Health	0.24/0.78	.069/.000	[−0.024; 0.49]
	Relational	0.22/0.68	.163/.000	[−0.082; 0.542]
Voice	Happiness	−0.04/0.79	.618/.000	[−0.207; 0.122]
	Health	−0.14/0.67	.286/.000	[−0.393; 0.105]
	Relational	0.41/0.78	.009/.000	[0.102; 0.716]

Table A2. Formative measurement models – moderation analysis.

Construct and indicators (My organization…)	Well-being dimension	Weights/loadings	p-values	95% CI weights
Investing in employees
…uses a rigorous selection process to select new recruits for a position similar to mine.	Happiness	0.17/0.63	.000/.000	[0.098; 0.243]
	Health	0.24/0.68	.000/.000	[0.160; 0.338]
	Relational	0.51/0.82	.029/.000	[0.163; 0.933]
…invests considerable time and money in my training and development.	Happiness	0.41/0.87	.000/.000	[0.328; 0.489]
	Health	0.39/0.87	.000/.000	[0.294; 0.492]
	Relational	0.12/0.74	.590/.000	[−0.490; 0.422]
… actively promotes my career (e.g. through career planning, mentoring).	Happiness	0.59/0.92	.000/.000	[0.515; 0.655]
	Health	0.55/0.90	.000/.000	[0.457; 0.630]
	Relational	0.57/0.86	.004/.000	[0.124; 0.974]
Providing engaging work
…allows me to make a lot of job decisions on my own.	Happiness	0.32/0.78	.000/.000	[0.254; 0.388]
	Health	0.33/0.78	.000/.000	[0.228; 0.431]
	Relational	0.07/0.58	.282/.000	[−0.067; 0.207]
…ensures that my job includes an appropriate level of challenge and variety to keep my work interesting.	Happiness	0.19/0.84	.000/.000	[0.122; 0.260]
	Health	0.09/0.81	.080/.000	[−0.028; 0.177]
	Relational	0.24/0.76	.000/.000	[0.131; 0.406]
…keeps me informed about business issues and about how well it’s doing.	Happiness	0.20/0.72	.000/.000	[0.140; 0.267]
	Health	0.17/0.69	.001/.000	[0.070; 0.267]
	Relational	0.34/0.78	.000/.000	[0.180; 0.471]
…provides meaningful feedback to me, regarding my performance at least once a year.	Happiness	0.21/0.70	.000/.000	[0.144; 0.280]
	Health	0.17/0.67	.005/.000	[0.039; 0.268]
	Relational	0.56/0.88	.000/.000	[0.435; 0.743]
…ensures that I can use my skills and abilities appropriately.	Happiness	0.34/0.87	.000/.000	[0.261; 0.409]
	Health	0.49/0.91	.000/.000	[0.372; 0.642]
	Relational	0.01/0.70	.885/.000	[−0.209; 0.158]
Positive social and physical environment
…does everything possible to prevent accidents and health hazards in the workplace.	Happiness	0.09/0.65	.003/.000	[0.032; 0.152]
	Health	0.20/0.72	.001/.000	[0.102; 0.328]
	Relational	0.05/0.59	.541/.000	[−0.124; 0.241]
…ensures that I am not overloaded by my work.	Happiness	0.24/0.76	.000/.000	[0.175; 0.299]
	Health	0.26/0.72	.001/.000	[0.151; 0.450]
	Relational	0.12/0.72	.219/.000	[−0.134; 0.254]
…offers practices and incentives to make me take more care of my health (e.g. sports courses, providing fruit).	Happiness	0.03/0.58	.258/.000	[−0.027; 0.094]
	Health	−0.15/0.43	.063/.000	[−0.311; −0.006]
	Relational	0.35/0.77	.003/.000	[0.184; 0.649]
…ensures that all employees have the same opportunities and that no one is disadvantaged.	Happiness	0.15/0.79	.000/.000	[0.083; 0.210]
	Health	0.13/0.77	.022/.000	[0.003; 0.237]
	Relational	0.19/0.77	.004/.000	[0.082; 0.340]
…does not tolerate any forms of bullying or discrimination at work.	Happiness	0.09/0.66	.015/.000	[0.015; 0.155]
	Health	0.20/0.75	.010/.000	[0.052; 0.353]
	Relational	0.01/0.56	.897/.000	[−0.174; 0.136]
…strongly encourages teamwork.	Happiness	0.36/0.84	.000/.000	[0.301; 0.431]
	Health	0.35/0.82	.000/.000	[0.250; 0.459]
	Relational	0.25/0.75	.000/.000	[0.099; 0.355]
…provides an average pay level (including bonuses) to our employees that is higher than that of our competitors.	Happiness	0.23/0.64	.000/.000	[0.173; 0.281]
	Health	0.04/0.47	.620/.000	[−0.130; 0.183]
	Relational	0.33/0.72	.000/.000	[0.206; 0.508]
…ensures that I can keep my job in the long term.	Happiness	0.18/0.65	.000/.000	[0.125; 0.244]
	Health	0.24/0.71	.001/.000	[0.113; 0.402]
	Relational	0.07/0.54	.257/.000	[−0.093; 0.173]
Organizational support
…supports me when I have a problem.	Happiness	0.39/0.77	.000/.000	[0.323; 0.452]
	Health	0.37/0.78	.000/.000	[0.260; 0.455]
	Relational	0.64/0.89	.000/.000	[0.439; 0.901]
…often asked employees to participate in work-related decisions.	Happiness	0.64/0.89	.000/.000	[0.578; 0.698]
	Health	0.53/0.83	.000/.000	[0.432; 0.624]
	Relational	0.49/0.80	.001/.000	[0.137; 0.709]
…enables me to combine my work with my private life very well (e.g. through flexible working hours, telework).	Happiness	0.21/0.63	.000/.000	[0.162; 0.268]
	Health	0.37/0.74	.000/.000	[0.288; 0.471]
	Relational	0.07/0.55	.366/.000	[−0.120; 0.170]
Voice
…encourages an open and comprehensive exchange of information between me and my supervisor.	Happiness	0.69/0.93	.000/.000	[0.626; 0.741]
	Health	0.81/0.97	.000/.000	[0.707; 0.905]
	Relational	0.29/0.72	.098/.000	[−0.149; 0.509]
…has formal grievance procedures to take care of employee complaints or appeals.	Happiness	0.30/0.69	.000/.000	[0.243; 0.359]
	Health	0.21/0.61	.000/.000	[0.117; 0.296]
	Relational	0.46/0.82	.000/.000	[0.338; 0.628]
…conducts regular employee surveys that give all team members a voice.	Happiness	0.23/0.71	.000/.000	[0.172; 0.290]
	Health	0.13/0.65	.008/.000	[0.034; 0.231]
	Relational	0.41/0.82	.000/.000	[0.286; 0.619]
…has a works council or similar body that actively represents my interests.	Happiness	−0.02/0.31	.362/.000	[−0.069; 0.027]
	Health	0.01/0.30	.731/.000	[−0.045; 0.076]
	Relational	0.16/0.53	.294/.000	[0.005; 0.685]
Well-being oriented HRM
Investing in employees	Happiness	−0.05/0.67	.305/.000	[−0.161; 0.045]
	Health	−0.12/0.52	.158/.000	[−0.294; 0.046]
	Relational	0.55/0.90	.003/.000	[0.212; 0.943]
Providing engaging work	Happiness	0.41/0.93	.000/.000	[0.281; 0.55]
	Health	0.36/0.85	.002/.000	[0.127; 0.59]
	Relational	0.02/0.74	.919/.000	[−0.4; 0.407]
Positive social and physical environment	Happiness	0.46/0.94	.000/.000	[0.307; 0.609]
	Health	0.75/0.97	.000/.000	[0.544; 0.951]
	Relational	0.42/0.84	.022/.000	[0.052; 0.751]
Organizational support	Happiness	0.29/0.89	.000/.000	[0.156; 0.414]
	Health	0.09/0.78	.431/.000	[−0.138; 0.307]
	Relational	−0.19/0.62	.323/.000	[−0.568; 0.2]
Voice	Happiness	−0.05/0.78	.469/.000	[−0.185; 0.076]
	Health	−0.06/0.71	.610/.000	[−0.266; 0.158]
	Relational	0.32/0.81	.084/.000	[−0.045; 0.682]

Table A3. Correlations – multi-group analysis.

	Mean	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1 Age	45.37	12.86	--
2 Branch	0.76	0.43	−0.07*	--
3 Care responsibilities	0.24	0.43	−0.10**	0.04	--
4 Change in telework intensity	0.76	0.43	0.09**	0.08*	−0.14**	--
5 Gender	0.52	0.50	−0.06**	−0.08*	−0.11**	0.04	--
6 Educational level	4.89	1.14	−0.02	−0.03	0.08*	0.03	−0.11**	--
7 Employment status	0.94	0.24	0.21**	0.03	0.01	0.10**	−0.06	0.08*	--
8 Leadership responsibilities	0.22	0.42	−0.02	0.07*	0.16**	−0.15**	−0.11**	0.15**	0.06*	--
9 Marital status	2.55	1.62	−0.16**	0.01	−0.18**	0.06	0.07*	−0.02	−0.08*	−0.10**	--
10 Prior experiences with telework	2.02	1.76	0.04	0.08*	0.16**	0.09**	−0.02	0.05	0.00	0.08*	−0.04	--
11 Telework intensity	4.39	2.34	0.03	0.06	0.39**	−0.05	−0.06*	0.11**	0.01	0.02	−0.03	0.37**	--
12 WBHRM	5.76	1.58	−0.01	0.06	0.10**	−0.06	−0.12**	0.06	0.05	0.19**	−0.07*	0.10**	0.21**	--
13 Strain	2.42	0.94	−0.13**	−0.02	0.11**	−0.07*	0.12**	0.01	−0.06	0.05	−0.04	0.02	−0.04	−0.29**	--
14 Social Isolation	2.80	1.18	−0.07*	0.00	0.17**	−0.17**	−0.04	0.09**	−0.05	0.11**	−0.01	0.01	0.31**	0.19**	0.06	--
15 Happiness well-being	3.12	0.86	0.08**	0.02	0.04	−0.03	−0.10**	−0.02	0.05	0.16**	−0.05	0.07*	0.10**	0.55**	−0.44**	0.04

* = correlation is significant at the 0.05 level;** = correlation is significant at the 0.01 level; n = 931.

Table A4. Correlations – moderation analysis.

	Mean	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14
1 Age	44.88	12.78	1
2 Branch	0.70	0.46	0.02	1
3 Care responsibilities	0.42	0.49	−0.12**	0.05	1
4 Change in telework intensity	0.57	0.50	0.11**	0.03	−0.11**	1
5 Gender	0.49	0.50	−0.05	−0.06**	−0.08**	0.05*	1
6 Educational level	4.99	1.15	−0.03	−0.04	0.04	0.05	−0.07**	1
7 Employment status	0.94	0.24	0.21**	0.04	0.00	0.10**	−0.04	0.08**	1
8 Leadership responsibilities	0.35	0.48	0.02	0.10**	0.25**	−0.10**	−0.13**	0.14**	0.10**	1
9 Marital status	2.35	1.58	−0.15**	−0.01	−0.19**	0.04	0.08**	−0.11**	−0.12**	−0.16**	1
10 Prior experiences with telework	2.02	1.56	0.03	0.07**	0.06**	0.12**	−0.02	−0.01	0.01	0.05*	−0.02	1
11 Telework intensity	4.52	1.41	0.03	0.12**	−0.06**	0.20**	0.03	0.03	0.01	−0.18**	0.05	0.27**	1
12 WBHRM	5.99	1.41	−0.01	0.08**	0.09**	0.01	−0.11**	0.06*	0.06*	0.23**	−0.08**	0.07**	0.02	1
13 Strain	2.42	0.88	−0.09**	0.01	0.10**	−0.07**	0.10**	−0.01	−0.04	0.03	−0.02	0.02	−0.02	−0.28**	1
14 Social Isolation	3.08	1.09	−0.02	−0.02	0.09**	−0.01	−0.04	0.06*	−0.03	0.07**	0.05*	−0.05	0.02	0.15**	0.10**	1
15 Happiness well-being	3.23	0.79	0.06**	0.02	0.06*	−0.02	−0.07**	0.00	0.05*	0.16**	−0.06*	0.04	−0.06*	0.51**	−0.43**	0.02

* = correlation is significant at the 0.05 level;** = correlation is significant at the 0.01 level; n = 1680.