1 Introduction

Psychosocial risks have important adverse consequences on the health of people, organisations and economies (Schaufeli & Enzmann, 1998). Employees are undergoing turbulent times at work not only because of the high levels of psychosocial stressors such as workload, role stress and workplace mistreatment but also because of the high job insecurity and burnout (Jiang & Probst, 2017; Karatepe et al., 2020). Paradoxically, we know more about the important benefits of occupational health and uch research points out the importance of unceasing organizational efforts to resolve this problem, including more integrated perspectives to understand employees’ experiences both in and outside of the work environment, as well as caring practices for employees, their families and organizations (Braganza et al., 2020; Tetrick & Winslow, 2015). From this perspective, it makes sense for the employees to identify proactive self-care behaviours, to effectively solve work problems and remain healthy, and to reduce the negative consequences in their personal lives. Workers also need to be physically and psychologically fit in order to maintain high levels of energy, effort, and engagement over time (Bakker, 2011; Casper & Sonnentag, 2020, Sonnentag, 2015) and to competently resolve daily work problems with a sense of commitment. However, little research has been carried out concerning the study of energy loss due to the combination of two relevant stressors in the field of burnout, such as role stress and work incivility, specifically taking into account personal resources and its mechanisms as workers’ recovery experiences, controlling as well the effect of optimism. Surprisingly, despite its benefits, optimism has received little attention in work settings. Therefore, the current study explores the relationships between role stress, work incivility and energy loss—including emotional exhaustion and affect—using a diary design. This study also examines the underlying mechanisms of the association between theses variables, incluiding personal resources (i.e., optimism and recovery). From this perspective, we study the specific effect of the analyzed variables, i.e. whether the relationship of role stress and work incivility differs when the moment of the day is controlled and its predictive effect depends on the criteria variable (i.e. exhaustion and negative and positive affect at night). In addition to this, considering both positive and negative criteria variables allow us to deepen in the knowledge of preventative programs. Thus, taking into consideration optimism and recovery as protective factors provide us insight into how these variables could help to deal with negative situations by re-establishing energy levels. On the contrary, this allow us to study whether the analyzed job stressors may have a profound impact on workers’ well-being. The knowledge about how role stress and work incivility affect at a daily level lead to determine how to prevent this within the workers’ routine. These facts also boost the development of changes that organizations and workers need to put in practice. Working with the optimism as a trained skill help to protect the workers’ well-being, facilitating healthy environments and positive expectations, similar to recovery experience and its role in the increase of energy levels. Hence, both constitute resources that could be used in a personal and organizational level and boost the understanding of how employees can become positive, energetic and engaged, even when facing undesirable work conditions.

2 Theoretical framework and hypotheses

Our study is built on the occupational well-being theory of the job demands–resources model (JD–R; Bakker & Demerouti 2007; 2017). According to this model, daily job stressors and workplace mistreatment (i.e., role stress and work incivility) explain the daily loss of energy through high exhaustion and negative affect, and low positive affect at night. Within the theories of organizational stress, exhaustion is described as a loss of energy and this loss is well-known derived from emotional process (Cooper, 1998). Exhaustion linked with negative emotions involve a loss of energy that could weaken the person. Such losses of energy are critical because they are associated with burnout (Maslach et al., 1997). Burnout is a syndrome consisting predominantly of chronic exhaustion and negative attitudes toward work and it predicts adverse functioning in the workplace (Bakker et al., 2014). Furthermore, personal resources are expected to buffer the undesirable impact of job demands on strain, and increase the desirable power of job demands on commitment. Research has provided only reduced support for this suggestion, which means that more research is needed to test the potential interacctions of personal resources (Bakker & Demerouti, 2016; Bakker & van Wingerden 2020). From this perspective, our proposal aims to study how this composition of high job demands and low personal resources (JD-R model; Xanthopoulou et al., 2009) constitutes a work environment of high stress, possibly leading to a daily emotional exhaustion and lack of energy. This is a process where the workers’ job stress may depend on the job demands and resources of the daily work routine (Bakker, 2014), being the work environment a critical factor to understand the workers’ well-being (Cooper, 1998). However, workers do not linearly or simply react to the work environment (Bolger & Zuckerman, 1995). In this sense, the characteristics of the work and how these impact on them could actively intervene, for example, through self-regulation strategies, recovery experiences and their own personality (Prem, 2017).

2.1 Role stress and work incivility

Role stress is one of the most important job stressors predicting job-related exhaustion and negative affect (Garrosa et al., 2011; Jin et al., 2020; Lee & Ashforth, 1996; Schmidt et al., 2014; Shaufeli & Enzmann, 1998). Thus, from the perspective of the JD-R theory, this stressor is considered a job demand that involves continued emotional and cognitive effort (Demerouti et al., 2001), and its associated with physiological and psychological costs (e.g., fatigue, depression), and increases the probability of burnout (Alarcon, 2011; Bakker & Schaufeli, 2000). Emotional exhaustion is often cited as a consequence of role stress and energy loss, thus experiences of stress finally diminish emotional resources (Posig & Kickul, 2003). In addition, role stress in women is related to discrepancy strain (Gillespie & Eisler, 1992), which explains the relationship of role congruity with beliefs about expectations for women (Eagly & Karau, 2002). As such, altogether the evidence suggests the likelihood of a direct relation between these variables. This hypothesis help to understand the role of time concerning the drawbacks of daily role stress in the workers’ well-being at bedtime. In this sense, we deepen understanding of how daily symptoms of exhaustion related to the daily role stress suffered could have a long-lasting and severe levels of exhaustion and energy loss through an accumulative process.

Hypothesis 1

Day-specific role stress will be positively related to: (H1a) exhaustion at bedtime and (H1b) negative affect (NA) at bedtime, and negatively related to (H1c) positive affect (PA) at bedtime.

Workplace mistreatment has received an increasing amount of attention from researchers (Aquino & Thau, 2009; Barling et al., 2009; Garrosa et al., 2015; Li et al., 2020; Yang et al., 2012). Workplace mistreatment comprises different forms, ranging from incivility to physical aggression. In this study, we are specifically interested in workplace incivility, a low-intensity mistreatment with an ambiguous intent to harm the target and violating norms for mutual respect at the workplace (Andersson & Pearson, 1999). Despite its vague intention, it has important consequences (e.g., burnout, dissatisfaction, lack of creativity and cooperation, psychological distress, negative affect) and, according to Cortina’s theory, it involves selective incivility toward women and could be considered a “modern” expression of sexism in the workplace (Cortina et al., 2011). From the perspective of JD-R theory, workplace incivility is a job demand, an interpersonal stressor that most likely demands and depletes resources and energy at work (Rhee et al., 2017). Considering the previous arguments, we predict that when employees experience daily work incivility, they are exhausted at bedtime. This hypothesis provides information on how workers would use their resources to face the threatening situations and defend them from work incivility (e.g., disrespect). This daily work incivility is associated with negative emotions, including as well the time effect to know more about this mechanism among these variables. Specifically, work incivility may hurt psychological needs like the need to be both respected and appreciated by others (Baumeister & Leary, 1995) and, consequently, initiate different emotional and affective responses with daily consequences in energy loss (i.e., high exhaustion, high NA, and low PA, everyday). Studies where short-term effects are considered can help us to understand that daily consequences are not small and provide information about how everyday incidents disturb us.

Hypothesis 2

Day-specific work incivility will be positively related to (H2a) exhaustion at bedtime and (H2b) NA at bedtime, and negatively related to (H2c) PA at bedtime.

In view of these undesirable job stressors, we wished to investigate mechanisms that assist workers’ self-care and self-defence from these toxic conditions.

2.2 Emotional wellness: Optimism and recovering

In this study, we focused on two variables that are well known to have a positive influence on emotional wellness: Optimism and recovery. According to JD-R theory and its further developments and applications, these personal resources provide resilience and protection, as they an positively moderate the relationship between the loss of energy, derived from daily job stressors, and negative consequences (Xanthopoulou et al., 2009; Kinnunen et al., 2011). As an example, Van Yperen and Snijders (2000) have observed the moderator role of self-efficacy and optimism in the relationship between job demands and psychological health symptoms. It seems that in situations of work overload and high job demands, personal resources diminish the workers’ negative perception of the stressors suffered due to the effort of work (Buruck et al., 2016; Salminen et al., 2014). These studies suggest that workers with high levels of personal resources posit a great domain over themselves, and in turn, this helps them to effectively handle hazardous working conditions such high workload and work incivility, avoiding the negative outcomes. Thus, personal resources such as optimism and recovery not only could contribute to achievi a positive environment, but also could determine the way that workers perceive, react and contribute to the work environment.

Optimism research reflects a broad view of self-regulation, planning and managing achievable goals, and positive expectancies when confronted with obstacles (Carver & Scheier, 2014; Neff & Geers, 2013) found that optimists engaged in more constructive problem solving than did pessimists. If a person, for instance, feels insecure and susceptible, problems are more difficult to resolve, and she or he may feel stress and disengagement (Carver & Scheier, 1998; Garrosa et al., 2011; Mappamiring & Putra, 2021). Additionally, optimists have a better profile of emotional responses to adversity and their coping responses are healthier. Interest in daily optimism at the workplace is increasing due to its positive relation with the dynamics of well-being (Casper et al., 2019; Littman-Ovadia & Nir, 2014; Sonnentag, 2015), however, there are few studies about it.

Therefore, optimism after work may also act as a buffer, reducing the consequences of daily job stresors, leading to protected levels of energy, and increasing PA, as we propose in the next hypotheses:

Hypothesis 3

Day-specific levels of optimism will be negatively related to (H3a) exhaustion at bedtime and (H3b) NA at bedtime, and positively related to (H3c) PA at bedtime.

Hypothesis 4

Optimism after working hours will moderate the within-person spillover of (H4a) exhaustion at bedtime and (H4b) NA at bedtime, and (H4c) it will increase PA at bedtime. The relation between job stressors and energy loss will be stronger when optimism is low than when optimims is high.

Lack of daily recovery leads to an accumulation of exhaustion and fatigue, and subsequently to chronic loss of energy and illness (Geurts & Sonnentag, 2006). In contrast, effective recovery increases energy, productive behaviours and well-being (Parker et al., 2020; Sonnentag, 2003; Xanthopoulou et al., 2014). Research about recovery from work can help to understand how employees can become energetic, engaged, and healthy, even when confronting poor work environments. In addition, recovery can moderate the daily effect of spillover from work to home (Sonnentag & Fritz, 2015; Trougakos et al., 2008). This mechanism refers to the process of decreasing or eliminating physical and psychological symptoms of strain produced by job demands and job stressors (Craig & Cooper, 1992; Meijman & Mulder, 1998), by distancing oneself from these negative emotions in one’s personal life. Therefore, this process is important to keep an inner balance and energy in other personal areas, and also to distancing oneself from a problem to better understand it and choose the best actions to resolve a toxic situation. Sonnentag and Fritz (2015) consider psychological detachment as a core recovery experience. For example, in Spanish workers, experiences of relaxation have produced important benefits in the intensification of positive emotions (Garrosa et al., 2013). This may occur through the prolonged activation of the functional system, neutralizing the effects of negative affect (Parker et al., 2020; Sonnentag & Fritz, 2007; Xanthopoulou et al., 2014). Thus, recovery is understood as a way of energy activation. In this vein, the workers have the needed energy to healthfully handle job demands (Quinn et al., 2012). This fact of workers feeling recovered and rested may lead them to be more proactive and adaptatively face job demands (Lazarus & Folkman, 1991). In other words, workers with high punctuation in recovery are likely to face job demands as their energy levels allow them to make an effort in solving job problems. Also, energy activation could amplify the thoughts repertoire with innovative and creative ideas (Quinn et al., 2012), resulting in a protector of the negative consequences of daily stress derived from high workload and work incivility. From the perspective of JD-R theory, recovery experiences constitute a personal resource by which the energy levels invested in the working day are recovered (Moreno-Jiménez et al., 2012). We propose the following hypotheses:

Hypothesis 5

Day-specific levels of psychological detachment after working hours will be negatively related to (H5a) exhaustion at bedtime and (H5b) NA at bedtime, and positively related to (H5c) PA at bedtime.

Hypothesis 6

Day-specific levels of relaxation after working hours will be negatively related to (H6a) exhaustion at bedtime and (H6b) NA at bedtime, and positively related to (H6c) PA at bedtime.

Hypothesis 7

Recovery from work will moderate the within-person spillover of (H7a) exhaustion at bedtime, (H7b) NA at bedtime, and (H7c) PA at bedtime. The relation between job stressors and energy loss will be stronger when recovery is low than when recovery is high.

To sum up, this study will contribute to the occupational health and personality literature in at least four ways. Firstly, by conducting a daily study, we will contribute to the burnout literature by shedding light on how employees through role stress and work incivility may affect personal resources and energy loss, with a day-to-day consequences. Secondly, since we utilize both an inter-individual and an intra-individual perspective, we will postulate insights on how frequency of emotional wellness actions impact the magnitude of change in levels of energy, thus increasing our understanding of how optimism and recovery (i.e., psychogical detachment and relaxation) fluctuate, and if these variables could diminish the daily consequences of role stress and work incivility. Research on flexibility patterns of different types of resources is scarce and critical to further increase our understanding of dynamic processes that occur at work (Casper et al., 2019; Salanova et al., 2010; Prem 2017). The introduction of this intra-individual approach and diary design in the field of occupational health allows us to go deeper into workers’ psychological processes (i.e. the sequence of their psychological states) and to identify antecedents and consequences of the central phenomena we are exploring, taking into account their temporal dimension (Navarro et al., 2015). It is important to note that when we approach the study of workers’ psychological processes, as is the case with this study, the emphasis has to be on states rather than traits. In this sense, diary studies are one of the most appropriate methodologies to assess those states and their relationships over time, and in the workers’ natural contexts (Navarro et al., 2015).

Thirdly, because we examine three variables for emotional wellness (i.e., optimism, psychological detachment and relaxation), we will elucidate their different processes in the workplace context. Fourthly, we expand the Job Demands-Resources theory (JD-R; Bakker & Demerouti, 2014; Bakker & Demerouti 2017) to develop a theoretical framework that allows us to explain how employees can improve daily energy at work through optimism and recovery, and to reduce job stressors. Figure 1 represents the proposed research model.

Fig. 1
figure 1

Energy loss after daily role stress and work incivility. The role of daily emotional wellness

3 Method

3.1 Procedure and sample

The participants of this study were employees in the service sector from a broad range of professional backgrounds, including trade retail, health, social work, consulting and education. We recruited study participants by contacting them through different associations from this sector in Madrid. After the associations had expressed interest in the study, information packages describing the study procedure were sent to the employees. As an incentive for participation, we offered feedback about the study results. Study participants filled in the general questionnaire and subsequently, they completed daily questionnaires three times a day, for five consecutive working days (Monday–Friday): (1) a morning item of sleep quality (to be completed after waking up in the morning, before going to work), (2) an afternoon questionnaire (to be completed after work), and (3) a night time questionnaire (to be completed before bedtime). Participants received clear instructions on how to complete the questionnaires at these three times, and the researchers emphasized the importance of following this procedure, encouraging the use of reminder alarms (e.g., watch and phone alarms) and maintaining regular contact (Green et al., 2006).

Of the 135 survey packages distributed, 121 were returned (86.6% response rate). In 4 of these diaries, the participants did not fill in all the days, so finally, 117 general questionnaires and diaries were used for the current study.

The final sample comprised 117 workers (74% female and 26% male). Mean age was 33.7 years (SD = 9.1), and mean professional experience was 18.0 years (SD = 39.6). Mean working time was 40.2 h per week (SD = 8.1), 51.3% reported having contact with people more than 80% of their working day, and 62% had a university degree.

3.2 Measures

We collected the data with the same general and daily pencil-and-paper questionnaires. Specifically, we assessed (1) daily sleep quality in the morning, (2) role stress and work incivility at work in the afternoon, and (3) optimism and recovery from work after working hours with the bedtime survey. Exhaustion, negative and positive affect were also reported at night. Daily measures of all the variables used modifications of items from the corresponding general-scale, which were reworded for daily administration. Moreover, for daily measures, we used the same response categories as for the general measure. This method of developing state-level analogues of general measures has been used successfully in the past (Blanco-Donoso et al., 2015; Nezlek, 2012; Parker et al., 2020). Table 1 shows means, standard deviations, zero-order correlations between study variables, as well as the means of Cronbach’s alphas averaged across the days of data collection.

Table 1 ICCs, means, standard deviations, Cronbach´s alphas and correlations (N = 117 individuals, N = 585)

3.2.1 Daily role stress at work

We measured role stress at work with the end-of workday survey, using scale items from the Role Stress Scale developed by Rizzo et al., (1970). Participants had to report how they had experienced role stress “today, during work” by responding to the items on 4-point Likert scales from 1 (not true at all) to 4 (totally true). In order to limit participant burden, we used 8 items to assess role stress (e.g.,‘‘I have had to work under vague directives or orders.”, “I think I have had many tasks in the same time.”).

3.2.2 Daily incivility at work

We used 6-items from the Workplace Incivility Scale (WIS; Cortina et al., 2001) to measure personal experiences of uncivil behavior. Participants responded on a 5-point scale from 1 (not true at all) to 5 (totally true), describing whether they had experienced each behavior by a coworker or supervisor (e.g., “Today you have been in situations in which one of your superiors or coworkers doubted your judgment on a matter over which you had responsibility. Gave you hostile looks, stares, or sneers”).

3.2.3 Optimism

We assessed optimism with the bedtime survey, using the 10-item Life Orientation Test- Revised (LOT-R; Scheier et al., 1994). This time, we instructed our participants to report how they felt at this moment. Each item is rated on a 7-point Likert scale ranging from 1 (not true at all) to 7 (totally true). A sample item is: “I don’t get upset too easily”.

3.2.4 Recovery from work

We measured psychological detachment and relaxation from work after working hours with the bedtime survey, using the 6-item Recovery Experience Questionnaire (Sonnentag & Fritz, 2007) in its Spanish version (Sanz-Vergel et al., 2010), rated on 8-point Likert items from 1 (not true at all) to 8 (totally true). A sample item is: “Tonight, I forgot about work”.

3.2.5 General and daily exhaustion

The Job-Related Exhaustion Scale (Wharton, 1993) provided a measure of emotional exhaustion, the primary dimension of job burnout (Lee & Ashforth, 1996; Wright & Cropanzano, 1998). The 6-item scale refers to the frequency of feeling “emotionally drained” and “burnt out”, and ranges from 1 (never felt this way) to 6 (felt this way every day). The scale was modified so that the items refer to nighttime. An example item is: ‘‘At this moment, I feel burnt out’’.

3.2.6 General and daily affect

Positive and negative affect were measured using the 10-item International Positive and Negative Affect Schedule Short-Form (I-PANAS-SF; Thompson 2007) with five items measuring PA and five items measuring NA. Participants rated the extent to which they generally experience each of the 10 emotions on a 5-point scale ranging from 1 (never) to 5 (always). The scale was modified so that PA and NA referred to nighttime. An example item is: ‘‘At this moment, I feel inspired’’.

For all these expected relationships, we took into account the potential effects of different control variables: Sex, the trait aspect of the corresponding criterion variable and sleep quality. Sleep quality was evaluated in the morning with the item “How do you evaluate your last night’s sleep?” derived from the Pittsburgh Sleep Quality Index (Buysse et al., 1989), rated on a 6-point Likert scale ranging from 1 (very bad) to 6 (very good). This measure correlates highly with the full Pittsburgh Sleep Quality Index (Sonnentag & Binnewies, 2013). Since working conditions are currently different depending on sex, it is necessary to control this variable, the general measure of the criterion variable is used to carry out multilevel analysis, and sleep quality is the typical control variable in burnout studies as it affects to daily energy levels.

3.3 Data analysis

Day-level data were nested within persons. We then analyzed the data with a random-coefficient modeling approach (Bryk & Raudenbush, 1992) using the software package MLwiN (Rasbash et al., 2000). To guarantee that the analyses reflected our concern about within-person variation and used unbiased estimates of the day-level variables (Enders & Tofighi, 2007), we centered day-level predictor and moderator variables at the person mean. We tested the proposed hypotheses in a sequence of multilevel models, with the first ones including the control variables (Model 1), then, the main effects of role stress and incivility at work (Model 2), then, the main effects of the moderator variables of optimism (Model 3) and recovery (Model 4), and then the interaction effects (Model 5). We compared these models with Chi-square tests, using the difference of the − 2 × log likelihoods of the nested models (with degrees of freedom corresponding to the number of new parameters added).

4 Results

4.1 Preliminary analyses

Prior to testing our hypotheses, we scanned the degree of within-person and between-person variations of our day-level variables. Table 1 shows the percentage of within-person variance (1–ICC = 1 − (between-person variance /[between-person variance + within-person variance]) derived from intercept-only models. All predictor variables showed an ICC above 25% (Hox & Roberts, 2011) with the exception of exhaustion at night (20%). Largely, these findings suggest that a substantial portion of the variance in our variables can be attributed to within-person variation across the five days. We have also considered exhaustion at night in a multilevel approach because, according to Nezlek (2008), even when the ICC approaches 0, multilevel modeling might still be essential, and this circumstance was coherent at the theoretical level.

4.2 Hypotheses testing

Exhaustion at bedtime.

For exhaustion at bedtime as an outcome measure (see Table 2), Model 1, with sleep quality, sex, and general exhaustion as control variables, improved the null model (difference − 2XLog = 98.58, df = 3, p < .001) with sex and general exhaustion as positive predictor variables (t = 2.23, p < .01, and t = 8.36, p < .001, respectively). By adding daily role stress and incivility at work (Model 2), the fit improved (difference − 2XLog = 28.38, df = 2, p < .01), with positive relationships with exhaustion at bedtime, role stress (t = 3.16, p < .01) and work incivility (t = 3.88, p < .01). Model 3, which included optimism, improved the fit, and showed a negative association with exhaustion at bedtime (t = -2.33, p < .01). Model 4 significantly improved the fit (difference − 2XLog = 36.24, df = 2, p < .001), also showing a negative relationship between relaxation and exhaustion at night (t = -4.46, p < .001). In Model 5, the effect of interaction term between role stress and psychological detachment increased the fit (t = -2.09, p < .05). To gain more insight into the nature of this interaction effect, we conducted simple slope tests, using the multilevel tool developed by Preacher et al., (2006). These simple slope tests showed that role stress was more positively related to exhaustion at night on days when people experienced less psychological detachment (γ = 0.457, SE = 0.118, z = 3.84, p < .001), whereas they were unrelated on days when psychological detachment was high (γ = -0.136, SE = 0.180, z = − 0.75, ns) (see Fig. 1). The hypothesis about the interaction between role stress and psychological detachment was also supported. In addition, when this interaction is significant, rol stress lost its significance as predictor.

Table 2 Multilevel estimates for models predicting exhaustion at night-time

NA at bedtime.

When NA at bedtime was included as an outcome variable (see Table 3), Model 1 showed a better fit than the null model, with sleep quality and general NA as predictors (t = -2.39, p < .05 and t = 2.94, p < .01, respectively). Model 2 improved the previous models, but role stress and work incivility had no significant influence on NA at bedtime. Model 3 increased the variance (-2XLog difference = 8.24, df = 1, p < .01) through optimism as a negative predictor of NA at night (t = -2.91, p < .01). Model 4 included recovery experiences, but there were no significant predictors. Finally, Model 5 increased the fit, with psychological detachment as a moderator of the relationship between work incivility and NA at night (t = -2.17, p < .01). Simple slope tests showed that workplace incivility was more positively related to NA at night on days when people experienced less psychological detachment (γ = 0.670, SE = 0.254, z = 2.64, p < .01), whereas they were unrelated on days when psychological detachment was high (γ = -0.362, SE = 0.242, z = -1.49, ns) (see Fig. 2). Overall, the results supported the hypothesis about the interaction, because psychological detachment is a negative predictor of NA at bedtime, with an interaction between incivility and psychological detachment.

Table 3 Multilevel estimates for models predicting NA at night-time
Fig. 2
figure 2

Interaction effects of role stress at work and psychological detachment on job related exhaustion at night-time

PA at bedtime.

Lastly, with PA at bedtime as an outcome measure (see Table 4), Model 1 fit the data better than the null model, and again, sex was a strong negative predictor of PA at bedtime (t =-3.17, p < .01), and general level of PA was a predictor of PA at bedtime (t = 2.16, p < .05). Role stress, work incivility, optimism included in Models 2 and 3 sequentially, were non-significant. Model 4 showed a highly significant improvement over the previous models (difference − 2XLog = 30.26, df = 2, p < .001), with relaxation as a robust predictor of PA at bedtime (t = 4.94, p < .001). Model 5 also increased the fit, and there was a significant effect of the interaction between work incivility and optimism (t = 2.00, p < .05). Simple slope tests (see Fig. 3) indicated that workplace incivility was more positively related to PA at night on days when people experienced more optimism (γ = 0.666, SE = 0.318, z = 2.09, p < .05), whereas they were unrelated on days when optimism was low (γ = -0.573, SE = 0.341, z = -1.67, ns). These results only supported the moderator effect between optimism and work incivility.

Table 4 Multilevel estimates for models predicting PA at night-time
Fig. 3
figure 3

Interaction effects of incivility at work and psychological detachment on NA at night-time

Fig. 4
figure 4

Interaction effects of incivility at work and optimism on PA at night-time

5 Discussion

Our day-level study showed that predictor variables that are implicated in daily energy have different components in relation to variability, affect symmetry that involve the reactions to positive events and experiences are reflected mainly in positive affective states and reactions to negative events and experiences are reflected mainly in negative affective states (Sonnentag, 2015) and the involved mechanism. These findings show that exhaustion is more stable and does not undergo so many fluctuations at the daily level as NA and PA. Daily exhaustion depends to a great extent on the general levels of chronic exhaustion. Exhausted workers are probably already burnt out (Bakker et al., 2014), disengaged and are lacking energy (Casper & Sonnentag, 2020; Demerouti et al., 2001). Our study provides new insights, showing that daily work incivility presented a significant effect in daily exhaustion, even more than role stress, and could have negative consequences for workers’ emotional health. An intensification of physiological dysregulation is likely to accelerate the development of disease (McEwen & Wingfield, 2003), especially when workers are already exhausted in a general way and the effect of work incivility adds up (e.g., disrespect among other elements), producing consequences for daily exhaustion at night time. Exhaustion fluctuates within shorter periods of time in the week and can increase over longer periods of time. In this sense, this accumulative increase of exhaustion may clarify the relation between burnout and depressive symptoms and life dissatisfaction (Hakanen & Schaufeli, 2012).

Exhaustion compounds sad affect and, according to Teasdale’s (1988) work, these feelings can lead to a self-defeating cycle of escalating negative thoughts and emotions, which, in turn, can increase the risk of an eventual major depressive episode (Beevers & Carver, 2003; Lau et al., 2004). It has been confirmed that burnout predicted antidepressant treatment, with a stronger association in men than women (Madsen et al., 2015). This study extends earlier burnout research and shows that, to prevent exhaustion at the day level, zero tolerance of work incivility and respect for both the self and others in everyday tasks are neededas, once workers experience loss of energy, the recovery is complex. Daily work incivility activates a stress reaction (exhaustion at night-time) that remains persistent, even after working hours have ended and the initial stressor is no longer present (Sonnentag, 2015).

Concerning the included variables, only daily relaxation and optimism contribute to the mitigation of this adverse cycle in workers’ everyday life, for example, by facilitating positive expectation, effective mood repair, adaptive self-regulation strategies, as well as enhanced psychological functioning (Bakker & van Wingerden, 2020; Carver & Connor-Smith, 2010; Casper et al., 2019; McHale et al., 2015; Scheier & Carver, 1992). Relaxation facilitates recovery because it limits the prolonged activation of the functional system and counteracts the effects of emotional exhaustion (Parker et al., 2020; Sonnentag & Fritz, 2007). These results contribute to the notions that asymmetry is not persistent, and that personal resources can also contribute to offset the negative strain process (Sonnentag, 2015). Thus, daily emotional wellness had direct effects on daily exhaustion at bedtime, specifically with optimism and relaxation. Also, a moderator effect of recovery through detachment in the relationship between role stress and exhaustion at bedtime was found.

Unexpected results revealed that job stressors had no direct effect on NA and PA at bedtime. According to Friedman and Kern (2014), NA would be more related to individual dispositions. These results, especially those non-significant, support that emotional exhaustion is a consequence more related to stressful job situations such as work incivility, whereas NA and PA are more related to workers’ capacity of self-regulation and their recovery possibilities. Our findings support this idea and extend the power of daily optimism to its negative association with NA at night. An explanation for this could be that optimists work determinedly to counteract these negative moods and achieve mood repair, which contributes to enriched emotional functioning and prevents burnout (Garrosa et al., 2011; Mappamiring & Putra, 2021; McHale et al., 2015). This study extends previous research showing that these mechanisms occur at the daily level and are essential to maintain energy during the working week and not allow NA to affect one’s personal life. Furthermore, daily optimism is more likely to develop positive evaluations, such as the belief that their problems will have solution, which in turn reduce the impact of NA.

Regarding PA, we observed a major effect of symmetry due to the positive valence of the predictor variables and a valuable relationship with proactive behaviors such as relaxation, which has also been shown to predict an increase in work engagement (Simbula et al., 2011; Sonnentag, 2003). Furthermore, our results show that relaxation is an important resistance factor against energy loss. However, the result of no direct relationship between optimism and PA was unexpected. This may be partially due to the multilevel study methodology employed, in which multicollinearity is diminished. However, optimism resulted in being a positive predictor in the interaction with work incivility, observing its protective effect as an emotional equilibrium. This protective effect maintaining PA could be explained because optimistic workers see negative events (i.e., work incivility) as not being their fault. They also see them as being flukes that have nothing to do with other areas of their lives or future events, linked with the use of better regulation strategies and effective coping skills against job stressors.

The moderating effect of psychological detachment as a recovery experience from adverse conditions is confirmed, as it buffers the negative effects of role stress in the face of exhaustion, and of incivility and NA. However, it had no effect on the increase of energy through PA. On the days when workers did not achieve psychological detachment, role stress was more positively related to exhaustion at night. Equivalently, workplace incivility was more strongly associated with NA at night on days when people experienced less psychological detachment. These results support the stressor-detachment model (Sonnentag, 2010), which states that psychological detachment from work during nonworking time is not only a direct predictor of increased well-being and energy but could also buffer the negative impact of job stressors (Sonnentag & Fritz, 2015). Thus, our study extends previous research by revealing that the effect of psychological detachment can moderate the daily effect of role stress and work incivility. This may occur through an attentional process, for example, by focusing thoughts toward other more beneficial life areas which in turn increases self-confidence to cope with job stressors and decreases negative thoughts during nonworking time. This is particularly important in the face of work incivility, such as low-intensity mistreatment, which undeniably has consequences on self-esteem (Andersson & Pearson, 1999) and which may be a precursor of other forms of aggression of greater intensity, such as bullying (Cortina & Magley, 2003). By disconnecting, a worker’s self-esteem does not suffer as much, and he or she may be capable to stop being mistreated at the workplace.

This study provides more evidence about relaxation as a positive experience at the emotional and energy level. In our study, relaxation had a main effect, contributing to avoiding exhaustion. After general exhaustion, relaxation is the most significant predictor, as well as the most significant at the level of direct effects of PA at night, but it does not have moderator effects. Our study shows that psychological detachment and relaxation are predictors of daily energy, even after controlling for daily optimism, but they play different roles in this process because they intervene with different mechanisms. Relaxation would have a direct effect of reducing activation and increasing PA, perhaps contributing to, or ameliorating physiological dysregulation and thereby playing a key role in exacerbating or counteracting allostatic load/overload (McEwen, 2007). Therefore, processes that reduce this prolonged activation are crucial in order to restore an organism’s pre-stress state, whereas psychological detachment would be a buffer. Relaxation probably generates the prior emotional state necessary to distance oneself psychologically, thereby contributing to breaking ruminative thoughts associated with depression (Nolen-Hoeksema et al., 1994). Some people have trouble disconnecting, probably because it requires a prior state of calmness and vitality that can be achieved through relaxation. This process would facilitate distancing and could be considered a precedent in future studies. It would be interesting to explore this relationship to include practical resources to help workers to distance themselves from their work in stress prevention programs.

Sex is a significant predictor of exhaustion, and PA. Our results show higher scores on exhaustion and lower PA scores for female workers at the daily level. This is in line with results found in similar studies (Marinaccio et al., 2013). Future studies should investigate whether the cause is women’ worse working conditions and/or that more difficulties to practice recovery. In addition to this, gender perspective – based program for women would take a relevant place in order to prevent these work stress-related issues. In this sense, the promotion of optimism as self-efficacy beliefs is related to important cognitions that may protect women from negative outcomes of job stress (Brown & Harris, 2012; Mappamiring & Putra, 2021). Women with a positive viewpoint have less stress, better creative problem-solving skills, and better health outcomes than less optimistic women. Similarly, the uses of time and work-life balance for women still have more barriers (Clercq & Brieger, 2021). In the same vein, poor quality sleep also explains NA, because sleep helps to preserve emotional well-being (Sonnentag & Binnewies, 2013) and could promote physiological balance and resilience (McEwen, 2006). In short, enhancing women’s skills to better deal with time to boost recovery and co-responsibility in men are necessary aspects.

Finally, daily optimism has been shown to be an important variable as a driver of energy, both at the level of main effect, decreasing exhaustion and NA, as well as moderating the relation between work incivility and PA. It has been observed that workers with high optimism maintain a high level of PA at nighttime, even with high work incivility. Nevertheless, it would be interesting to confirm the long-term effect of this interaction and determine whether it is related to actions such as denouncing these toxic situations at work, or whether the prolonged strain could end up exhausting the worker. Ultimately, whether it is an expression of active resource to achieving positive outcomes in the face of adversity, like coping mechanisms that operate at the time of trauma (Karatsoreos & McEwen, 2011). Concretely, studies have supported optimism’s protective mechanisms against stress and its associations with skills to connect and obtain social support in stressful situations (Andersson, 2012). Thus, social support can both help define the extent to which the event is appraised as stressful and mitigate negative outcomes. We understand that feeling socially supported can protect individuals’ well-being from the stress associated with incivility.

In conclusion, this study contributes to the research on burnout and recovery, revealing the daily and undesirable loss of energy in workers due to work incivility and role stress, perhaps with chronic consequences. Conversely, the combination of optimism, relaxation and psychological detachment can contribute to reducing stressful circumstances and to improving daily energy. Likewise, the moderator resources help to reduce the consequences of negative emotions, daily spillover from work to home. For example, daily psychological detachment helps in the process of decreasing exhaustion and NA related to work incivility and role stress. This role is relevant even after controlling for optimism. Daily optimism could also increase energy through the increase in PA. In sum, these mechanisms are relevant to maintain an adequate energy balance and more healthy lifestyles in our personal life, and to distance oneself from work stress and improve active responses to adverse situations, such as workplace mistreatment and role stress, thereby avoiding energy loss, which predicts depressive symptoms and life dissatisfaction.

5.1 Limitations and implications for future research

The current study has some limitations: Firstly, we assessed all the data with self-reports because they are suitable to assess private experiences (Conway & Lance, 2010), but they involve the problem of common-method variance. However, collecting data with general and daily surveys and using person-centered scores in the analyses ensured the measurement of predictor and criterion variables at different time points each day (morning, afternoon, and night) and included control variables according to previous research. Although the temporality dimensions we focused in this article are limited, they provide us a valuable field to explicitly consider the role of time in work settings and health of workers. Future research should collect an in-depth analysis of time more extensively with the incorporation of the multiple dimensions of temporality (George & Jones, 2000), data from several sources (i.e., colleagues, family, and clients), more objective variables (e.g., biomarkers) and the extension of research findings from a study conducted on a sample to the population at large. Secondly, we used survey packages, following previous daily studies, which have the advantage of being low-cost, easy to administer, and being psychometrically equivalent to data obtained through electronic assessment, but they also have some disadvantages. For this reason, the participants received clear instructions to complete the questionnaire in the specified three moments and, for each moment, they had to indicate the exact time they finished. They were also encouraged to use reminder alarms (e.g., watch and phone alarms) and to maintain regular contact (Green et al., 2006). Although all participants reported a completion time within the corresponding hours, future research could use more creative procedures and a longer assessment period (i.e., two weeks or more) to determine how consistent these processes are over time (Schwarz, 2012). Thirdly, we explored the moderating role of daily optimism and recovery. Future research should also explore mediation within the multilevel context, for example, the sequential relation between relaxation, psychological detachment, and energy. Finally, like most researchers, we assumed that the relationships between the within-person variables are linear. However, many psychological variables do not follow a linear pattern (Burke et al., 2007). It is possible, for example, that the effect of work incivility is nonlinear because different uncivil behaviors are not equivalent. Further research on modeling nonlinear effects is necessary (Deboeck, 2011; Ladstätter et al., 2015).

5.2 Practical implications

Primary interventions may be more operative and effective in the workplace because, when loss of energy occurs, there are negative consequences on daily well-being, which can become chronic. In this case, recapturing initial motivation is more difficult. Therapeutic assistance may be required in this situation, revealing unsuccessful prevention of psychosocial risks. In addition to creating a healthy workplace with a suitable adaptation between demands and resources in the workers and the organization, a clear commitment to ethical values ​​(i.e., zero tolerance of workplace mistreatment and discrimination) and supportive relationships in organizations will help workers to feel healthier (Morrison & Cooper-Thomas, 2013). The implementation of workplace wellness programs including learning skills to develop psychological detachment, personal resources such as optimism, and the use of relaxation techniques could be a self-care practice. Such a wellness program should necessarily include a better adaptation of working time to contribute to an optimal balance of work and life. Organizations can also generate principles for respectful and caring interpersonal interactions, growth in cooperating work and collective goals among employees, and build social support and commitment.