Telework during the Covid-19 pandemic and the work-nonwork conflict

2.1 Telework

Allen et al. (2015) attest that different definitions of telework exist because it is of interest to different areas, such as transport, information systems and management. They rest on conceptualizations widely adopted to define telework as:

a work practice that involves members of an organization substituting a portion of their typical work hours […] to work away from a central workplace—typically principally from home—using technology to interact with others as needed to conduct work tasks (Allen et al., 2015, p. 44).

In this definition, four characteristics stand out and will be considered in the theoretical model supporting this article:

1. space flexibility: part of the hours that would be worked at the main office is worked outside the office (it is not overtime);

2. telework intensity: portion of hours worked in telework;

3. telework takes place mainly in the worker's home;

4. use of technologies for interaction among teleworkers.

Messenger and Gschwind (2016) consider the use of these technologies as a milestone for 21st-century telework, increasing time and space flexibilities. These flexibilities, in turn, can confound the work-nonwork boundaries (Raghuram & Wiesenfeld, 2004) and increase WNWC (Keeney et al., 2013), as will be explored later (subchapter 2.3).

A fifth characteristic is time flexibility, as telework can be done beyond the hours typically allocated to work. It is beneficial to teleworkers when it means autonomy to fulfill their tasks (Nakrošienė, Bučiūnienė, & Goštautaitė, 2019), but harmful if it means overtime or invades nonwork moments (Kossek, Lautsch, & Eaton, 2006).

2.2 Work-nonwork conflict

WNWC is defined as “[…] difficulty participating in nonwork domains due to participation in the work domain” (Keeney et al., 2013, p. 221). It is an expansion from WFC (Greenhaus & Beutell, 1985), as they have the same conceptual basis of role conflict (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964), but WNWC includes other nonwork roles. The eight resulting nonwork dimensions are health, family, household management, friendships, education, romantic relationships, community involvement and leisure.

Validating the WNWC scale, Keeney et al. (2013) replicated the WFC dominant framework, whose main sources are time (worked hours) and stress (Greenhaus & Beutell, 1985), and proposed a multidimensional construct with 16 factors – 8 nonwork dimensions and 2 sources. This model was also validated for a Brazilian sample (Oliveira, 2017) and it is partially illustrated in Figure 1.

Identifying which other nonwork dimensions besides family the time spent teleworking influences is essential to develop more policies and practices for preventing and combating WNWC among teleworkers. Interests change from person to person (Bailyn, 2006) as shown by studies about workers (not teleworkers yet): the North Americans allocated more time to household management, leisure and health (Keeney et al., 2013); the Brazilians, to education, family and romantic relationship (Oliveira, 2017).

Telework has been positively associated with the hours worked (Hill, Ferris, & Märtinson, 2003) and WFC (Russell, O’Connell, & McGinnity, 2009). So, it can also consume time desired for other nonwork dimensions, originating the first hypothesis:

2.3 Telework and WNWC

Telework is theorized to implement time and space flexibility policies to help workers manage work and nonwork responsibilities. Empirical evidence, however, shows a weak association specifically between telework and lower WFC (Allen et al., 2015). The WFC is just one of the WNWC factors, but the only one found in quantitative telework studies, the reason why WFC predominates in this theoretical subchapter.

As the objective of this article is to expand the investigation from the WFC to the WNWC for teleworkers, the hypotheses presented below consider time-originated WNWC and its eight factors (Figure 1). Therefore, for hypotheses 2–10, the investigation considered 8 derived hypotheses. For example, for hypothesis 2, we investigated from “H2a: Telework intensity is negatively associated with the work-health conflict” to “H2h: Telework intensity is negatively associated with the work-leisure conflict”.

The weak association between telework and lower WFC was found in two meta-analyses of 46 and 58 studies. In the first meta-analysis, Gajendran and Harrison (2007) investigated possible moderating variables and found that, when the intensity was high or there was more experience in this format, the WFC was lower. These results ground hypotheses 2 and 3:

In the second meta-analysis, Allen, Johnson, Kiburz, and Shockley (2013) focused on the effects of time and space flexibilities on the WFC, discerning whether they were available or effectively used. The negative association was low and only between: (a) availability of time flexibility and WFC, and (b) use of space flexibility and WFC. Due to the low effects, the authors hypothesized that telework would increase the risk of confounding boundaries between work and family, a perception already verified in Hill, Miller, Weiner, and Colihan (1998) and Hill et al. (2003).

This confusion between boundaries may be unwanted, as these boundaries are the main factors workers can manage to reduce the WFC (Ashforth, Kreiner, & Fugate, 2000). According to the boundary theory, people prefer to exercise their roles considering some point between total integration (no distinction between what belongs to work or family or when or where each role manifests itself) and total segmentation (the two dimensions as separate worlds, that occur in distinct time and space) (Nippert-Eng, 1996). Furthermore, if the work environment demands the opposite of what the worker prefers, there is a greater chance of conflict.

Convergence between worker preference and environment demands is decisive, as more integration between roles can both reduce the conflict, by allowing an easier transition between dimensions, or increase it, by confounding the most salient role for the individual at any given time. The opposite occurs when there is more segmentation (Ashforth et al., 2000): it increases the conflict by making the transition difficult or decreases it by explicitly expressing the priority at each moment.

Thus, as telework tends to enhance the integration between roles, teleworkers who prefer more segmentation may have a higher perception of WNWC, which generates the fourth hypothesis:

Time and space flexibilities, telework characteristics, can affect teleworkers’ mental health positively, by allowing them to balance working hours with family demands, or negatively, by causing conflicts between work and family demands (Kattenbach, Demerouti, & Nachreiner, 2010; Johnson et al., 2020), So, it is important to investigate whether these and other telework-related characteristics are associated with WNWC. Hornung, Rousseau, and Glaser (2008) found that time flexibility was negatively related to WFC. Regarding space flexibility, while Hill et al. (1998) found specific places associated with more hours worked, which could result in higher WFC, Solís (2016) found that such places were related to lower WFC. Thus, two other hypotheses will be investigated:

Several studies indicate that social support works better at combating WFC than formal organizational practices (Anderson, Coffey, & Byerly, 2002; Oliveira, Cavazotte, & Paciello, 2013), even for teleworkers. Bentley et al. (2016) found an association between manager support and lower psychological stress. These studies support the following hypotheses:

It is also important to examine teleworkers’ characteristics, starting with sex. Telework can impact household and people care, especially for women, as they usually spend more hours per week on household tasks than men. In Brazil, before the Covid-19 pandemic, women spent 21.4 hours on these tasks, while men spent 11.0 hours (IBGE, 2019), even though they (women) have participated increasingly in the labor market for a long time. Only from 2012 to 2019, there was a 2.9 pp increase, from 51.6% to 54.5% (IBGE, 2019).

The sexual division of labor stems from men being responsible for working outside the home and ensuring financial support for the family, while women work at home, taking care of people and domestic chores (Hirata & Kergoat, 2003). Even when the assumptions for such a division no longer exist, with both sexes working outside the home, women still seem to be considered the main responsible for domestic tasks. A study by Oliveira and Casado (2021) found women with more WNWC than men for all eight nonwork dimensions, especially if they were married or had no children. This sexual division of labor can increase in telework (Ellison, 1999), which motivates the following hypotheses:

Investigation concerning telework and life stages suggested that children were a key component, increasing WFC (Zhang, Moeckel, Moreno, Shuai, & Gao, 2020). Hilbrecht, Shaw, Johnson, and Andrey (2008) explored the experience of school-age children’s mothers and found tensions as working hours followed children's school rhythm, with no time for leisure or personal care. They also expressed living the sexual division of labor and being responsible for household and child care. Russell et al. (2009) found similar results: a positive relationship between telework at home and WFC, especially for mothers of under 5-year-olds. These results generate the following hypotheses:

Solís (2016) discovered that having more personal responsibilities, such as taking care of an elderly person, was associated with higher WFC. One last group of hypotheses, then, should be investigated:

Investigating telework during the Covid-19 pandemic, Nguyen and Armoogum (2021) found middle-aged women less likely to prefer home-based telework than younger women. Having more household and children responsibilities was the most probable reason. Çoban (2022) encountered women expressing a greater risk of being detached from professional work and of consolidating their role as housewives.

Shirmohammadi, Au, and Beigi (2022, p. 1) analyzed 40 studies concerning work-life balance for teleworkers during the pandemic. They found four misfits between desirable expectations and undesirable realities: “(1) flextime versus work intensity, (2) flexplace versus space limitation, (3) technologically-feasible work arrangements versus technostress and isolation, and (4) family-friendly work arrangement versus housework and care intensity”.

There are recent Brazilian studies with teleworkers whose results brought WFC up. Filardi, Castro, and Zanini (2020) discovered perceptions of low WFC and higher interaction with the family, although domestic chores interfered with telework. During the pandemic, the studies reported: (1) women with varied perceptions about WFC, some mentioning work overload, some highlighting time gained for other activities (Lemos et al., 2021); (2) difficulties to reconcile different roles at home and how positive time flexibility was (Pandini and dos Santos Pereira, 2020); (3) WFC as one of the disadvantages, in addition to the distraction caused by household chores (Leite & Lemos, 2020).

At Latin America’s biggest university, São Paulo University, professors conducted two studies, whose results are comparable due to the similar samples: balance between men and women, most respondents married and with a postgraduate degree, an average income of United States (US)$ 1,700.0 [1]. From the first to the second survey, people’s desire to keep teleworking increased from 70% to 78%, as well as their perception about sharing tasks to take care of the family while teleworking or telework being compatible with their physical health care, as shown in Table 1 (PROGEP, 2020, 2021).

Figure 1 shows the theoretical model that supports the hypotheses investigated.

4.1 Characteristics of the sample

The sample includes 56.9% women and 43.1% men, with an average age of 42 years (standard deviation (SD) = 10), concentrated in the Brazilian Southeast (97.9%), and mostly married (68.6%). There is a similar number of respondents with and without children (respectively, 52.5% and 47.5%) and 33.8% (of all respondents) have children until 14 years old. As the focus is on investigating the whole WNWC (its 8 factors), and not just the WFC, it is interesting to have diverse family configurations in the sample.

Concerning the respondents’ professional occupations, they are mostly specialists, analysts, assistants (26.1%), managers (18.4%) or autonomous (15.4%). They work 46.1 hours per week on average (SD = 12.5) and 40.2 hours (SD = 15.4) in telework, with no statistical difference between men and women. Almost a third of the respondents (31.8%) were studying at the time of the survey.

On telework, the average intensity is 87.2%; 47.2% of respondents had previous experience with this format before 2020, 53.5% have a specific place at home to telework, and 63.2% work for companies with telework policies. Everyone uses a computer for telework. During telework, 15.1% are alone at home, 56.9% have the company of their spouses, 43.1% of their children, 8.4% of elderly people under their care and 22.7% of someone who does the household tasks.

Regarding their income, the average is US$ 2,139.0 [2], with greater concentration in the ranges between US$ 813.1 and US$ 1,219.4 (13.7%), US$ 1,219.6 and US$ 1,625.8 (13.4%), and above US$ 3,658.1 (21.7%), which makes the sample a very specific portion of Brazilian population, whose average income in 2019 was US$ 448.9 (in the Southeast, US$ 515.4) (IBGE, 2019).

4.2 Time allocation to nonwork dimensions and the more affected ones in WNWC

Family, household management, and education are the three nonwork dimensions to which teleworkers allocate more time (4.96, 4.53, and 4.52, respectively – Table 3), while leisure (3.71) and community involvement (2.38) have less time spent. It is important to point out that previous experience did not generate statistical differences in the time allocated to family and household management. Only having children up to 14 years old or at home or being married did.

Compared to what North American and Brazilian workers declared (Keeney et al., 2013; Oliveira, 2017), in this study, the two dimensions related to the home environment stand out by figuring, together, among the three to which teleworkers dedicate more time (Table 3). This could come from the Covid-19 scenario, as people had to adapt abruptly to telework at home and take care of the family and the house at the same time. Brazilian teleworkers had highlighted performing domestic chores even before the pandemic (Filardi et al., 2020) – and just continued in the Covid-19 scenario (Leite & Lemos, 2020; Pandini and dos Santos Pereira, 2020; PROGEP, 2020, 2021).

Another highlight: all dimensions have time allocation varying among respondents (coefficient of variation (CV) from 28% to 56%). It reinforces the multiplicity of interests to which teleworkers dedicate themselves in addition to their jobs and families and, consequently, the need to study nonwork besides the family dimension.

As for the nonwork dimensions more affected in the WNWC, leisure, family and education stand out (4.63, 4.34 and 4.32, respectively – Table 4), which led to the acceptance of hypothesis 1. This result strengthens the possibility that telework invades leisure time (Kossek et al., 2006) and does not help to manage work and family demands (Allen et al., 2015), at least not as intensely as desired (WFC is the second most intense factor). Similar perceptions appeared in Pandini and dos Santos Pereira (2020), since people declared difficulties to reconcile different roles at home, and in one of the misfits found by Shirmohammadi et al. (2022): family-friendly work arrangements versus care intensity.

As happened with North American and Brazilian workers (Keeney et al., 2013; Oliveira, 2017), the nonwork dimensions affected vary among respondents, as well as within each dimension (CV from 32% to 42%; Table 4). This result yet again emphasizes the need to study nonwork besides the family dimension.

Just like in Oliveira (2017), leisure figures as a key dimension because it is the second with less time allocated, the most affected in WNWC, with correlations between allocated time (low correlations) and all eight WNWC factors (Table 5). In other words, allocating less time to leisure is associated with a more intense general WNWC perception for teleworkers. It seems that when teleworkers spend less time at leisure, they feel like not being able to dedicate themselves as much as they want to the other seven nonwork dimensions too. That might happen because leisure is the freest of the nonwork dimensions, the one that can be customized to what each person wants to do.

Friendships and health are the other dimensions whose allocated times are correlated with more WNWC factors: seven and six, respectively (Table 5).

4.3 Telework characteristics and WNWC

Concerning telework characteristics, the average intensity in the sample is 87.2% and teleworkers report a slightly higher perception of manager support (average = 4.99 and SD = 1.57) than of time flexibility (mean = 4.57 and SD = 1.81). This last characteristic was also slightly less mentioned by PROGEP (2020, 2021) respondents from 2020 to 2021 (62% versus 59%).

Investigating the association between these three telework characteristics and WNWC (Table 6), time flexibility stands out due to a negative (low) correlation with all eight WNWC factors. There is lower WNWC among teleworkers who declare higher time flexibility (Table 6), which leads to the acceptance of hypotheses 5a–h. These results converge with Pandini and dos Santos Pereira (2020), whose respondents pointed out time flexibility as an advantage, overcoming the misfit “flextime versus work intensity” (Shirmohammadi et al., 2022). It indicates that when people work in their own homes, having some autonomy to decide on their time becomes especially important.

As manager support is negatively associated (low correlation) with six WNWC factors (Table 6), hypotheses 7a, c, d, e, g, and h are also accepted. It reinforces the importance of social support to prevent WNWC (Anderson et al., 2002; Oliveira et al., 2013). Finally, hypotheses 2a–h were rejected, as there is no association between telework intensity and WNWC (Table 6).

In other words, the percentage of hours spent teleworking does not affect the surveyed teleworkers’ WNWC, but the perception of manager support and time flexibility does so. Regarding only time flexibility, it does not seem to confound boundaries, differently from what was found in Hill et al. (1998, 2003).

Having a specific place in the house to telework is not related to WNWC intensity. Thus, hypothesis 6 is refuted. A similar thing happens to having previous experience with telework, which is associated with only two factors, household management and community involvement. So, only hypotheses 3c and g are accepted.

4.4 Teleworkers’ characteristics and WNWC

Regarding the preferred boundary management strategy, the mean indicates a balance between integration and segmentation (mean = 4.02 and SD = 1.05). This characteristic stands out for presenting a positive (low) correlation with all eight WNWC factors (Table 7). It indicates a higher perception of this conflict for teleworkers who prefer segmentation, leading to the acceptance of hypotheses 4a–h. This result strengthens the role of individual preference for a specific boundary management strategy. Such preference is fundamental for the positive or negative meaning attributed to greater permeability between work and nonwork, an intrinsic characteristic of telework at home. So, people who prefer to separate work from nonwork tend to find it harder to telework at home.

Characteristics such as teleworkers’ sex, having elderly people under their care, and, specifically for women, having children up to 14 years old, are not associated with different levels of WNWC. These results lead to refuting hypotheses 8a–h, 9a–h, and 10a–h.

Complementing these final analyses, there is no difference in the WNWC factors either between mothers and fathers of children up to 14 years old. Focusing specifically on the WFC and comparing people with and without children up to 14 years old, there is a difference for men. The fathers declare higher WFC when compared to the other men (4.90 versus 3.98). These results differ from Nguyen and Armoogum (2021) and Çoban (2022), as their female respondents were less likely to telework or feared consolidating their role as housewives, probably because of children and household responsibilities. This difference may be a consequence of the sexual division of labor, when women are held responsible for domestic tasks and incorporate this role, finding it difficult to admit that work can interfere negatively.

Table 8 summarizes the results found.