When considering recovery as the process of restoring employees' resources (Zijlstra and Sonnentag, 2006; Sonnentag and Fritz, 2007), the importance of work-to-family balance for the recovery process seems crucial. Until now, most findings have focused on the positive outcomes of daily recovery experiences for employees' well-being instead of their antecedents (Blanco-Donoso et al., 2017; Wendsche and Lohmann-Haislah, 2017; Bennett et al., 2018). Additionally, experimental studies have been conducted to test the affective consequences of daily recovery experiences (Sonnentag and Niessen, 2020). Currently, new insights suggest the pertinence of exploring daily fluctuations in the antecedents of recovery experiences due to emerging trends in the work sphere (Rodríguez-Muñoz et al., 2018; Chawla et al., 2020; Pfaffinger et al., 2020). A meta-analysis of antecedents and outcomes of detachment from work found that this recovery experience is influenced by work-related and personal characteristics (Wendsche and Lohmann-Haislah, 2017). Moreover, work events and physiological indicators of parasympathetic regulation have been identified as important antecedents of off-the-job relaxation (Parker et al., 2020).

The W-HR model (ten Brummelhuis and Bakker, 2012) may help explain how short-term work-home processes of day-level WFF can benefit recovery processes on the daily basis. Day-level WFF can be described as a process of resource accumulation by means of work resources that can increase personal resources and where those personal resources, in turn, can be utilized to improve home system functioning, such as through recovery experiences and being healthy. Thus, day-level WFF occurs when the developmental, affective, capital and efficiency gains acquired in the work system are transferred to and enhance the functioning of the family system, with day-level WFF being positively related to the day-to-day recovery process (Wayne et al., 2007). Based on the W-HR model (ten Brummelhuis and Bakker, 2012), the present study examines the beneficial effects of day-level WFF on recovery experiences, suggesting that this positive process may operate within a relatively short time frame on a daily basis, fostering subsequent general health at night. We examine whether employees experience a higher level of detachment from work and relaxation during off-job time in the evening and increase their level of general health at night on days involving higher levels of WFF. Thus, we propose the following (please, see Figure 1):

Work engagement is a well-documented predictor of important outcomes at the employee, team, and organizational levels (Bakker and Albrecht, 2018). Currently, it is assumed that work engagement fluctuates among employees over time (Sonnentag et al., 2010b). For example, daily fluctuations are found in workers who are engaged during workdays when they have reached a high level of recovery during evenings (Sonnentag, 2003). In fact, a reciprocal relationship has been demonstrated between daily work engagement and daily recovery (Bakker, 2014). This means that on days on which employees recover well, they are also more engaged, and on days on which employees feel engaged during the day, they experience subsequent daily recovery (Sonnentag et al., 2012). Daily work engagement refers to daily levels of vigor, dedication, and absorption that may fluctuate as a function of daily demands, resources, and proactive behaviors (Bakker and Albrecht, 2018). From this perspective, it can reasonably be expected that daily work engagement promotes recovery experiences and fosters general health as a consequence. On the basis of the W-HR model (ten Brummelhuis and Bakker, 2012), gain spirals of daily work engagement may explain how on days on which employees are engaged in the work sphere, they can use the rest of their energy during off-job time through recovery experiences, transferring engagement to participation in the home sphere. In this sense, we expect to find positive direct effects of daily work engagement on the level of recovery during leisure time and on general health at night. Thus, we propose the following:

Empirical research has reported the moderating role of work engagement in recovery, for example, in buffering the negative effect of workload on psychological detachment from work (Clauss et al., 2020). However, little is known about the adequate level of daily work engagement that moderates relationships, resulting in positive outcomes such as recovery and general health (Blanco-Donoso et al., 2021). An emerging line of research has introduced the debate on a “dark side of engagement” in the literature, which posits that more work engagement is not necessarily preferable (Bakker and Leiter, 2010; George, 2010; Bakker et al., 2011; Leiter, 2019). A study supporting this idea reveals adverse effects of high work engagement on psychological distress in the short term, showing a curvilinear relationship between work engagement and psychological distress (Shimazu et al., 2018). Hence, high levels of daily work engagement might have detrimental effects on individuals' well-being.

Drawing on the W-HR model (ten Brummelhuis and Bakker, 2012), short-term work engagement may reflect daily processes by means of employees who spend their volatile contextual resources in the work sphere (e.g., being highly engaged at work) and may affect daily outcomes in the home domain through a change in volatile personal resources (e.g., for recovering well during off-job time; Blanco-Donoso et al., 2021). To allow employees to use their gains from work to recover during off-job time in the evenings and even experience higher daily levels of general health at night, they might require a low level of daily work engagement during the workday instead of a high level, as it may drain their resources. Based on this idea, we examine the moderating role of daily work engagement and propose the following hypothesis:

Detachment from work has been considered a mediator of the relations between employees' responses to stressful work-related experiences and health (Sonnentag and Fritz, 2015). Beyond physical distance to the workplace, detachment from work refers to a withdrawal from work not only physically but also psychologically. Thus, psychological processes are key to detachment from work, implying not thinking about work or doing work-related activities outside of the workplace (Sonnentag and Fritz, 2007). The benefits of detachment from work go further than the individual emotional outcomes experienced on the same day at night (e.g., Garrosa et al., 2015) and affect the next day's level of well-being. In particular, several studies have shown the lagged effects of detachment from work in the following morning. For example, the level of work detachment in the evening predicts low levels of negative affect and fatigue (Sonnentag et al., 2008), high levels of vigor (ten Brummelhuis and Bakker, 2012), and feelings of recovery (Volman et al., 2013), and a lack of psychological detachment has been related to near accidents when commuting to work the next morning (Pereira et al., 2016).

Despite previous evidence, it is still unknown whether the lagged effect of detachment from work in the evening on the level of well-being persists at the end of the following day. Following the W-HR model (ten Brummelhuis and Bakker, 2012), the daily level of detachment from work during off-job time may reflect short-term processes of gaining spirals within the home sphere to maintain optimal general health. Day-level detachment may act as a daily outcome that operates on the next day. Thus, on days on which employees have more gains and reach high levels of detachment from work during off-job time, they are also more likely to experience better general health due to the daily process of restoring resources, with benefits persisting to the next day. Considering the daily process, the present study examines whether detachment from work during off-job time in the evening on day 1 predicts employees' general health the following day at night. Thus, we propose the following hypothesis:

We collected data from 104 Chilean employees from a variety of industries, including education (26%), administration (19.2%), services (21.2%), health (6.7%), banking (1.9%), and Information Technology and Communications (1.9%), among others (23.1%). Data were collected before the COVID-19 outbreak. The sample was composed of 43.3% men and 56.7% women with a mean age of 42.3 (SD = 16.5) of a range of ages including 20–30 (32%), 31–40 (20%), 41–50 (23%), and 51–60 (25%). With respect to family status, 47.1% of the participants were married or living with a partner, and 35.6% were single. Among the participants, 63.4% had children (16.3% had one child, 26% had two children, 16.3% had three children, and 4.8% had four children). Regarding education level, 37.1% held a university degree, and 36.5% held a technical degree. Most of the participants worked for small and medium enterprises (54.8%), and the mean job tenure was 18.0 years (SD = 26.7). Most of the participants had a permanent contract (63.5%) and a full-time schedule (78.8%) and worked 44 or 45 h per week (46.2%). A snowball sampling technique was used to recruit participants from several local organizations with the following inclusion criterion: working with a timetable of 5 consecutive days during the work week. We based this criterion on methodology recommendations made by Ohly et al. (2010) for diary studies in organizational research. Following similar diary studies evaluating employees, we selected serial temporary data for 5 working days to capture an entire weekly work cycle (i.e., Sanz-Vergel et al., 2010a; Garrosa et al., 2015). After the participants expressed their willingness to voluntarily participate in the study, the researchers gave them and explained a paper-based package that included (a) a letter describing the objective of the study, instructions on the completion of the surveys, and simple questions used to create a personal code to preserve anonymity (i.e., the year of birth); (b) two informed consent forms, with one copy for researchers and another copy for the participants; and (c) the general and daily paper-based questionnaires. The participants did not receive a reward for taking part in the study. Ethical permission was obtained and approved by the Research and Ethics Committee at Temuco Catholic University (Chile). Participants were instructed to complete the general questionnaire (Level 2, Person-level) before starting on the daily questionnaires (Level 1, Day-level) for 5 consecutive working days twice a day (N = 520 observations).

To capture daily dynamic processes (Bolger et al., 2003), the researchers designed diary entry intervals to better understand work and non-work periods of the participants' schedules. Specifically, the participants completed Monday through Friday daily work engagement in the afternoon, immediately after returning home from work and referring to the workday; day-level WFF at night refers to home activities performed in the evening; day-level detachment from work and relaxation at night refers to off-job time in the evening; and general health at night was measured before bedtime. To ensure that the participants appropriately completed the questionnaires, the researchers underlined the importance of following the procedure, and the participants were also given a reminder to make daily entries. The time taken to complete the daily surveys was also recorded. The participants completed the afternoon form at 18:18 p.m. on average (SD = 43 min) and the nighttime form at 23:04 p.m. on average (SD = 1 h and 11 min).

We used a general questionnaire to collect data at the individual level (Level 2, general) and daily questionnaires at the day level (Level 1, daily). As a previous step before administering the general and daily questionnaires, a group of academicians selected the Chilean versions of the scales and adapted the language of those for which there were no versions. Based on guidelines provided by Hambleton and De Jong (2003), all of the scales were revised by academicians focused on the Chilean context and on psychology and were further analyzed through group discussions to test their adequacy for studying Chilean employees (Willgerodt, 2003; Vogt et al., 2004). Then, experts reviewed the appropriateness of the items' use over the established timeline according to the sample characteristics. At the day level (Level 1), daily measures of predictor and criterion variables were modified from the corresponding general scale to the specific diary entries made at night (Ohly et al., 2010; Nezlek, 2012). Following previous studies, responses measured on 5-point Likert scales of 1 = I fully disagree to 5 = I fully agree were given for all items (Rodríguez-Muñoz et al., 2012).

We used hierarchically structured data with repeated measures at the day level nested among individuals at the individual level simultaneously (Hox, 2010). The model has two levels: repeated measures at the day level (Level 1: within-person variation; N = 520 diary entries) and participants at general Level 2 (between-person variation; N = 104). To analyze the two-level dataset, we conducted a multilevel analysis using a hierarchical linear modeling approach (Nezlek, 2012). Based on Ohly et al. (2010), we centered the predictor daily variables of Level 1 (day-level WFF and daily work engagement) using the person-mean centering method, which considers the mean scores of each participant. The grand-mean centering method, which considers the mean scores of all participants, was used for the general or trait criterion variables and for predictor variables of Level 2. This procedure eliminates all of the variance between subjects so that it does not influence the interpretation of the results. Data were analyzed using MLwiN software (Leckie and Charlton, 2013).

Before testing our hypotheses, we examined within- and between-person variation in our variables. When calculating the intraclass coefficient for the outcome variables, we found that within-person variance reached 66.5% for detachment from work, 64.5% for relaxation, and 71.1% for general health. The intraclass coefficient of the predictor variables was measured at 67.8% for day-level WFF and at 77.7% for daily work engagement. All predictor and criterion variables showed an intraclass correlation coefficient of above 25% (Hox and Roberts, 2011). Overall, these findings suggest that a substantial portion of the variance in our outcome variables can be attributed to within-person variation across the 5 days, which supports the use of multilevel analysis (Fisher and To, 2012). Table 1 shows the means, standard deviations, reliability of day-level and general measures, and bivariate correlations among the study variables.

To test our hypotheses, we compared four nested models that included the specific general or trait (Level 2 person-level: between-person variation) and daily (Level 1 day-level: within-person variation) variables. The null model included only the intercept. In Model 1, we entered control variables at the individual level for Level 2 (e.g., general or trait criterion variables and day). In Model 2, we added daily WFF (Level 1) for testing H1. In Model 3, we included daily work engagement (Level 1) to test H2. In Model 4, we included the interaction term of the daily level of WFF (Level 1) as a predictor variable, and daily work engagement (Level 1) as a moderator for testing the moderation hypothesis (H3). In addition, we applied two nested hierarchical linear models to test the lagged effect hypothesis (H4). In this case, we included the null model with the intercept. In Model 1, we included employees' general health or health traits and day person-level variables (Level 2), and in Model 3, we added detachment from work as a predictor day-level variable (Level 1). To assess the improvement of each model over the previous one, we explored the differences between the respective likelihood ratios. Tables 2–5 present model fit information (difference of −2 x Log) estimates for the fixed parameters and estimates for the variance components. As a measure of effect size, we computed pseudo-R² following the recommendations of Singer and Willett (2003). The pseudo-R² statistic is used to quantify the incremental variance in the dependent variable predicted by adding a new set of predictors to a given model.

For day-level detachment during off-job time in the evening as an outcome variable (see Table 2), Model 2 provided the best model fit with the positive role of day-level WFF (B = 0.13, SE = 0.05; t = 2.61, p < 0.01) contributing to an increased model fit (difference of −2 × log = 6.65, df = 1, p < 0.01), in line with Hypothesis 1a. However, contrary to Hypothesis 2a, daily work engagement did not have a significant positive relation with detachment from work at night (B = 0.14, SE = 0.08; t = 1.71, ns), but Model 3 was not found to be significant (difference of −2 × log = 2.93, df = 1, ns). Model 4 fit the data significantly better than Model 3 (difference of −2 × log = 6.04, df = 1, p < 0.05) with the positive relation for day-level WFF (B = 0.10, SE = 0.05; t = 2.05, p < 0.05) and the interaction effect on detachment from work during off-job time in the evening. The interaction term between day-level WFF and daily work engagement was significant (B = −0.36, SE = 0.14; t = −2.47, p < 0.01).

To obtain more insight into the role of this interaction, we performed simple slope tests (Preacher et al., 2006). As Figure 2 illustrates, congruent with Hypothesis 3a, the relationship between day-level WFF and detachment from work during off-job time in the evening was stronger when the employee's level of daily work engagement was lower (y = −0.629; SE = 0.301; z = −2.089; p < 0.05) rather than higher. Thus, on days on which employees experienced low levels of daily work engagement, day-level WFF was strongly related to detachment from work during off-job time in the evening as opposed to on those days with high levels of daily work engagement. Moreover, surpassing our expectations, simple slope tests showed a significant result for a high slope of daily work engagement (y = −0.629; SE = 0.301; z = −2.089; p < 0.05). This result means that on days on which employees experience high levels of daily work engagement, the daily level of WFF was weakly related to detachment from work during off-job time in the evening as opposed to on those days with low levels of daily work engagement at work. With regard to pseudo-R², all predictor and control variables entered in the models explained 61.9% of the variance at Level 1 [0.346 – (0.334/0.346) = 0.619] and 14.4% of the variance at Level 2 [0.686 – (0.372/0.686) = 0.144].

For daily relaxation during off-job time in the evening as an outcome variable (see Table 3), congruent with Hypothesis 1b, Model 2 showed the best model fit with the positive relation for daily WFF (B = 0.17, SE = 0.05; t = 3.33, p < 0.001) contributing to an increased model fit (difference of −2 × log = 11.00, df = 1, p < 0.05). In Model 3, daily work engagement did not have a significant positive relation with relaxation during off-job time in the evening (B = 0.14, SE = 0.08; t = 1.64, ns), contrary to Hypothesis 2b, but the model was not found to be significant (difference of −2 × log = 2.70, df = 1, ns). Model 4 showed a further improvement over Model 3, with daily WFF (B = 0.14, SE = 0.05; t = 2.71, p < 0.05) and the interaction term (B = −0.45, SE = 0.15; t = −3.05, p < 0.01) was significant.

Simple slope tests for the interaction (Preacher et al., 2006) show that congruent with Hypothesis 3b (see Figure 3), the relationship between daily WFF and relaxation during off-job time in the evening was stronger when the employee's daily work engagement was lower (y = −0.789; SE =0.308; z = −2.56; p < 0.01) rather than higher. Thus, on days on which employees experienced low levels of daily work engagement, the daily WFF was strongly related to relaxation during off-job time in the evening as opposed to on those days with high levels of daily work engagement. Supporting our expectations, simple slope tests show a significant result for a high level of daily work engagement (y = −2.092; SE = 0.724; z = −2.88; p < 0.01). This result shows that on days on which employees experienced high levels of daily work engagement, daily WFF was weakly related to relaxation during off-job time in the evening as opposed to on those days with low levels of daily work engagement. With regard to pseudo-R², all predictor and control variables entered in the models explained 58.7% of the variance at Level 1 [0.369 – (0.353/0.369) = 0.587] and 4% of the variance at Level 2 [0.665 – (0.416/0.665) = 0.040].

For daily general health at night (see Table 4), Model 2 was not found to be significant (difference of −2 × log = 3.34, df = 1, ns), and Hypothesis 1c is not supported. Daily WFF did not have a significant positive relation with general health at night (B = 0.06, SE = 0.03; t = 1.83, ns). However, Model 3 provided the best model fit with the positive relation for daily work engagement (B = 0.30, SE = 0.05; t = 6.36, p < 0.001) contributing to an increased model fit (difference of −2 × log = 37.91, df = 1, p < 0.001), in line with Hypothesis 2c. Model 4 fit the data significantly (difference of −2 × log = 7.61, df = 1, p <0.01) with the positive relation for daily work engagement (B = 0.30, SE = 0.05; t = 6.36, p < 0.001) and the interaction effect between daily WFF and work engagement on general health at night (see Figure 4) (B = −0.46, SE = 0.08; t = −5.57, p < 0.001).

Following Preacher et al. (2006) and congruent with Hypothesis 3c (see Figure 4), simple slope tests for the interaction show that the relationship between daily WFF and general health at night was stronger when the employee's daily work engagement was lower (y = −0.939; SE =0.175; z = −5.36; p < 0.001) rather than higher. Thus, on days on which employees experienced low levels of daily work engagement, daily WFF was strongly related to general health at night as opposed to on those days with high levels of daily work engagement. Additionally, simple slope tests show a significant result for a high level of daily work engagement at work (y = −2.270; SE = 0.414; z = −5.47; p < 0.001), which exceeds our expectations. This means that on days on which employees experienced high levels of daily work engagement, daily WFF was weakly related to general health at night as opposed to on those days with low levels of daily work engagement. With regard to pseudo-R², all predictor and control variables entered in the models explained 74.7% of the variance at Level 1 [0.127 – (0.111/0.127) = 0.747] and 7.5% of the variance at Level 2 [0.312 – (0.121/0.312) = 0.075].

For predicting day-level general health at night on the following day (see Table 5), according to Hypothesis 4, day-level detachment from work during the previous evening during off-job time (day 1) showed a positive relationship to next-day general health at night (B = 0.07, SE = 0.03; t = 2.18, p < 0.05), and Model 2 also provided a significant model fit (difference of −2 × log = 4.72, df = 1, p < 0.05). With regard to pseudo-R², all predictor and control variables entered in the models explained 84.6% of the variance at Level 1 [0.130 – (0.127/0.130) = 0.846] and 6.1% of the variance at Level 2 [0.311 – (0.116/0.311) = 0.061].

We used the W-HR model (ten Brummelhuis and Bakker, 2012) to explain the daily process of WFF—recovery—general health among Chilean employees. The aim of the present study was threefold. Our first goal was to test whether the beneficial effect of daily WFF and daily work engagement impacts recovery experiences during off-job time in the evening and subsequently employees' general health at night. Multilevel analyses show positive direct effects of daily WFF on detachment from work and relaxation but not on employees' general health at night. Our results extend previous research on the positive synergistic potential of work to home, which highlights the positive view of daily WFF as an antecedent of recovery experiences. Most research focuses on the positive side of WFF in terms of well-being outcomes (Wendsche and Lohmann-Haislah, 2017; Bennett et al., 2018; Carlson et al., 2019; Jiao and Lee, 2020), while less attention has been given to exploring the positive aspects of WFF as an antecedent of recovery. Our results reveal that WFF is crucial to the day-to-day recovery process, implying that employees who achieve a positive transference of system functioning from work to the home domain may experience a daily process of restoring resources during off-job time. According to the W-HR model (ten Brummelhuis and Bakker, 2012), daily WFF as part of a process of resource accumulation contributes to individual psychological well-being. Thus, our findings support this model by means of the intrapersonal process of recovery, which is influenced by the positive short-term gain spirals of daily WFF.

Unexpectedly, our findings do not support our postulates related to the W-HR model, and we found that daily WFF does not have a significant positive relationship with employees' general health at night, even though it is considered a precursor of health and well-being (Jones et al., 2020), while it is significant in predicting recovery. In contrast, daily work engagement is a precursor of employees' general health at the end of the day but does not predict recovery. These results run contrary to previous evidence showing a reciprocal relation between daily work engagement and recovery (Bakker, 2014). Surprisingly, these findings could indicate the presence of different daily mechanisms depending on the nature of psychological constructs involved. It seems that the positive interaction between work and the family sphere by means of the work sphere facilitates system functioning at home and improves aspects in terms of performance. That is, people can feel better by engaging in activities with a partner, family or friends; achieving better performance in domestic tasks; or duly fulfilling family responsibilities because they have acquired the required skills or competencies in the work sphere, allowing them to recover well. In contrast, the beneficial effects of daily WFF might not be related to general health because they could depend on other diary aspects. Examples may include other contextual variables of work that have not been measured or personal variables such as emotional regulation and personality characteristics. Thus, in line with the W-HR model developed by ten Brummelhuis and Bakker (2012), contextual work resources seem to be relevant to gaining short-term personal resource processes.

Our second goal was to examine the moderating role of daily work engagement in the relationship between recovery experiences and employees' general health. The results indicate that on days on which employees are low in daily work engagement, the daily level of WFF is strongly related to recovery and general health. Our results also exceed expectations by showing that on days on which employees are high in daily work engagement, the daily level of WFF is weakly related to recovery and general health. Curiously, these results seem to be in line with the debate on a “dark side of engagement,” which posits that employees could ascribe excessive importance to their work that impacts their lives (Bakker and Leiter, 2010; George, 2010; Bakker et al., 2011; Leiter, 2019; Blanco-Donoso et al., 2021). In the case of our study, excessive daily work engagement might be an obstacle to recovering and experiencing general health at the end of the day. What our results seem to indicate is that when employees have already earned positive transference from work competencies to implementation in home-related tasks, it might not be necessary to be highly engaged during the workday. Thus, in accordance with emerging evidence, daily work engagement may have curvilinear—instead of linear—effects on recovery and mental health (Shimazu et al., 2018).

Finally, our third goal was to shed light on the lagged effect of psychological detachment from work as a core recovery experience. Multilevel results show that its positive impact on employees persists to next-day general health at night. Employees who detached from work during off-job time the previous evening experienced better general health at bedtime on the following day. Most research on psychological detachment from work has focused on its mediational role (Sonnentag et al., 2010a; Sonnentag and Fritz, 2015; Sonnentag and Lischetzke, 2018). Despite the relevance of this knowledge for theory and practice, our study suggests the relevance of exploring lagged effects that might have an association with an individual's well-being. Until now, few studies have explored the lagged effect of detachment from work at the daily level, and our results not only are congruent with previous evidence on its benefits the following morning (Sonnentag et al., 2008; ten Brummelhuis and Bakker, 2012; Volman et al., 2013; Pereira et al., 2016) but also go further by measuring the effect on the following day at bedtime.

The importance of the beneficial effects of day-level WFF on daily recovery experiences is supported by this study. Theoretically, this study is based on the W-HR model focusing on the positive synergistic potential of the work–family interface. Due to the benefits of daily WFF for recovery experiences, employees should be informed about the potential positive effect of daily processes on their lives to learn how to regulate their work and home spheres and to become aware of the importance of actively seeking recovery experiences. Following experts' recommendations, inclusive organizations should provide employees with resources to effectively perform in the private sphere, fostering daily recovery from work stress (Martinez-Corts and Demerouti, 2017). Thus, we encourage organizations to evaluate psychosocial risks using specific measures [for example, see Wood et al. (2019)] and to even expand training and coaching programs, for example, by integrating a preventive module on healthy work-home balance techniques. In such training [for example, see Bisschoff et al. (2019) and Wu and Chang (2020)], employees should be asked to apply gains acquired at work to the home sphere to be more efficient during off-job time and recover well. Employees can offset the gains of the work domain (e.g., management strategies to very high job demands) by using opportunities for recovery during daily off-job time.

According to our findings, in such training, employees should also be urged to implement strategies promoting psychological detachment from work during their off-job time. For example, a recent intervention revealed that employees with more leader-member exchange and a higher need for recovery benefit most from learning strategies for detaching at work (Clauss et al., 2018). Furthermore, organizations can adopt mindfulness training at the team and individual levels to regulate employees' levels of work engagement because its positive influence may be helpful in affecting the curvilinear trend of work engagement (Liu et al., 2020). Doing so in an early stage might have beneficial effects on employees' general health and in managing healthy organizations. These practical implications are particularly beneficial for employees who have to deal with emerging trends in current job-related demands, such as telepressure and technostress, which make work–life balance and employee well-being more difficult to achieve (Barber et al., 2019; Pfaffinger et al., 2020).

This research has several limitations. The first limitation stems from the self-report measures used. These measures relied on the sincerity of the participants' responses and may have been subject to social desirability bias. Although researchers provided each of the participants with an envelope to maintain their privacy and confidentiality, in future studies, it might help to incorporate the use of online questionnaires instead of paper surveys and envelopes to improve the anonymity of data collection. Another limitation relates to the variety of occupations our participants were engaged in (e.g., services, administration, banking, Information Technology and Communications, and education). The participants were recruited via snowball sampling, but other sampling techniques could be used to render sociodemographic characteristics uniform in future research. In addition, while women and men differ in their experiences of combined conflict and facilitation balance (Boz et al., 2016), this study did not adopt a gendered perspective. Future studies should differentially analyze the daily process of WFF—recovery—general health for female and male Chilean employees to better understand the specific relationships involved, as well as considering socio-demographic variables such the age of employees' children. Moreover, considering the relevance of specific recovery experiences, future studies should also focus on other types of recovery experiences, such as mastery. The present study involved collecting data before the COVID-19 outbreak. Future research could further pursue daily process exploration to understand how pandemic measures affect working modes (e.g., smart working and the closure of schools with subsequent effects for working parents), which may have an impact on the beneficial effects of daily WFF.