Work–family conflict and spouse’s job performance: when detaching from home is key

ABSTRACT

Work–family conflict (WFC) is a common experience which frequently affects employees’ performance and wellbeing. But among dual-earner couples, is it possible that employees’ WFC relates to spouses’ job performance? Why does this occur and what are the associated consequences? Drawing on crossover literature and the stressor–detachment model, we explore a number of crossover and spillover effects. Specifically, we propose that employees’ WFC negatively affects spouses’ home detachment and job performance on the next day (crossover effects), which relates to spouses’ lower work detachment (spillover effects). We conducted a diary study among 145 working couples who responded to a daily survey over five consecutive working days (N = 1450 occasions). Our results from multilevel analyses largely supported our hypotheses. Employees’ WFC predicted lower home detachment of spouses on the following day, which, in turn, negatively affected spouses’ performance and work detachment. These findings show that WFC affects spouses beyond the family domain and that detachment from family-related issues is key to performing well and disconnecting from work.

KEYWORDS:

• Crossover
• spillover
• work–family conflict
• home detachment
• work detachment

Introduction

In the twenty-first century, despite the development of technology, better flexible working arrangements, or improvements in sharing family responsibilities, people continue to find it difficult to juggle work and family. This problem has become more evident with the COVID-19 pandemic (Vaziri et al., 2020). When family life breaches the boundary of work, or vice versa, work–family conflict arises . Work–family conflict has been defined as “a form of interrole conflict in which the role pressures from the work and family domains are mutually incompatible in some respect” (Greenhaus & Beutell, 1985, p. 77). This definition suggests a bidirectional relationship between the two domains (Netemeyer et al., 1996), that is, work-to-family conflict (WFC) and family-to-work conflict (FWC). WFC takes place when work-related demands interfere with family responsibilities (e.g. working in the evening instead of spending time with the family), whereas FWC occurs when family responsibilities interfere with employees’ working life (e.g. having to leave work because a family member becomes ill). The conflict between these two domains appears to be a key stressor (Judge & Colquitt, 2004) that negatively affects employees’ work- and family-related outcomes (Steiner & Krings, 2016). In this study, because we are examining the experience of employees while they are at home, we analyse whether work interferes with family life. Therefore, the direction of interest is WFC.

The study of work and family involves significant others, so the effects of WFC on family members cannot be ignored (Bakker & Demerouti, 2013; Westman, 2001). For instance, there is evidence of the crossover of WFC among couples, that is, how employees’ WFC affects spouses’ stress or strain reactions (Westman, 2001). When employees’ work responsibilities interfere with their family life, spouses report higher exhaustion (Bakker et al., 2008), reduced family satisfaction (Lu et al., 2016), and higher distress (Cowlishaw et al., 2010). But what are the implications for spouses beyond family- or strain-related outcomes? How does employees’ WFC affect spouses at work? In this study, we answer this question by examining crossover and spillover effects.

Firstly, to explain the crossover effect, we introduce home detachment as an explanatory mechanism. This concept is defined as “disengaging oneself mentally from private issues and stopping to think about personal or family-related problems’ (Sanz-Vergel et al., 2011, p. 777). We propose that the reason as to why spouses are affected by the focal employee’s WFC is their lack of home detachment; that is, they remain thinking about family-related issues even when they are not with their partner. This argument is based on the stressor–detachment model (Sonnentag & Fritz, 2015), which posits that constant cognitive activation, even when the stressor is no longer present, hinders the experience of psychological detachment. Secondly, we focus on task performance, which refers to activities that “are recognized by formal reward systems and are part of the requirements as described in job descriptions’ (Williams & Anderson, 1991, p. 606). For organisations, task performance is a key element with which to achieve their goals (Campbell, 1983), and for employees, it helps to gain career opportunities (Campion et al., 1994), detach from work (Rodríguez-Muñoz et al., 2018), and reduce turnover intentions (Zimmerman & Darnold, 2009). Thirdly, in addition to crossover, it is widely known that experiences in one role affect experiences in the other, rendering the roles more similar (Edwards & Rothbard, 2000). This phenomenon is called spillover and refers to associations found in the same individual (Bakker & Demerouti, 2013). Whereas the spillover of work detachment (the “individual’s sense of being away from the work situation”, Etzion et al., 1998, p. 579) has been widely analysed (Gaudiino & Di Stefano, 2021; Kondrysova et al., 2022; Sonnentag et al., 2017), the spillover of home detachment remains unexplored. In this respect, an appropriate outcome that aligns with the concept of home detachment is work detachment. Sonnentag and Fritz (2007) consider the ability to detach to be a core psychological experience that helps to recover after being exposed to work-related stressors. When individuals are not at work, if they are still thinking about work-related issues, recovery cannot occur. In the present study, and also based on the stressor–detachment model, we propose that employees’ WFC affects spouses’ home detachment and performance and that these effects spill over in the form of spouses’ poor work detachment. For clarity purposes, the conceptual model of the study has been depicted in Figure 1.

Figure 1. Conceptual model of the study.

Our study contributes to the literature in three ways. Firstly, our study will help to gain understanding of the relationship between employees’ WFC and spouses’ work-related outcomes. So far, studies have been mainly focused on how employees’ WFC affects spouses in the home domain (e.g. reducing family satisfaction), whereas the study of how employees’ WFC affects spouses in the work domain has been largely neglected (Booth-LeDoux et al., 2020; Hammer et al., 2003). Incorporating spouses’ job performance helps to understand the ample consequences that WFC has for significant others beyond the family sphere. If we ignore these consequences, the picture that we obtain is incomplete (Li et al., 2021). Secondly, we contribute to the study of crossover (Westman, 2001) by incorporating a new mechanism that explains why employees’ WFC is linked to spouses’ performance: a lack of home detachment. This is the first study to highlight the crucial role that cognitive aspects have in the explanation of crossover processes. Moreover, we respond to recent calls to focus on home detachment and how it affects performance (Haun et al., 2018). Thirdly, our study adds to research on spillover by analysing the link between two types of detachment – one taking place at work (home detachment) and another taking place at home (work detachment). The relationship between these two types of detachment has not been previously examined. This means that our study may shed light on why employees enter into a vicious cycle of a lack of detachment in which they are not able to focus on the appropriate task at the appropriate time. Our results in this respect can inform the literature on recovery and encourage scholars to further examine potential antecedents and consequences of home detachment, as well as associated crossover and spillover effects.

In terms of the research design, we use a within-person approach (daily diary study) because, as Maertz and Boyar (2011) pointed out, the study of WFC benefits from a dynamic view. The incompatibilities between the roles take place at specific episodes that may vary within individuals, rather than being a consolidated experience with reliable meaning across individuals.

Theoretical framework and hypotheses

We use the stressor–detachment model as a framework for our study (Sonnentag & Fritz, 2015). These authors take insights from the cognitive activation theory (Meurs & Perrewé, 2011; Ursin & Eriksen, 2010) to explain the link between stressors and a lack of psychological detachment. The two main insights taken from the cognitive activation theory are: firstly, that stressors generate an immediate response at both the physiological and the psychological levels, and, secondly, that it is not primarily the acute stress reaction that is detrimental to an organism, but rather the sustained cognitive activation, even when the stressor is no longer present. For example, according to Sonnentag and Fritz, employees may remain cognitively activated after work (i.e. recalling a negative event that happened in the morning), which makes it difficult for them to detach in the evening. It is the constant activation, without the stressor present, which hinders the experience of psychological detachment.

Applying this reasoning to our model, the experience of employees’ WFC acts as a stressor for spouses. In fact, previous studies in the field have provided evidence that employees’ WFC is linked to negative reactions of spouses, such as reduced physical and mental wellbeing (Lu et al., 2016; Ratnaningsih & Idris, 2023), depression (Hammer et al., 2005), exhaustion (Bakker et al., 2008), or relationship tension (Matthews et al., 2006). Beyond the stressor–strain link, other studies have examined how employees’ WFC may affect spouses in practical terms (i.e. generating more work at home or more family interruptions at work). For example, Bakker et al. (2008) found that employees’ WFC was positively related to social undermining, which, in turn, related to partners’ perception of higher home demands and higher FWC. Hammer et al. (2003) found that wives’ FWC was positively related to husbands’ interruptions at work. These findings suggest that employees’ experiences of conflict have an impact on the ability of the spouses to cope with work due to family-related issues. We build on these studies to analyse a related form of crossover in which employees’ WFC impairs spouses’ ability to cognitively detach from home. In the context of marital interactions, cognitions are key influences on the behaviour of couples (Neff & Karney, 2004), and as Meier and Cho (2019) pointed out, poor psychological detachment is an additional cognitive mechanism that links work and family lives. We argue that when the employee is no longer present, his or her spouse may remain cognitively active on the following day, thinking about the practical implications derived from the employee’s WFC (e.g. who may take care of the children, cook, clean) or thinking about the wellbeing of their partner. According to Sonnentag and Fritz’s stressor–detachment model (2015), this activation would impair the ability to detach. Therefore, in our first hypothesis we propose that:

Hypothesis 1: Employees’ WFC in the evening is negatively related to spouses’ home detachment on the next day.

The stressor–detachment model (Sonnentag & Fritz, 2015) also states that a lack of detachment, in turn, explains how stressful experiences translate into poor health and wellbeing. In this study, we go one step further in the application of this model and propose that spouses’ lack of home detachment may be linked to spouses’ lower performance. In fact, there is evidence that cognitive aspects related to the private domain affect performance. For example, Beal et al. (2005) suggested that when an individual cannot allocate the maximum amount of resources to the task at hand, the individual cannot perform at his or her best. This is in line with Kahn’s (1990) proposition in respect of the investment of energy. In particular, investing cognitive energy in work roles contributes to organisational goals, as it promotes behaviour that is more vigilant, attentive and focused. If cognitive energy is spent on thinking about home-related issues, it cannot be directed towards a task. In addition, Demerouti et al. (2007) found that a lack of concentration adversely affected task performance over time. In the only study that we found on home detachment, the inability to detach from family-related issues was related to lower task performance on a daily basis (Sanz-Vergel et al., 2011). According to these authors, in the same way in which detaching from work is important in recovering from stress, disconnecting from home is crucial in performing well. Therefore, in our second hypothesis we propose that:

Hypothesis 2: Spouses’ home detachment is positively related to spouses’ job performance (same person, same-day effects).

In the first two hypotheses, we proposed that an experience occurring in the family domain (WFC) crosses over to the spouse in the form of reduced home detachment and performance during the following day at work (in the afternoon). Our third hypothesis focuses on how spouses’ low performance, in turn, affects spouses’ ability to detach from work during the evening. This is known as spillover; that is, experiences in one role affect experiences in the other (Edwards & Rothbard, 2000). To develop our third hypothesis, we also build on Sonnentag and Fritz’s (2015) stressor–detachment model and more specifically, on their explanation about why employees may not detach from work. They point out that a lack of detachment may occur a) when individuals recall a negative experience lived at work or b) when they think about the next day’s work. Based on this, we argue that spouses may not be able to detach from work because they remain cognitively activated in the evening due to thinking either that they have not performed well or about the accumulated tasks that they will have on an upcoming day. Indeed, there is evidence that daily low performance hampers work detachment during leisure time (Rodríguez-Muñoz et al., 2018). Furthermore, it has been demonstrated that full detachment may be hampered when an employee thinks about an upcoming work situation (e.g. accumulated or incomplete tasks) (Smit, 2016). In the same vein, Leroy (2009) found that disengaging from a task and redirecting the attention towards a different matter was difficult for employees with unfinished tasks. Therefore, in our third hypothesis we propose that:

Hypothesis 3: Spouses’ job performance in the afternoon is positively related to spouses’ work detachment in the evening (same person, same-day effects).

Taking all of these arguments together, we propose a vicious cycle that starts with employees experiencing WFC, continues with spouses’ inability to detach from home and perform at work on the next day, and finishes in the evening by spouses not being able to detach from work. In our final hypothesis, we propose that:

Hypothesis 4: Employees’ WFC is negatively related to spouses’ work detachment on the next day (evening) through spouses’ low home detachment and job performance on the next day (afternoon).

Method

Procedure and sample

We recruited participants from different organisations and sectors. The inclusion criteria were as follows: both members of the couple had to be employed, living together, and interacting for at least one hour during the evening. We excluded employees who worked a night shift or did not work five consecutive working days from Monday to Friday.

Participants came from a broad range of occupational backgrounds, with most of them working in the following sectors: health (13.2%), financial institutions (12.9%), trade (11.5%), industry (10.8%), and education (6.3%). The researchers organised research seminars as part of an introductory course on Organisational Psychology. In those seminars, they explained to students the advantages of diary studies, the objectives of this particular study, and the benefits that this study may have for employees. Students also benefitted from an explanation as to how to design and fill in a diary study. Those who were interested were invited to participate in the process of data collection. Each student contacted at least one dual-earner couple from their social network, and they were granted extra course credits. Scholars in the field of organisational behaviour recognise these benefits and remind that this is a common and appropriate method, especially when dealing with difficult designs involving high effort from participants (e.g. Demerouti & Rispens, 2014; Matta et al., 2015). Please note that students were not disadvantaged in any way, as those who were not interested in participating or who thought that they could not recruit even one couple were offered an alternative activity to gain extra course credits (i.e. reading selected papers on diary studies and spillover–crossover effects and writing an essay reflecting and critically discussing the studies).

We obtained formal approval from the Deontological Ethics Committee of the institution in which the study was carried out. We collected data in Spain in 2014, following two different phases. Firstly, all participants completed a general questionnaire including demographics (e.g. gender, age, and organisational information). Secondly, participants completed diary booklet surveys during five consecutive working days, twice a day (i.e. afternoon and evening). We chose to use paper-and-pencil surveys in order not to restrict the sample to a potentially preselected group of participants who had internet access or basic computer/smartphone skills (cf. Ohly et al., 2010). We asked employees to specify the time at which they were filling in the survey. The afternoon survey was filled in immediately after work and included measures of job performance and psychological home detachment. This survey was completed at 18:31 on average. The evening survey was filled in before going to bed and included measures of psychological work detachment and WFC. This survey was completed at 22:47 on average. Responses of partners were linked by using anonymous codes provided by the participants. All surveys plus written consent forms were returned via the students. Finally, because being aware of each other’s emotions and detachment while filling in the diaries could affect their self-ratings, one of the instructions for couples was not to discuss their answers between them. Students recruiting the sample emphasised the importance of answering the surveys separately.

In total, 380 employees voluntarily agreed to participate. Of these employees, 306 valid questionnaires were returned (80.5% response rate). Questionnaires with at least one day missing or answered at inappropriate times were excluded (e.g. afternoon questionnaire answered immediately before the evening questionnaire). As a result, only eight couples had to be removed. The final sample comprised 145 heterosexual couples (N = 290 participants and N = 1450 occasions). The total number of occasions is appropriate according to the guidelines by Ohly et al. (2010) and Gabriel et al. (2019). Specifically, Ohly et al. suggest a minimum of 100 participants over five days, and Gabriel et al. recommend a sample size of at least 83 at the person level and 835 observations at Level 1.

The mean age was 43.74 (range 19–63) years (SD = 9.96), and the mean job tenure was 20.44 (range 1–45) years (SD = 11.22). On average, they worked 40.05 (range 12–60) hours per week (SD = 8.43). Seventy point two per cent of the couples had at least one child, and 51.5% of the sample had a university degree or postgraduate studies.

Measures

Daily survey data

Daily job performance was measured with four items from the performance scale of Williams and Anderson (1991). We examined aspects of task performance (performance in required duties and responsibilities). A sample item is “Today, I have adequately completed assigned duties’. Items were rated on a six-point scale, ranging from 1 = not true at all to 6 = totally true. The average internal consistency (Cronbach’s alpha) across five days was .80.

Daily work–family conflict was measured with three items from the Survey Work–Home Interference-Nijmegen (SWING; Geurts et al., 2005), modified to measure daily experience. Participants had to indicate how often they had experienced each situation (e.g. “During the evening, my work obligations have made it difficult for me to feel relaxed at home”). Items were rated on a six-point scale, ranging from 1 = not true at all to 6 = totally true. The average Cronbach alpha across five days was .75.

Daily psychological home and work detachment were measured with six adapted items of the daily version (Bakker et al., 2015) of the Recovery Experience Questionnaire (Sonnentag & Fritz, 2007). Participants had to indicate how often they had experienced each situation at work (e.g. “Today, during my off-job time … , I didn’t think about work at all”) and at home (e.g. “Today, at work, I haven’t thought about my home responsibilities at all”). Items were rated on a six-point scale, ranging from 1 = not true at all to 6 = totally true. The average Cronbach alpha was .90 for work detachment and .80 for home detachment.

General survey data

Sociodemographic information was collected in the general questionnaire (e.g. age, gender, number of children, educational level, job tenure, supervisory position, number of hours worked per week).

Strategy of analysis

We relied on the actor–partner interdependence model (APIM; Kenny et al., 2006) to analyse our data. This is an appropriate method when individuals cannot be considered as independent from one another – in other words, there is common variance shared by the couples (Kashy & Kenny, 2000). With APIM, we can explore how employee’s WFC affects spouse’s home detachment (this is called “partner effect”), but also how spouse’s home detachment affects spouse’s performance and spouse’s work detachment (this is called “actor effect”). This allows us to analyse the two processes presented in Figure 1: first, the crossover, where we explore partner effects, and second, the spillover, where we explore actor effects.

We took into account the hierarchical structure, as we used a research design with repeated measurements from each member of the couple, resulting in a dataset with a multilevel structure: the first level (within-person) refers to the daily measurements, while the second level (between-person) refers to each individual included in the sample. Following a recent recommendation in the methodological literature (Bliese et al., 2018), we decided not to centre our variables, as we were not interested in testing cross-level interactions (for which it would have been of essential utility). To analyse the data, we employed Mplus 8 (Muthén & Muthén, 2012), and followed Preacher et al. (2010) to test our sequential-mediation model.

Preacher et al. recommend that mediation with two levels of analysis (within-person and between-person in the same model) is carried out using multilevel structural equation modelling or multilevel path analysis. This framework is useful to assess all types of multilevel mediation, including the one proposed in this study, in which all variables are measured at Level 1.

Results

Preliminary analyses

Prior to the multilevel analyses, we calculated the intraclass correlations for all of the study variables. Thus, we can demonstrate that variability exists at the person and day levels. According to Byrne (2011), when ICC values are larger than .10 and smaller than .90 there is a substantive amount of variance at that level of analysis and, therefore, the use of multilevel analysis is appropriate. In our case, the person-level variance explained by each variable clearly supports the use of multilevel modelling: 42.2% for employees’ WFC, 50% for spouses’ home detachment, 53.1% for spouses’ job performance, and 52.7% for spouses’ work detachment. Next, we calculated means, standard deviations, and between-person and within-person correlations for all of the study variables (see Table 1). The pattern of correlations was in the expected direction.

Table 1. Means, standard deviations, and correlations.

Variable	M (SD)	1	2	3	4
1. Employees’ WFC (evening)	2.24 (1.21)	–	–.08**	–.05	.01
2. Spouses’ home detachment (afternoon)	3.36 (1.41)	.00	–	.18**	.06*
3. Spouses’ job performance (afternoon)	5.16 (0.83)	–.19**	.23**	–	.09**
4. Spouses’ work detachment (evening)	4.46 (1.46)	–.12*	.17**	.30**	–

* p < .05, ** p < .01. Note: Correlations below the diagonal are at the between-person level. Correlations above the diagonal are at the within-person level (calculated on the within-person centred variables to account for the non-independence of measures).

Hypotheses testing

As recommended by Zhang et al. (2009), we modelled our hypotheses at both the within-person and the between-person levels of analysis. Within-person effects refer to fluctuations within a person over time. Participants’ scores are compared with their own average across the days. Between-person effects refer to stable differences between persons, and participants’ scores are compared with the average of the sample. We report both levels, but as we are conducting a daily diary study, we are mainly interested in the within-person effects.

Regarding Hypothesis 1, as shown in Table 2 (within-person level effects), we found a negative and significant relationship between employees’ WFC in the evening and spouses’ home detachment on the next day (Estimate = -.092, p < .05). This provides support for Hypothesis 1. Moreover, Hypothesis 2 was supported because, as shown in Table 2, we found a significant and positive relationship between spouses’ home detachment and spouses’ job performance (same person, same-day effects, Estimate = .106, p < .001). Hypothesis 3 was also supported: spouses’ job performance in the afternoon was positively related to spouses’ work detachment in the evening (same person, same-day effects, Estimate = .247, p < .01). Finally, Hypothesis 4 stated that employees’ WFC would be negatively related to spouses’ work detachment on the next day (evening) through spouses’ low home detachment and job performance on the next day (afternoon). We tested this indirect effect but it was non-significant (Estimate = -.002, p = .09). In summary, at the within-person level of analysis, we found support for Hypotheses 1, 2 and 3, while the full mediation model was not supported.

Table 2. Multilevel estimates at the within-person level of analysis.

WITHIN-PERSON LEVEL MODEL	Estimate	S.E.	Est./S.E.
Employees’ WFC (evening) → Spouses’ home detachment on the next day (afternoon)	−0.092	0.042	−2.20*
Spouses’ home detachment (afternoon) → Spouses’ job performance (afternoon) (same person, same-day effects)	0.106	0.022	4.77***
Spouses’ job performance (afternoon) → Spouses’ work detachment (evening) (same person, same-day effects)	0.247	0.075	3.30**
Employees’ WFC (evening) → Spouses’ work detachment on the next day (evening)	0.031	0.035	0.90
Indirect within-person level effects
Employees’ WFC (evening) → Spouses’ home detachment (next day) → Spouses’ job performance (next day) → Spouses’ work detachment (next day)	−0.002	0.001	−1.67
Variances
Employees’ WFC (evening)	0.857	0.064	13.30***
Residual variances
Spouses’ work detachment on the next day (evening)	0.994	0.077	12.85***
Spouses’ home detachment on the next day (afternoon)	0.987	0.068	14.60***
Spouses’ job performance on the next day (afternoon)	0.318	0.027	11.75***

Note: N (observations) = 1450; n (participants) = 290. * p < .05, ** p < .01, *** p < .001.

At the between-person level of analysis, the only two significant relationships that we found referred to spouses’ variables. As shown in Table 3, spouses’ home detachment was positively related to spouses’ job performance (Estimate = .149, p < .001), and spouses’ job performance was positively related to spouses’ work detachment (Estimate = .550, p < .001). This means that those who have difficulties in detaching from home report lower levels of job performance and poor work detachment not only on a daily basis (compared with themselves) but also on average (compared with the rest of the sample). The crossover and the sequential indirect effects, however, were non-significant at the between-person level of analysis.

Table 3. Multilevel estimates at the between-person level of analysis.

BETWEEN-PERSON LEVEL MODEL	Estimate	S.E.	Est./S.E.
Employees’ WFC (evening) → Spouses’ home detachment on the next day (afternoon)	0.021	0.093	0.22
Spouses’ home detachment (afternoon) → Spouses’ job performance (afternoon) (same person, same-day effects)	0.149	0.038	3.91***
Spouses’ job performance (afternoon) →Spouses’ work detachment (evening) (same person, same-day effects)	0.550	0.113	4.89***
Employees’ WFC (evening) → Spouses’ work detachment on the next day (evening)	−0.118	0.094	−1.24
Indirect between-person level effects
Employees’ WFC (evening) → Spouses’ home detachment (next day) →Spouses’ job performance (next day)→ Spouses’ work detachment (next day)	0.002	0.008	0.23
Variances
Employees’ WFC (evening)	0.627	0.064	9.75***
Residual variances
Spouses’ work detachment on the next day (evening)	1.000	0.107	9.31***
Spouses’ home detachment on the next day (afternoon)	0.998	0.093	10.67***
Spouses’ job performance on the next day (afternoon)	0.351	0.043	8.22***

Note: N (observations) = 1450; n (participants) = 290. * p < .05; **p < .01, *** p < .001

Discussion

The objective of this daily diary study was to examine a sequential indirect effect in which employees’ experience of WFC affected spouses’ home detachment and job performance, which, in turn, related to lower work detachment of spouses. Based on the stressor–detachment model (Sonnentag & Fritz, 2015) and the cognitive activation model (Meurs & Perrewé, 2011; Ursin & Eriksen, 2010), we proposed that when employees experience high WFC in the evening, on the next day their spouses remain cognitively activated with family-related thoughts, which, in turn, affects their performance at work. We found evidence of this crossover effect. These findings are in line with previous studies showing that cognitive aspects related to the private domain affect performance. For example, Nohe et al. (2015) found that daily FWC related to lower performance via a lack of concentration. Additionally, our findings are in line with Kahn’s (1990) proposition regarding the investment of energy. If cognitive energy is spent on thinking about home-related issues, it cannot be directed towards a task, thereby affecting performance. Therefore, crossover via cognitive mechanisms is possible, and it is not only social aspects such as undermining or an actual increase in home demands that can explain the crossover of experiences between couples (Westman, 2001). We consider this to be a key contribution of our study and call for more studies to examine the antecedents and outcomes of home detachment in the context of WFC among couples. The consequences of WFC for spouses have lasting effects, and when the partner is no longer present, spouses cannot disconnect from the implications derived from this situation. Remaining cognitively activated makes the experience of the partner’s WFC more proximal. Therefore, not only work-related attitudes (Li et al., 2021) but also cognitions (i.e. detachment) and work-related behaviours (i.e. job performance) are linked to the focal employee’s WFC.

The second theoretical implication of this study relates to the spillover effects, that is, how spouses’ low home detachment and performance spill over in the form of low work detachment. This is relevant from a theoretical perspective and advances recovery literature by showing that employees enter into a daily cycle of “a lack of detachment”, being unable to focus on the appropriate task at the appropriate time. More specifically, we found that these two types of detachment are linked via job performance. This is in line with the stressor–detachment model (Sonnentag & Fritz, 2007). When individuals do not perform well, they remain cognitively activated and think about work (instead of engaging in other activities to disconnect psychologically). This is an important contribution to the literature on recovery – being able to detach at work is related to the ability to detach at home on a daily basis, not directly, but rather indirectly (via job performance). This means that the relationship between these two types of detachment is complex and depends on different mechanisms that should be further explored in future studies. For example, apart from reduced job performance, it is possible that employees who do not disconnect from home have more conflicts with colleagues or their supervisor, which affects their ability to detach from work during the evening. In this line, Sanz-Vergel et al. (2015) found that when family problems affect work, employees report more conflicts with colleagues at work and with spouses at home. Furthermore, it might be that employees who do not disconnect at work and do not perform well engage in work-related tasks during the evening so as to compensate for their low performance, thereby impairing the ability to disconnect. We encourage scholars to further examine the link between these two types of detachment as well as ways of breaking such “lack of detachment” cycles. For example, support from colleagues could buffer the negative association between a lack of home detachment and performance. Positive activation could also facilitate detachment from home and then promote engagement and job performance at work. Although previous studies have examined the relationship between positive affect and work detachment (e.g. Sonnentag et al., 2008), its impact on home detachment in particular has not been analysed.

Limitations and suggestions for future research

This study presents some limitations that could be overcome in future studies. The first limitation relates to the tested model. As mentioned above, there are additional explanatory mechanisms which may add complexity and help us to further understand the whole process of recovery. Although the concepts of psychological detachment and rumination may overlap (Sonnentag & Fritz, 2015), future studies could include rumination (Cropley & Purvis, 2003) instead of detachment to find out whether results are similar and to disentangle the possible differences between these two concepts.

Secondly, most of the empirical evidence on recovery is gathered through a five-day diary design (e.g. Germeys & De Gieter, 2018; Sonnentag et al., 2008; Volman et al., 2013). However, in their review on 20 years of recovery, Sonnentag et al. (2017) pointed out that recovery can follow different trajectories, and scholars should pay more attention to the variable time, exploring optimal temporal patterns of daily recovery. From this perspective, one could argue that detaching from home may be necessary only during specific episodes at work, and employees can try to solve any family-related matter during work breaks.

Thirdly, objective ratings of performance would help us to find out whether colleagues and supervisors also perceive the same. Although self-reported performance indicators are generally significantly correlated with objective measures of performance (Fletcher et al., 2008; Wall et al., 2004), future studies should consider objective performance indicators as outcomes.

Fourthly, home detachment and job performance were both measured in the afternoon because both take place in the work domain, so in order to reduce retrospective bias, participants have to respond immediately after the working day has finished. Future studies could overcome this issue by using designs such as event sampling, in which the responses are directly related to a particular period of time or a particular event (see Ohly et al., 2010 for a review).

Finally, we used paper-and-pencil diaries as a method of data collection. This method has several benefits, such as the fact that it is an inclusive method (Ohly et al., 2010). This means that individuals with no access to internet or those lacking technological skills can participate. However, one of the potential limitations of this method is that participants may not answer the booklet at the time required. Although we asked employees to specify the time at which they were filling in the survey, it may be more convenient to use electronic devices so that the timing of responses is automatically registered. Moreover, Ohly et al. (2010) emphasise that by using web-based surveys, data are already in the computer, and resources for data entry are spared.

Practical implications

These findings have implications for managers and policymakers. Job performance depends not only on job characteristics per se, but also on how these characteristics affect family life. When employees experience a conflict between these two important domains, spouses also suffer consequences in the work domain. In this respect, adequate support at work becomes crucial. Work–family policies need to be in place in order to help employees and their families to concentrate on their responsibilities at the appropriate time (Garraio et al., 2022).

Moreover, organisations should aim at providing training that helps employees to be engaged at work and detached at home (Sonnentag et al., 2008). The traditional interventions on work detachment have proven to be effective (Karabinski et al., 2021) and could be adjusted to improve home detachment too. In this line, being mindful both at work and at home has been proposed as a useful recommendation (Haun et al., 2018). However, not every person has the ability to totally separate work and family, detach, and be mindful at the appropriate time. Organisations would benefit from training programmes focused on these cognitive aspects. For example, Althammer et al. (2021) designed an intervention based on mindfulness as a cognitive-emotional segmentation strategy and found positive effects on psychological detachment, psychological WFC, and work–life balance satisfaction. These types of interventions will help employees to disconnect from a specific domain when necessary, as mindfulness provides individuals with skills with which to focus on the present moment. What is more, mindfulness facilitates the reduction of negative emotions while increasing positive emotions (Li et al., 2023; Mäkiniemi & Heikkilä-Tammi, 2018).

In the same vein, tailored couple-oriented interventions have proven to be effective for couples’ life satisfaction (Schaer et al., 2008). The objective of this type of programme is to provide couples with strategies with which to cope together with stress. We propose that these couple-oriented interventions should incorporate training on psychological detachment and mindfulness as essential coping strategies in the context of work–family balance. In addition, it has been demonstrated that engaging in absorbing activities together, as a couple, can also help to foster detachment (Hahn et al., 2012).

Finally, individuals may cope with conflicts between work and private life by segmenting life domains. Behavioural segmentation strategies, such as avoiding the use of work-related communication technology at home, have proven to help employees to detach and recover from work demands (Park et al., 2011). Therefore, organisations should share these tips with employees and encourage them to follow them as part of a comprehensive wellbeing organisational strategy.

Conclusion

All in all, our study sheds light on crossover and spillover processes, with a lack of home detachment and job performance as key mechanisms. This extends the stressor–detachment model (Sonnentag & Fritz, 2015) by showing that a) stressors affect not only work detachment but also home detachment, b) stressors affecting detachment not only are one’s own job demands but also can be experiences lived by one’s partner (i.e. employee’s WFC as a stressor affecting the spouse), and c) there are additional crossover processes beyond the traditional “stressor–lack of detachment–wellbeing” cycle. We encourage scholars to further explore how employees’ WFC affects spouses beyond the family domain.

Disclosure statement

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