Detach to Thrive: Psychological Detachment from Work and Employee Well-Being
- Open Access
- 01.04.2025
- Research Paper
Erschienen in:
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by (Link öffnet in neuem Fenster)
Abstract
1 Introduction
Nowadays, the focus of employees in Western societies has shifted from financial survival to the search for happiness at work. For organizations, this implies that employee well-being is a critical issue in the war for talent. On the one hand, employee well-being affects several organizational outcomes, such as employee performance (Wright & Cropanzano, 2000) and turnover (Wright & Bonett, 2007). On the other hand, employee well-being also affects employee health and life expectancy (Howell et al., 2007). Moreover, ensuring good health and well-being was set as a Sustainable Development Goal in the Agenda 2030 by the United Nations. Accordingly, organizations and policy makers may seek to enhance employee well-being. For this purpose, it is crucial to identify the key drivers of well-being.
We study whether and how psychological detachment from work translates into employee well-being. Psychological detachment from work refers to not thinking about work during leisure time (Sonnentag & Fritz, 2007). This implies not only physically but also mentally distancing oneself from work (Etzion et al., 1998). If detachment enhances well-being, then employers as well as employee representatives could take measures to foster detachment in order to improve employee well-being. This is not only crucial for employers, but also for employees and society as a whole.
Anzeige
Our study contributes to the literature in many ways. Previous research often explored the relation between psychological detachment and indicators of impaired well-being. These indicators include burnout, health complaints, depressive symptoms, need for recovery, and emotional exhaustion (e.g., Fritz et al., 2010b; Santuzzi & Barber, 2018; Sonnentag & Fritz, 2007). We add to the literature on the negative relation between detachment and “dark” outcomes by analyzing sadness, worry, and anger as dependent variables. Apart from this, we explore the “bright” side of consequences, such as feeling happy, life domain satisfactions, and global life satisfaction. We are the first to investigate the effect of psychological detachment on job satisfaction, a critical proxy used for on-the-job utility (Cornelissen et al., 2011). In addition, this is the first study to compare the effect during the Covid-19 pandemic with the effect in the pre-pandemic time. This is also the first study that explores the relevance of detachment for well-being among various employee subgroups, investigating whether the effect is universal.
Importantly, previous studies on the relation between detachment and well-being are mostly based on small cross-sectional datasets (e.g., Hamilton Skurak et al., 2021; Kinnunen et al., 2010; Sonnentag & Fritz, 2007), homogenous occupational cohorts (e.g., Dettmers, 2017; Fritz et al., 2010a; Kühnel et al., 2009), or obtained inconsistent results (e.g., Cheng & McCarthy, 2013; de Bloom et al., 2013; Flaxman et al., 2012). We use a unique and representative large-scale panel dataset from Germany to close the gaps in the literature, namely the German Socio-Economic Panel (GSOEP). The GSOEP was developed specifically for the purpose of investigating well-being and is therefore a suitable data basis for this study (Goebel et al., 2019). We use all subjective well-being measures available in the GSOEP in the relevant survey waves, thereby capturing the multi-dimensional nature of subjective well-being (Diener et al., 1999). Our broad set of well-being indicators covers affective (pleasant and unpleasant feelings) and cognitive (life satisfaction, job satisfaction, and satisfaction with other life domains) well-being. These indicators are established in the literature on subjective well-being (Diener, 1984; Diener et al., 1999). Using multiple well-being measures allows us to analyze whether psychological detachment affects these different variables to a similar extent, i.e., to test whether the effect is evident across different well-being dimensions. Additionally, the dataset contains a heterogeneous sample of employees from various occupations. Therefore, we can explore whether the influence of detachment on well-being is universal or whether it depends on employees’ characteristics. Finally, the uniqueness of the dataset allows us to move closer towards causality by using individual fixed-effects panel estimations.
The remainder of this paper is structured as follows. The next section includes the theoretical background and relevant previous empirical work. Section 3 describes the data, variables, and method. Section 4 presents the results, robustness checks, limitations, and directions for future research. Finally, the findings are discussed and concluded in Sect. 5.
2 Theoretical Background and Previous Empirical Work
Etzion et al., (1998, p. 579) introduced the term psychological detachment from work as an “individual’s sense of being away from the work situation.” They revealed that it was not the rest itself, but rather the psychological distance from work during the rest period that was important for recovery. Following this pioneering work, psychological detachment has been established as a core mechanism contributing to employee recovery (Sonnentag & Fritz, 2007).
Anzeige
We analyze the impact of psychological detachment on employees’ subjective well-being. Well-being is defined as preference realization in the literature (Schimmack, 2009). Researchers often consider subjective well-being as multi-dimensional, distinguishing affective and cognitive dimensions of well-being (Diener, 1984; Diener et al., 1999). Affective well-being includes emotional responses (pleasant and unpleasant feelings) while cognitive well-being covers life domain satisfactions and global judgements of life satisfaction (Diener et al., 1999). Hedonic measures of feelings can indicate preference realization, because affective responses depend on individuals’ preferences. A more direct approach of measuring preference realization is to ask individuals to evaluate their life or life domains based on their personal preferences, for example assessing their satisfaction with life and various life domains (Schimmack, 2009). Overall, we follow the conceptualization of subjective well-being given by Diener et al., (1999, p. 277): “Subjective well-being is a broad category of phenomena that includes people’s emotional responses, domain satisfactions, and global judgments of life satisfaction.” Thus, Diener et al. (1999) categorize the components of well-being into four dimensions: pleasant affect, unpleasant affect, life satisfaction, and domain satisfactions. We investigate the relation between psychological detachment and all four components in our study.
We use two complementary theories for explaining how detachment may affect well-being. The effort-recovery model (Meijman & Mulder, 1998) proposes that employees mobilize their resources for goal-attainment at work and these resources are depleted throughout the workday (Quinn et al., 2012). Psychological detachment from work is a recovery experience which can help restore depleted resources (Newman et al., 2014). Detachment implies not thinking about work during leisure time (Sonnentag & Fritz, 2007). As a result, job demands cease to impact individuals such that recovery can occur. This helps individuals unwind and rebuild resources (Sonnentag et al., 2010). Accordingly, the effort-recovery model describes recovery as a process during which an individual’s strain level returns to a baseline level through relief from job demands (Meijman & Mulder, 1998). On the contrary, a lack of detachment inhibits relief from job demands and impairs recovery.
The stressor-detachment model (Sonnentag & Fritz, 2015) extends the effort-recovery model by proposing that psychological detachment from work is crucial for employee well-being (Santuzzi & Barber, 2018). According to the stressor-detachment model, psychological detachment plays a mediating role in the relation between work stressors and well-being. The model suggests that detachment enhances employee well-being by buffering the negative effects of job stressors. Thus, those who psychologically detach from work during non-work hours have better recovery, reduced stress, and lower exhaustion levels, which in turn can be translated into better well-being.1 In line with these theoretical underpinnings, we hypothesize that psychological detachment from work is positively associated with employee well-being. Additionally, certain dimensions of well-being may exhibit varying levels of resilience to the effects of low psychological detachment (e.g., Wettstein et al., 2022). Thus, we also explore whether the impact is consistent across our 12 well-being indicators, thereby assessing the generalizability of the effect.
A growing body of literature already examines the association between psychological detachment and different well-being indicators. However, these studies have many shortcomings. First, previous studies mostly focus on the association between psychological detachment and indicators of impaired well-being such as burnout, health complaints, depressive symptoms, need for recovery, and emotional exhaustion (among others, see Fritz et al., 2010b; Santuzzi & Barber, 2018; Sonnentag & Fritz, 2007). Second, most studies investigate a limited number of well-being indicators (among others, see Sonnentag & Bayer, 2005; Sonnentag & Fritz, 2007). Thus, these studies do not analyze whether psychological detachment affects multiple well-being outcomes to a similar extent, i.e., whether the effect is evident across different well-being dimensions. Third, most studies are based on small cross-sectional datasets and do not provide causal interpretations (among others, see Burke et al., 2009; de Jonge et al., 2012; Donahue et al., 2012; Fritz et al., 2010b; Moreno-Jimenez et al., 2009; Shimazu et al., 2012; Siltaloppi et al., 2009). Fourth, while some studies use longitudinal instead of cross-sectional datasets, their observations are based on very small samples of employees (e.g., Feuerhahn et al., 2014; Korunka et al., 2012; Sonnentag & Bayer, 2005). Fifth, most studies focus only on specific homogenous occupational cohorts such as managers (Burke et al., 2009; Hahn & Dormann, 2013), service workers (de Jonge et al., 2012), nurses (Donahue et al., 2012; Kühnel et al., 2009; Ten Brummelhuis & Bakker, 2012), teachers (Cropley & Millward Purvis, 2003; Fritz et al., 2010a), railway controllers (Korunka et al., 2012), public-service employees (Sonnentag et al., 2008), etc. Finally, while most studies show a positive association between detachment and well-being, there also exist studies reporting an insignificant (e.g., de Jonge et al., 2012; Hahn et al., 2012; Querstret & Cropley, 2012) or even a negative relationship (Shimazu et al., 2012) between detachment and their well-being indicators. Thus, Sonnentag and Fritz (2015, p. S85) stress that the “findings regarding well-being indicators seem less consistent.”
Our study contributes to the literature on detachment and well-being by addressing all these shortcomings. First, we add to the literature on the negative association between detachment and “dark” outcomes by analyzing sadness, worry, and anger as outcome variables. We additionally study “bright” outcomes, such as feeling happy, life domain satisfactions, job satisfaction, and global life satisfaction. Second, we use a unique, representative, and large-scale longitudinal dataset from Germany to examine the association between detachment and 12 different well-being indicators. Considering the multi-dimensional nature of well-being (Diener et al., 1999; Wettstein et al., 2022), we utilize all available well-being measures in our dataset. The representative dataset consists of employees with different occupations and different socio-demographic characteristics from all over Germany. Moreover, the longitudinal nature of the dataset allows the use of individual fixed-effects panel estimations bringing us a step closer to causality. The dataset even allows us to compare the effect of detachment before and during the 2020 global pandemic on these 12 well-being indicators. Finally, we also investigate whether the impact of detachment on well-being is universal or depends on employees’ characteristics.
3 Data, Variables, and Method
3.1 Data and Sample
The empirical analysis is based on data from the GSOEP, an annual population representative survey of about 30,000 individuals (Goebel et al., 2019). The dataset is highly suitable for the purpose of this study for several reasons. First, the GSOEP is a well-established and widely utilized dataset in well-being research.2 The GSOEP contains multiple well-being measures, including indicators of affective and cognitive well-being. Second, previous studies are often based on homogeneous occupational cohorts (e.g., Dettmers, 2017; Fritz et al., 2010a), while our dataset offers a large sample from various occupational groups. Third, the GSOEP provides a rich set of individual- and job-related controls. Fourth, its panel structure allows for longitudinal statistical analyses.
While the well-being variables were included in every survey year, psychological detachment was assessed less regularly. In the main analysis, we therefore use the survey waves 2011, 2016, and 2021. Our sample consists of employees aged between 18 and 67 years who work full- or part-time.3 We exclude apprentices and marginally employed individuals as work is not their primary purpose. Thus, we also obtain a more homogeneous sample. Moreover, as our estimations are based on a fixed-effects model, we exclude singleton observations. After excluding individuals with only one observation as well as observations with missing information, the analysis uses 12,045 observations from 5,511 employees.
3.2 Variables
3.2.1 Affective Employee Well-Being
Our dependent variables capture multiple dimensions of subjective well-being that are established in the literature (Diener, 1984; Diener et al., 1999).4 Affective employee well-being was assessed in the GSOEP through the following questions: “For each of the following feelings, please state how often you experienced this feeling in the last four weeks. How often have you felt…—angry?—worried?—happy?—sad?”5 Answers were given on a five-point Likert scale including “very rarely”, “rarely”, “occasionally”, “often”, and “very often.” Thus, we use four distinct measures of affective employee well-being: anger, worry, happiness, and sadness. We also create a measure of affect balance by subtracting the average of the three negative items (angry, worried, sad) from the positive item (happy) (Schimmack, 2009).
3.2.2 Cognitive Employee Well-Being
Satisfaction with different areas of life was assessed in the GSOEP. Individuals were asked: “How satisfied are you today with the following areas of your life? How satisfied are you with…—your health?—your sleep?—your job?—your leisure time?—your family life?—your social life?” The answer categories ranged from 0 “completely dissatisfied” to 10 “completely satisfied.” We focus on these six domain satisfactions because only these may reasonably be related to detachment. In addition, we use the average of these domain satisfactions as a rough measure of overall satisfaction (Schimmack, 2009). Apart from domain satisfactions, cognitive well-being also includes global judgements of life satisfaction. This was measured in the GSOEP through the following item: “Now we would like to ask you about your satisfaction with your life in general. How satisfied are you with your life, all things considered?” Possible answers ranged from 0 “completely dissatisfied” to 10 “completely satisfied.” Table 1 provides the definitions and descriptive statistics of the dependent variables.
Table 1
Descriptive statistics of outcome variables
Variable | Definition | Mean | Std. Dev | |
|---|---|---|---|---|
Affective well-being | Angry | Score of feeling angry ranging from 1 “very rarely” to 5 “very often.” | 2.82 | 0.98 |
Worried | Score of feeling worried ranging from 1 “very rarely” to 5 “very often.” | 1.88 | 0.90 | |
Happy | Score of feeling happy ranging from 1 “very rarely” to 5 “very often.” | 3.64 | 0.79 | |
Sad | Score of feeling sad ranging from 1 “very rarely” to 5 “very often.” | 2.23 | 0.96 | |
Affect balance (AB) | Score of employee affect balance constructed from subtracting the average of the three negative items (angry, worried, sad) from the positive item (happy) | 1.33 | 1.26 | |
Cognitive well-being | Health satisfaction (HS) | The overall health satisfaction scored on an eleven-point Likert scale ranging from 0 “completely dissatisfied” to 10 “completely satisfied.” | 7.02 | 1.86 |
Sleep satisfaction (SS) | The overall sleep satisfaction scored on an eleven-point Likert scale ranging from 0 “completely dissatisfied” to 10 “completely satisfied.” | 6.92 | 2.02 | |
Job satisfaction (JS) | The overall job satisfaction scored on an eleven-point Likert scale ranging from 0 “completely dissatisfied” to 10 “completely satisfied.” | 7.17 | 1.83 | |
Leisure time satisfaction (LTS) | The overall leisure time satisfaction scored on an eleven-point Likert scale ranging from 0 “completely dissatisfied” to 10 “completely satisfied.” | 6.78 | 1.98 | |
Family life satisfaction (FLS) | The overall family life satisfaction scored on an eleven-point Likert scale ranging from 0 “completely dissatisfied” to 10 “completely satisfied.” | 7.91 | 1.74 | |
Average domain satisfaction (DS) | The average of six domain satisfactions (health, sleep, job, leisure time, family life, social life) measured on an eleven-point Likert scale ranging from 0 “completely dissatisfied” to 10 “completely satisfied.” | 7.16 | 1.30 | |
Life satisfaction (LS) | The global overall life satisfaction scored on an eleven-point Likert scale ranging from 0 “completely dissatisfied” to 10 “completely satisfied.” | 7.44 | 1.48 |
3.2.3 Psychological Detachment from Work
Detachment is the main explanatory variable in this study. Two items were used to measure the extent to which employees think about job-related matters in their free time: (1) “When I come home, it is very easy to switch off from thinking about work,” and (2) “Work seldom lets go of me, it stays in my head all evening.” Each item is measured on a four-point Likert scale ranging from 1 “strongly disagree” to 4 “strongly agree.” The intercorrelation of the two items is suitably high with a Cronbach’s alpha of 0.67. We create a single score of psychological detachment from work by dividing the sum of the two items by 2. Responses to the second item were recoded in inverse order before adding up. Overall, psychological detachment increases with higher numbers, i.e., fewer work-related thoughts during leisure time.6
3.2.4 Control Variables
The GSOEP provides a rich set of individual- and job-related controls. We control for age, years of education, marital status, number of children in the household, number of persons in the household, and region of residence to account for the socio-demographic background of employees. Moreover, job-related factors are kept constant by including variables for monthly gross income, working hours, fixed-term contract, tenure, public sector, perceived job insecurity, firm size, part-time contract, industry, and occupation. Finally, the year of observation is controlled for. Our choice of the control variables largely follows the existing literature on the relation between detachment and well-being (e.g., Fritz et al., 2010b; Sonnentag & Fritz, 2007) and takes determinants of well-being into account to obtain more precise estimates (e.g., Cornelissen, 2009; Gülal & Ayaita, 2020). Table 2 shows the definitions and descriptive statistics of the explanatory variables.7
Table 2
Descriptive statistics of explanatory variables
Variable | Definition | Mean | Std. Dev |
|---|---|---|---|
Psychological detachment | Score of psychological detachment from work constructed from adding up two items measured on a four-point Likert scale ranging from 1 “strongly disagree” to 4 “strongly agree.” The sum of items is divided by 2. The items are: “When I come home, it is very easy to switch off from thinking about work”, and “Work seldom lets go of me, it stays in my head all evening.” The second item was recoded in inverse order before adding up | 2.84 | 0.79 |
Age | The employee’s age by years ranging from 18 to 67 | 46.54 | 9.49 |
Years of education | The employee’s years of education ranging from 7 to 18 years | 13.20 | 2.72 |
Married | Dummy equals 1 if the employee is married | 0.67 | – |
Number of children in HH | The number of children in the household | 0.74 | 0.99 |
Size of HH | The number of persons in the household | 2.87 | 1.28 |
East Germany | Dummy equals 1 if the employee resides in one of the federal states located in East Germany (Berlin, Brandenburg, Mecklenburg-West Pomerania, Saxony, Saxony-Anhalt, Thuringia) | 0.26 | – |
Southern West Germany | Dummy equals 1 if the worker resides in one of the Southern federal states located in West Germany (Bavaria, Baden-Wuerttemberg) | 0.28 | – |
Northern West Germany | Dummy equals 1 if the worker resides in one of the Northern federal states located in West Germany (Schleswig–Holstein, Hamburg, Lower Saxony, Bremen) | 0.15 | – |
Log of gross income | Natural log of gross income received last month | 7.94 | 0.61 |
Working hours | The number of weekly hours the worker actually works including possible over-time | 38.75 | 10.18 |
Fixed-term contract | Dummy equals 1 if the employee holds a fixed-term contract | 0.07 | – |
Tenure | The number of years the employee is with their current firm | 13.24 | 10.51 |
Public sector | Dummy equals 1 if the employee is employed in the public sector | 0.31 | – |
Job insecurity | Dummy equals 1 if the employee is somewhat concerned or very concerned about his or her job security | 0.35 | – |
Firm size: 20–199 | Dummy equals 1 if the employee is employed in a firm with 20–199 employees | 0.27 | – |
Firm size: 200–1999 | Dummy equals 1 if the employee is employed in a firm with 200–1999 employees | 0.25 | – |
Firm size: 2000+ | Dummy equals 1 if the employee is employed in a firm with more than 2000 employees | 0.32 | – |
Part-time contract | Dummy equals 1 if the employee is employed part-time | 0.27 | – |
Industry dummies | Six broad industry dummies for manufacturing, construction, trade, transport, banking/insurance and services (reference group: agriculture, energy and mining) | ||
Occupation dummies | Eleven broad occupation dummies for semi-skilled blue-collar, skilled blue-collar, foreman/forewoman, semi-skilled white-collar, skilled white-collar, highly-skilled white-collar, white-collar with extensive managerial duties, middle-level civil servant, upper-level civil servant, executive-level civil servant, and self-employed (reference group: unskilled blue-collar, master craftsperson, unskilled white-collar, lower-level civil servant) | ||
Year dummies | Two dummies for the years 2016 and 2021 (reference year: 2011) | ||
3.3 Method
Our estimations are based on the fixed-effects regression approach making use of our longitudinal dataset (see Wooldridge, 2013, p. 484). This method has the advantage of controlling for unobserved time-invariant employee characteristics, and consequently, mitigates the endogeneity issue. The main estimation equations take the following form:where the dependent variable \(Wellbeing_{it}\) is an indicator of well-being (affective well-being, cognitive well-being) for employee \(i\) in year \(t\). The main explanatory variable \(PsychDetach_{it}\) is the detachment index for individual i in year t. \({\varvec{X}}\) is a vector of time-variant control variables, including socio-demographic and job-related factors. \(\alpha_{i}\) is the unobserved time-invariant employee effect, \(y_{t}\) is the year of observation fixed-effect, and \(u_{it}\) is the error term. Our parameter of interest is \(\gamma_{1}\), capturing the effect of psychological detachment from work on different measures of employee well-being. We also estimate variations of (1) with and without control variables.
$$Wellbeing_{it} = \gamma_{1} PsychDetach_{it} + \beta {\varvec{X}}_{it} + y_{t} + \alpha_{i} + u_{it}$$
(1)
4 Results
4.1 Psychological Detachment from Work and Affective Well-Being
Table 3 reports the initial estimates of the relationship between psychological detachment from work and affective well-being.8 We show estimates for five different measures of affective employee well-being: angry, worried, happy, sad, and affect balance. Panel A provides concise estimates without control variables. Psychological detachment from work is negatively associated with anger, worry, and sadness, while it is positively associated with happiness and affect balance. A one-point increase in psychological detachment from work decreases anger, worry, and sadness by 0.14, 0.10, and 0.13 points, respectively and increases happiness and affect balance by 0.11 and 0.23 points, respectively (p < 0.01). This provides a first indication that psychological detachment from work is positively associated with affective employee well-being.
Table 3
Psychological detachment from work and affective well-being
Panel A: Without controls | |||||
|---|---|---|---|---|---|
(1) | (2) | (3) | (4) | (5) | |
Angry | Worried | Happy | Sad | AB | |
Psychological detachment | − 0.143 (0.02)*** | − 0.103 (0.02)*** | 0.109 (0.01)*** | − 0.127 (0.02)*** | 0.233 (0.02)*** |
Socio-demographic characteristics | – | – | – | – | – |
Job-related characteristics | – | – | – | – | – |
Year fixed-effects | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.036 | 0.013 | 0.012 | 0.019 | 0.026 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 | 5511 |
Panel B: With controls for socio-demographic characteristics | |||||
|---|---|---|---|---|---|
(6) | (7) | (8) | (9) | (10) | |
Angry | Worried | Happy | Sad | AB | |
Psychological detachment | − 0.142 (0.02)*** | − 0.103 (0.02)*** | 0.109 (0.01)*** | − 0.127 (0.02)*** | 0.233 (0.02)*** |
Socio-demographic characteristics | ✓ | ✓ | ✓ | ✓ | ✓ |
Job-related characteristics | – | – | – | – | – |
Year fixed-effects | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.038 | 0.014 | 0.016 | 0.020 | 0.029 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 | 5511 |
Panel C: With controls for socio-demographic and job-related characteristics | |||||
|---|---|---|---|---|---|
(11) | (12) | (13) | (14) | (15) | |
Angry | Worried | Happy | Sad | AB | |
Psychological detachment | − 0.137 (0.02)*** | − 0.104 (0.02)*** | 0.107 (0.01)*** | − 0.126 (0.02)*** | 0.229 (0.02)*** |
Socio-demographic characteristics | ✓ | ✓ | ✓ | ✓ | ✓ |
Job-related characteristics | ✓ | ✓ | ✓ | ✓ | ✓ |
Year fixed-effects | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.046 | 0.029 | 0.021 | 0.030 | 0.042 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 | 5511 |
In the next step, we add controls for employees’ socio-demographic characteristics. Panel B presents the results. First and most importantly, the inclusion of socio-demographic characteristics does not change the association between psychological detachment from work and affective employee well-being. The association remains highly significant (p < 0.01) and of similar magnitude. Second, it shows that some socio-demographic characteristics emerge as significant predictors of affective well-being. Age is positively associated with affect balance but negatively associated with all the other four outcome variables. Education is positively associated with happiness and affect balance, while getting married is negatively associated with both. An increase in the number of children in the household increases the level of anger significantly. Moving from Western West Germany to East Germany increases happiness and affect balance,9 while a move to Southern West Germany only increases happiness.
Finally, Panel C additionally includes job-related controls. The job-related controls indicate that an increased monthly gross income decreases the sadness level of employees, while increased working hours decrease the worry level. Switching to a fixed-term contract increases the probability of being worried, while switching to employment in the public sector decreases the probability of feeling sad. One additional year of tenure with the current firm increases anger and decreases affect balance. Importantly, job insecurity increases anger, worry, and sadness, while it decreases happiness and affect balance. Employees choosing to work in firms with 200–1999 employees are less likely to experience anger and those choosing to work in firms with more than 2000 employees are more likely to feel happy and score higher in affect balance, while less likely to feel sad.
Most importantly, psychological detachment from work again emerges as a statistically significant determinant of the affective well-being measures. The coefficients are very similar to those without control variables and to those with only socio-demographic controls. Thus, the primary pattern of increased affective employee well-being remains, suggesting that the difference in well-being does not simply reflect differences in socio-demographic or job-related factors. A one-point increase in psychological detachment score decreases anger, worry, and sadness by 0.14, 0.10, and 0.13 points, respectively and increases happiness and affect balance by 0.11 and 0.23 points, respectively (p < 0.01). Taking the mean levels of the affective well-being variables into account, these could be translated into a 5–6 percent decrease in anger, worry, and sadness. It also implies a 3 percent increase in happiness and a 17 percent increase in affect balance. The effect of psychological detachment on these variables is similar to an inverse effect of job insecurity on the respective variables. For example, while a one-point increase in psychological detachment decreases sadness by 6 percent, having job insecurity increases sadness by the same magnitude.
4.2 Psychological Detachment from Work and Cognitive Well-Being
Having examined the impact of detachment on affective well-being, we now turn to cognitive well-being as a potential outcome of detachment. Table 4 shows the initial estimates of the association between psychological detachment from work and cognitive well-being using satisfaction with five different areas of life as outcome variables (i.e., health satisfaction, sleep satisfaction, job satisfaction, leisure time satisfaction, and family life satisfaction).10 Panel A presents concise estimates without control variables. Psychological detachment from work is positively associated with all the five satisfaction measures, with the largest impact on job satisfaction and lowest impact on family life satisfaction. A one-point increase in psychological detachment from work increases health satisfaction, sleep satisfaction, job satisfaction, leisure time satisfaction, and family life satisfaction by 0.17, 0.31, 0.47, 0.30, and 0.12 points, respectively (p < 0.01). This provides first evidence that psychological detachment from work not only increases affective well-being, but it also increases cognitive well-being significantly.
Table 4
Psychological detachment from work and cognitive well-being: satisfaction with various domains of life
Panel A: Without controls | |||||
|---|---|---|---|---|---|
(1) | (2) | (3) | (4) | (5) | |
HS | SS | JS | LTS | FLS | |
Psychological detachment | 0.167 (0.03)*** | 0.312 (0.03)*** | 0.470 (0.03)*** | 0.298 (0.03)*** | 0.124 (0.03)*** |
Socio-demographic characteristics | – | – | – | – | – |
Job-related characteristics | – | – | – | – | – |
Year fixed-effects | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.012 | 0.025 | 0.037 | 0.017 | 0.003 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 | 5511 |
Panel B: With controls for socio-demographic characteristics | |||||
|---|---|---|---|---|---|
(6) | (7) | (8) | (9) | (10) | |
HS | SS | JS | LTS | FLS | |
Psychological detachment | 0.166 (0.03)*** | 0.312 (0.03)*** | 0.469 (0.03)*** | 0.295 (0.03)*** | 0.125 (0.03)*** |
Socio-demographic characteristics | ✓ | ✓ | ✓ | ✓ | ✓ |
Job-related characteristics | – | – | – | – | – |
Year fixed-effects | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.013 | 0.029 | 0.038 | 0.022 | 0.005 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 | 5511 |
Panel C: With controls for socio-demographic and job-related characteristics | |||||
|---|---|---|---|---|---|
(11) | (12) | (13) | (14) | (15) | |
HS | SS | JS | LTS | FLS | |
Psychological detachment | 0.159 (0.03)*** | 0.298 (0.03)*** | 0.441 (0.03)*** | 0.270 (0.03)*** | 0.116 (0.03)*** |
Socio-demographic characteristics | ✓ | ✓ | ✓ | ✓ | ✓ |
Job-related characteristics | ✓ | ✓ | ✓ | ✓ | ✓ |
Year fixed-effects | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.025 | 0.037 | 0.085 | 0.034 | 0.012 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 | 5511 |
Panel B includes controls for socio-demographic characteristics. Age is negatively associated with sleep satisfaction and positively associated with leisure time satisfaction. Marriage is negatively associated with sleep and leisure time satisfaction. Education is a positive significant determinant of job satisfaction. An increase in the number of children in the household decreases sleep and family life satisfaction, while an increase in the size of the household decreases leisure time satisfaction but increases family life satisfaction. Moreover, psychological detachment from work continues to be a statistically significant positive determinant of all the five satisfaction scores (p < 0.01). The estimated coefficients are similar to those without controls for socio-demographic characteristics.
In the next step, we additionally control for job-related factors. The results are reported in Panel C. Monthly gross income is positively associated with job satisfaction, while working hours are negatively associated with sleep, job, and leisure time satisfaction. Employees with fixed-term contracts have higher job satisfaction and lower leisure time satisfaction. Tenure is negatively associated with health and job satisfaction. Employees choosing to work in the public sector have significantly higher job satisfaction, while those switching to a part-time contract have lower health, sleep, and job satisfaction. Moreover, job insecurity is a significant negative determinant of all the five satisfaction scores.
Importantly, while the inclusion of job-related controls slightly decreases the effect of psychological detachment, it does not alter the primary pattern of increased cognitive employee well-being. A one-point increase in psychological detachment score increases health, sleep, job, leisure time, and family life satisfaction by 0.16, 0.30, 0.44, 0.27, and 0.12 points, respectively (p < 0.01). Thus, the hypothesis that psychological detachment from work increases cognitive employee well-being cannot be rejected. Taking the mean levels of cognitive well-being variables into account, these could be translated into about a 2–6 percent increase in health, sleep, job, leisure time, and family life satisfaction. Family life satisfaction is influenced the least (2 percent), while employees’ job satisfaction is impacted the most (6 percent). The influence of psychological detachment on cognitive well-being measures is almost equivalent and in the opposite direction to the influence of job insecurity on the respective variables. For example, job insecurity decreases job satisfaction by about 0.5 points while a one-point increase in psychological detachment increases job satisfaction by 0.4 points.
Furthermore, in addition to satisfaction scores with each life domain, we also use the average of these domain satisfactions and global life satisfaction scores as outcome variables. Table 5 presents the estimates.11 Again, we start without control variables and sequentially add socio-demographic and job-related factors to our model. When no controls are used, a one-point increase in psychological detachment from work increases the average domain satisfaction and overall life satisfaction by 0.27 and 0.14 points, respectively (p < 0.01). Adding socio-demographic characteristics does not change this pattern and the psychological detachment coefficients remain unchanged. Age emerges as a negative determinant of average domain satisfaction and as a positive determinant of overall life satisfaction. Education is positively associated with both average domain and life satisfaction. Getting married decreases average domain satisfaction, while moving to East Germany (from Western West Germany) increases life satisfaction. Moreover, an increase in the size of the household decreases overall life satisfaction.
Table 5
Psychological detachment from work and cognitive well-being: average domain satisfaction and overall life satisfaction
(1) | (2) | (3) | (4) | (5) | (6) | |
|---|---|---|---|---|---|---|
DS | LS | DS | LS | DS | LS | |
Psychological detachment | 0.274 (0.02)*** | 0.144 (0.02)*** | 0.273 (0.02)*** | 0.145 (0.02)*** | 0.257 (0.02)*** | 0.133 (0.02)*** |
Socio-demographic characteristics | – | – | ✓ | ✓ | ✓ | ✓ |
Job-related characteristics | – | – | – | – | ✓ | ✓ |
Year fixed-effects | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.036 | 0.013 | 0.040 | 0.017 | 0.062 | 0.036 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 | 5511 | 5511 |
In the final step, we again add job-related controls. An increase in the monthly gross income increases overall life satisfaction. Working hours, tenure, part-time contract, and job insecurity are negatively associated with both average domain satisfaction and life satisfaction. More central to our topic, while psychological detachment from work has slightly lower coefficients, it continues to emerge as a statistically significant positive determinant of both average domain satisfaction and overall life satisfaction. A one-point increase in psychological detachment score is associated with 0.26 and 0.13 points higher domain satisfaction and overall life satisfaction, respectively (p < 0.01).12 Taking the mean levels of domain and life satisfaction into account, these could be translated into a 4 percent increase in average domain satisfaction and 2 percent increase in overall life satisfaction. The effect on overall life satisfaction is equivalent to a one-percent increase in gross income, while the effect on average domain satisfaction is equivalent to having job security. Finally, on the basis of the results shown in Tables 4 and 5, we can conclude that psychological detachment from work positively affects cognitive employee well-being, indicating that our hypothesis cannot be rejected.
4.3 Heterogeneity Analyses
The results so far showed a stable and robust relationship between psychological detachment from work and all indicators of employee well-being, including affective as well as cognitive well-being. This finding is notably robust to different specifications and holds even after controlling for other factors that may influence both psychological detachment and well-being. We further recognize that the association between psychological detachment and well-being may differ with workers’ circumstances and characteristics.
Firstly, since our sample consists of periods before and during the Covid-19 global pandemic, we are the first to explore the moderating role of the Covid-19 pandemic on the association between psychological detachment and well-being. Investigating the moderating role of Covid-19 provides a unique understanding of whether or not psychological detachment from work enhances employees’ well-being even in times of increased uncertainty, stress, and health risks (see Brodeur et al., 2021). From a theoretical viewpoint, the Covid-19 crisis could play both a positive and a negative moderating role. On the one hand, psychological detachment could be less relevant for employee well-being during the crisis than before. The Covid-19 pandemic was a threat to employee well-being. Layoffs, job uncertainty, increased workloads, and the use of remote work posed challenges on employees (De-la-Calle-Durán & Rodríguez-Sánchez, 2021). These other determinants of well-being—that are particularly evident during such difficult times—might overshadow the relevance of psychological detachment for well-being. On the other hand, psychological detachment could be particularly relevant for employee well-being during the crisis. According to the stressor-detachment model, psychological detachment might buffer the negative effects of stress on well-being during such uncertain and stressful times, reflecting its safeguarding role. Therefore, it becomes an empirical question to determine the moderating role of the Covid-19 crisis.
Table 6 presents the results differentiating between the period before and during the Covid-19 pandemic. We generate a dummy for Covid-19 equal to 1 if the survey year is during the Covid-19 pandemic years. The dummy equals zero if the survey year is before the global pandemic. To investigate the moderating role of the Covid-19 pandemic, we add the Covid-19 variable and its interaction with psychological detachment in our key regressions. While we find no significant moderating role of the Covid-19 pandemic on cognitive well-being variables, the interaction term emerges as a statistically significant determinant of affect balance. Column (1) shows that a one-point increase in psychological detachment from work increases affect balance by 0.20 points prior to pandemic and by 0.29 points (0.203 + 0.084 = 0.287) during the Covid-19 pandemic. This implies that psychological detachment from work impacts cognitive well-being to a similar degree before and during the crisis. However, the positive impact of psychological detachment on affect balance is larger during the crisis.
Table 6
Psychological detachment from work and employee well-being: before and during Covid-19
(1) | (2) | (3) | (4) | |
|---|---|---|---|---|
AB | JS | DS | LS | |
Psychological detachment | 0.203 (0.02)*** | 0.427 (0.04)*** | 0.261 (0.02)*** | 0.130 (0.03)*** |
Covid-19 | − 0.244 (0.10)** | − 0.133 (0.16) | 0.099 (0.10) | − 0.130 (0.12) |
Psychological detachment x Covid-19 | 0.084 (0.03)*** | 0.046 (0.05) | − 0.015 (0.03) | 0.008 (0.04) |
Control variables | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.044 | 0.085 | 0.062 | 0.036 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 |
At issue is which dimension of the affect balance drives this result. Therefore, we further investigate the moderating role of Covid-19 on the four dimensions of affective well-being separately, i.e., anger, worry, happiness, and sadness.13 While we find no significant moderating role of Covid-19 on anger, worry, and sadness, the interaction term emerges as statistically significant in the happiness regression. A one-point increase in psychological detachment increases happiness by 0.09 points prior to the pandemic and by 0.15 points (0.090 + 0.055 = 0.145) during the Covid-19 pandemic. This implies that the stronger role of psychological detachment on affect balance during the crisis is particularly due to its stronger impact on happiness. Overall, the findings indicate that psychological detachment from work is an important determinant of well-being regardless of the existence of a global pandemic and crisis. While boosting employee well-being can be highly challenging during such difficult times, psychological detachment from work proves as a highly important determinant of employee well-being, especially for employees’ happiness.
Secondly, to check whether the association between psychological detachment and well-being is universal, we also experiment with heterogeneity analyses by gender, sector, region, marital status, existence of children in the household, and age. Table 7 shows the results from explorative sample splits. Across all the 12 splits, psychological detachment from work is a significant positive determinant of the key well-being measures. The effect of psychological detachment on affect balance is consistently positive and significant, ranging from 0.20 (for those with children in the household) to 0.27 (for younger employees). The influence on job satisfaction is also consistently positive and significant ranging from 0.39 (for those without children in the household) to 0.54 (for those residing in East Germany). Moreover, the effect of psychological detachment on average domain satisfaction ranges from 0.21 (for those with children in the household) to 0.29 (for younger employees). Finally, the positive significant effect of psychological detachment on global life satisfaction ranges from 0.08 (for employees with children in the household) to 0.20 (for younger employees). Overall, these findings indicate that while the association between psychological detachment and well-being is always positive, the magnitude of the effect may vary slightly. Employees with children in the household have the lowest and younger employees or those residing in East Germany have the highest magnitudes. However, the differences in magnitudes are not significant. Thus, we find that the influence of psychological detachment on employee well-being is universal and is similar for all employees regardless of their socio-demographic characteristics. This further highlights the importance of psychological detachment for enhancing employee well-being.
Table 7
Psychological detachment from work and employee well-being: heterogeneity
Panel A: Gender | ||||||||
|---|---|---|---|---|---|---|---|---|
Female | Male | |||||||
(1) | (2) | (3) | (4) | (5) | (6) | (7) | (8) | |
AB | JS | DS | LS | AB | JS | DS | LS | |
Psychological detachment | 0.222 (0.03)*** | 0.448 (0.05)*** | 0.261 (0.03)*** | 0.136 (0.03)*** | 0.242 (0.03)*** | 0.429 (0.05)*** | 0.256 (0.03)*** | 0.137 (0.04)*** |
Control variables | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.047 | 0.086 | 0.065 | 0.045 | 0.048 | 0.095 | 0.076 | 0.042 |
Number of observations | 6086 | 6086 | 6086 | 6086 | 5959 | 5959 | 5959 | 5959 |
Number of employees | 2795 | 2795 | 2795 | 2795 | 2716 | 2716 | 2716 | 2716 |
Panel B: Sector | ||||||||
|---|---|---|---|---|---|---|---|---|
Public | Private | |||||||
(9) | (10) | (11) | (12) | (13) | (14) | (15) | (16) | |
AB | JS | DS | LS | AB | JS | DS | LS | |
Psychological detachment | 0.246 (0.04)*** | 0.428 (0.07)*** | 0.245 (0.04)*** | 0.173 (0.05)*** | 0.225 (0.02)*** | 0.458 (0.04)*** | 0.272 (0.02)*** | 0.118 (0.03)*** |
Control variables | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.054 | 0.095 | 0.070 | 0.049 | 0.043 | 0.092 | 0.070 | 0.038 |
Number of observations | 3193 | 3193 | 3193 | 3193 | 7876 | 7876 | 7876 | 7876 |
Number of employees | 1459 | 1459 | 1459 | 1459 | 3637 | 3637 | 3637 | 3637 |
Panel C: Region | ||||||||
|---|---|---|---|---|---|---|---|---|
East Germany | West Germany | |||||||
(17) | (18) | (19) | (20) | (21) | (22) | (23) | (24) | |
AB | JS | DS | LS | AB | JS | DS | LS | |
Psychological detachment | 0.223 (0.04)*** | 0.535 (0.06)*** | 0.283 (0.04)*** | 0.176 (0.05)*** | 0.230 (0.02)*** | 0.402 (0.04)*** | 0.245 (0.02)*** | 0.119 (0.03)*** |
Control variables | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.046 | 0.110 | 0.076 | 0.068 | 0.043 | 0.081 | 0.062 | 0.030 |
Number of observations | 3078 | 3078 | 3078 | 3078 | 8885 | 8885 | 8885 | 8885 |
Number of employees | 1387 | 1387 | 1387 | 1387 | 4090 | 4090 | 4090 | 4090 |
Panel D: Marital status | ||||||||
|---|---|---|---|---|---|---|---|---|
Non-married | Married | |||||||
(25) | (26) | (27) | (28) | (29) | (30) | (31) | (32) | |
AB | JS | DS | LS | AB | JS | DS | LS | |
Psychological detachment | 0.232 (0.04)*** | 0.479 (0.06)*** | 0.260 (0.04)*** | 0.157 (0.05)*** | 0.209 (0.02)*** | 0.398 (0.04)*** | 0.241 (0.02)*** | 0.105 (0.03)*** |
Control variables | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.067 | 0.098 | 0.077 | 0.076 | 0.037 | 0.090 | 0.063 | 0.030 |
Number of observations | 3355 | 3355 | 3355 | 3355 | 7531 | 7531 | 7531 | 7531 |
Number of employees | 1548 | 1548 | 1548 | 1548 | 3480 | 3480 | 3480 | 3480 |
Panel E: Children in HH | ||||||||
|---|---|---|---|---|---|---|---|---|
Without children | With children | |||||||
(33) | (34) | (35) | (36) | (37) | (38) | (39) | (40) | |
AB | JS | DS | LS | AB | JS | DS | LS | |
Psychological detachment | 0.229 (0.03)*** | 0.387 (0.05)*** | 0.233 (0.03)*** | 0.106 (0.04)*** | 0.200 (0.03)*** | 0.429 (0.05)*** | 0.213 (0.03)*** | 0.076 (0.04)** |
Control variables | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.053 | 0.089 | 0.062 | 0.044 | 0.044 | 0.104 | 0.064 | 0.040 |
Number of observations | 5740 | 5740 | 5740 | 5740 | 4280 | 4280 | 4280 | 4280 |
Number of employees | 2636 | 2636 | 2636 | 2636 | 2050 | 2050 | 2050 | 2050 |
Panel F: Age | ||||||||
|---|---|---|---|---|---|---|---|---|
Younger (Age ≤ 40) | Older (Age > 40) | |||||||
(41) | (42) | (43) | (44) | (45) | (46) | (47) | (48) | |
AB | JS | DS | LS | AB | JS | DS | LS | |
Psychological detachment | 0.265 (0.05)*** | 0.512 (0.09)*** | 0.294 (0.04)*** | 0.195 (0.05)*** | 0.205 (0.02)*** | 0.424 (0.04)*** | 0.246 (0.02)*** | 0.140 (0.03)*** |
Control variables | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.071 | 0.100 | 0.121 | 0.076 | 0.043 | 0.085 | 0.059 | 0.039 |
Number of observations | 2374 | 2374 | 2374 | 2374 | 8177 | 8177 | 8177 | 8177 |
Number of employees | 1138 | 1138 | 1138 | 1138 | 3761 | 3761 | 3761 | 3761 |
4.4 Robustness Checks
The study so far indicated a positive association between psychological detachment and different measures of employee well-being, and showed that this positive association holds regardless of the existence of the Covid-19 pandemic and is universal across different subgroups. In this section, we perform further robustness checks to mitigate potential concerns.
First, our findings might be subject to common method bias because we used self-reported data. Multiple procedural strategies (see Jordan & Troth, 2020) were implemented during the data collection to ensure accurate answers without systematic bias: (1) one reversed coded item of detachment, (2) different scale properties of the predictor and outcome variables (4-point and 11-point Likert scales), (3) clear separation of the dependent and independent variables in the survey, and (4) emphasizing the research purpose to participants. Furthermore, the fixed-effects approach may solve the common method bias to an extent if the bias is time-invariant and does not change across waves. Nonetheless, to address and mitigate concerns regarding common method bias, we temporally separate the measurement of the dependent and independent variables by introducing a time lag (Podsakoff et al., 2003). We exploit our unique longitudinal dataset and use measures of well-being from the next wave (\(t + 1\)) instead of the current wave (\(t\)) that the explanatory variables come from.
Table 8 reports the fixed-effects estimations using the lagged well-being measures. Since the last available wave in the GSOEP is 2021, we cannot observe the well-being measures for 2022, and hence, drop this wave and use the 2011 and 2016 waves. The results show a consistent and statistically significant positive association between psychological detachment from work at the time (\(t\)) and well-being measures at time (\(t + 1\)). A one-point increase in the psychological detachment score increases affect balance, job satisfaction, average domain satisfaction, and overall life satisfaction by 0.08, 0.18, 0.10, and 0.13 points, respectively. The effect sizes for affect balance, job satisfaction, and average domain satisfaction are smaller than our initial estimates, but still highly significant. However, the effect size for life satisfaction is even slightly larger than for our initial estimates. This indicates that the common method bias does not drive the positive association between psychological detachment and employee well-being. Previous literature also suggests that single source bias is not critical for the relationships of interests (Fritz et al., 2010b).
Table 8
Psychological detachment from work and employee well-being: lagged dependent variable
(1) | (2) | (3) | (4) | |
|---|---|---|---|---|
ABt+1 | JSt+1 | DSt+1 | LSt+1 | |
Psychological detachmentt | 0.076 (0.03)** | 0.175 (0.06)*** | 0.101 (0.03)*** | 0.134 (0.04)*** |
Control variables | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.020 | 0.045 | 0.038 | 0.022 |
Number of observations | 5482 | 5482 | 5482 | 5482 |
Number of employees | 2741 | 2741 | 2741 | 2741 |
Second, in our main estimations we eliminated the problem of time-constant unobserved heterogeneity by applying individual fixed-effects panel estimations. Despite controlling for time-variant factors that potentially influence both psychological detachment and well-being, we further mitigate concerns regarding unobserved time-variant heterogeneity. Thus, we return to our initial estimations and restrict the sample to those who did not change their firms. Employees who have changed firms might experience changes in their working conditions that are not controlled for, but that are related to their well-being. Therefore, we conduct a robustness check restricting the sample to employees working for the same firm over the relevant investigation period. Table 9 shows that the results remain robust to the exclusion of workers who switched their firms. As a final robustness check, we present separate regressions using each detachment item instead of the detachment index as the explanatory variable. Table 10 shows the results. Again, each detachment item illustrates a significant positive association with our key well-being measures supporting our main hypothesis.
Table 9
Robustness check: excluding switchers
(1) | (2) | (3) | (4) | |
|---|---|---|---|---|
AB | JS | DS | LS | |
Psychological detachment | 0.201 (0.02)*** | 0.366 (0.04)*** | 0.224 (0.02)*** | 0.121 (0.03)*** |
Control variables | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.034 | 0.056 | 0.049 | 0.029 |
Number of observations | 8297 | 8297 | 8297 | 8297 |
Number of employees | 3857 | 3857 | 3857 | 3857 |
Table 10
Robustness check: single detachment items
Panel A: Only item 1 | ||||
|---|---|---|---|---|
(1) | (2) | (3) | (4) | |
AB | JS | DS | LS | |
Psychological detachment (item 1) | 0.141 (0.02)*** | 0.236 (0.02)*** | 0.143 (0.01)*** | 0.078 (0.02)*** |
Control variables | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.035 | 0.069 | 0.049 | 0.034 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 |
Panel B: Only item 2 | ||||
|---|---|---|---|---|
(5) | (6) | (7) | (8) | |
AB | JS | DS | LS | |
Psychological detachment (item 2) | 0.163 (0.02)*** | 0.362 (0.03)*** | 0.203 (0.02)*** | 0.100 (0.02)*** |
Control variables | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.035 | 0.081 | 0.057 | 0.035 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 |
Panel C: Both items | ||||
|---|---|---|---|---|
(9) | (10) | (11) | (12) | |
AB | JS | DS | LS | |
Psychological detachment (item 1) | 0.107 (0.02)*** | 0.152 (0.03)*** | 0.097 (0.02)*** | 0.057 (0.02)*** |
Psychological detachment (item 2) | 0.124 (0.02)*** | 0.306 (0.03)*** | 0.167 (0.02)*** | 0.079 (0.02)*** |
Control variables | ✓ | ✓ | ✓ | ✓ |
Within R2 | 0.042 | 0.087 | 0.063 | 0.036 |
Number of observations | 12,045 | 12,045 | 12,045 | 12,045 |
Number of employees | 5511 | 5511 | 5511 | 5511 |
4.5 Limitations and Future Research Directions
Our study addressed a series of empirical gaps in literature. However, we are aware of some limitations and suggest paths for future research. First, we rely on self-reported data. Yet our results remained robust when separating the independent and dependent variables in time, mitigating concerns regarding common method bias (Podsakoff et al., 2003). Second, the results are based on data from Germany and may not be generalized to all countries. Future research should explore the role of psychological detachment for well-being in different cultural contexts. Third, our results are based on data collected before and during the Covid-19 pandemic. Future research should test whether the same patterns emerge on the basis of post-pandemic data. Fourth, although we mitigated endogeneity concerns in several ways, future research could consider using experimental approaches to move closer towards causality. Fifth, detachment can be regarded as multidimensional, consisting of physical, cognitive, and emotional detachment (de Jonge et al., 2012). We focused on cognitive detachment. Future research should capture the different dimensions of detachment and their respective associations with employee well-being (e.g., Balk et al., 2019).
Finally, while our longitudinal design has a series of advantages, it does not come without any limitations. For example, as with most longitudinal designs, our dataset is also subject to panel attrition. If the panel attrition is not random and associated with well-being, then it may introduce attrition bias. Another limitation of our longitudinal design could be the temporal spacing between the waves (2011, 2016, 2021). Thus, our dataset consisted of many singleton observations that were not used in our fixed-effects estimations. Moreover, although our fixed-effects estimations control for any time-invariant heterogeneity, omitting time-varying variables may still influence our findings. We controlled for a series of important worker and job characteristics. However, omitting variables such as work-from-home possibilities may still influence our findings. If working from home hinders detachment but enhances worker well-being, then the coefficients of our psychological detachment variable may be underestimated. Thus, our estimated effect of psychological detachment on well-being represents a lower-bound. These limitations demonstrate the need for future research with better designs. Despite these limitations, our findings provide multiple valuable insights and pave the way for future research.
5 Discussion and Conclusions
Using representative and longitudinal data from the GSOEP, we identified psychological detachment from work during non-work hours as a key driver of employee well-being. This finding held for all our 12 indicators of well-being, including emotional responses, job satisfaction, life domain satisfactions, and life satisfaction. Thus, detachment enhances affective as well as cognitive well-being. In contrast to previous studies which were mostly cross-sectional or focused on specific homogenous occupational cohorts, we moved closer towards causal inference by using a large representative dataset and applying individual fixed-effects panel estimations. We showed that the effect is universal across subgroups of employees. Across all the 12 sample splits, psychological detachment from work is a significant positive predictor of employee well-being. We further showed that detachment affects employee well-being to a similar extent before and during the Covid-19 pandemic. While ensuring employee well-being can be highly challenging during such difficult times, organizations and employee representatives could aim to foster detachment which in turn enhances every dimension of well-being.
Our study also strengthened and refined existent theory, especially the stressor-detachment model (Sonnentag & Fritz, 2015). Well-being is a core construct of the stressor-detachment model. In particular, the stressor-detachment model suggests that lack of detachment is a predictor of poor well-being. First, our study showed that the impact of psychological detachment on well-being holds for different employee subgroups. This points to the universality of the effect proposed in the theory. Second, we found that psychological detachment influences not only a few, but all the well-being measures investigated. Diener et al. (1999) already stressed the need to refine theories in order to better understand the varying impacts that one specific input variable may have on the distinct components of subjective well-being. Previous studies have focused on a narrow range of well-being measures, limiting the ability to make broad conclusions (among others, see Sonnentag & Bayer, 2005; Sonnentag & Fritz, 2007). Our study showed that psychological detachment is positively related to multiple components of affective as well as cognitive well-being. Third, we found that psychological detachment is relevant for employee well-being before as well as during the Covid-19 pandemic, which has not previously been tested. Fourth, Sonnentag and Fritz (2015, p. S96) stressed that “future research should pay more attention to issues of causality.” We followed their call for future research. By using individual fixed-effects panel estimations, we moved closer towards causality than previous studies that were mostly cross-sectional (e.g., Hamilton Skurak et al., 2021; Kinnunen et al., 2010; Sonnentag & Fritz, 2007). Thereby, we further strengthened the general trustworthiness of the theory. Given these aspects of our study, we make an important empirical contribution to the generalizability and trustworthiness of the stressor-detachment model, which suggests psychological detachment from work as a driver of employee well-being.
Recognizing the significance of psychological detachment for employee well-being, organizations, employee representatives, and employees themselves may seek to enhance psychological detachment. Two general strategies for improving detachment can be distinguished, namely individual- and organizational-level interventions (Wendsche & Lohmann-Haislah, 2017). Individual-level interventions include recovery trainings (e.g., Hahn et al., 2011), mindfulness trainings (e.g., Michel et al., 2014), positive work reflection interventions (e.g., Bono et al., 2013), and end-of-day planning (e.g., Smit, 2016). Moreover, individuals can enhance their detachment levels by engaging in distractive activities during off-job time, such as social, low-effort, or physical activities (e.g., Mojza et al., 2011; Ten Brummelhuis & Bakker, 2012; Wendsche et al., 2021). Regarding organizational-level interventions, reduced job demands and increased job resources might help improve detachment. Organizational interventions can also promote boundary management (Wendsche & Lohmann-Haislah, 2017). Karabinski et al. (2021) provide a meta-analysis on interventions for improving psychological detachment from work.
Furthermore, our study also contributed to the Sustainable Development Goals set by the United Nations. Identifying the antecedents of well-being is essential in order to develop suitable approaches to improve well-being. Through identifying detachment as a key driver of employee well-being, we directly contribute to Goal 3: Good Health and Well-being. Moreover, healthier and happier workers are likely to be more productive, fostering economic growth (Bryson et al., 2017; de Oliveira et al., 2023). Thus, enhancing employee well-being indirectly contributes to Goal 8: Decent Work and Economic Growth. Therefore, managers, employee representatives, and policy makers should be aware of the critical role of psychological detachment from work for employee well-being.
Considering the positive well-being consequences of detachment, more studies exploring detachment-enhancing interventions are needed. Karabinski et al. (2021) emphasized the need for further research on such interventions, especially on work-directed interventions. The positive well-being effects stemming from detachment can only be realized through effective interventions. Future research should devote further efforts towards detecting detachment-enhancing measures on individual, organizational, and societal levels.
Declarations
Conflict of interest
The authors declare that there is no conflict of interest.
Ethical Approval
The GSOEP places the highest priority on protecting the confidentiality of respondents’ data by ensuring strict adherence to European and German data protection regulations and maintaining the highest standards of research ethics.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.