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

Introduction

Although telework has been studied since the 1970s (Allen, Golden, & Shockley, 2015), the Covid-19 pandemic brought unprecedented challenges for at least 8.6 million Brazilians (IBGE, 2020). Telework was done initially full-time, regardless of the home infrastructure or people's previous experience. Schools were closed and the recommendation was not to leave home or receive other people. In March 2020, several Brazilian cities decreed quarantine, aiming at social isolation to stop the virus (OMS, 2020).

The spatial and temporal boundaries between work and nonwork were eliminated for those who could telework at home. People's private lives were possibly invaded, and Brazilian laws are very ambiguous when it comes to overtime worked at home (Pereira, Barbosa, & Saraiva, 2021).

With economic uncertainty enhanced by the pandemic (unemployment rose to 14.4% in 2020 (IBGE, 2020)), many people tried to be the ideal worker, who prioritizes professional life (Bailyn, 2006). But at which personal cost, not dedicating themselves as much as they would like to which nonwork dimensions? It is necessary to deepen the academic understanding of the work-nonwork conflict (WNWC) for teleworkers, recognizing it as a dilemma that affects people, organizations and society systemically (Bailyn, 2006).

Although nonwork has been treated as a synonym for family in studies about work and nonwork since the 1980s (Greenhaus & Beutell, 1985), it is necessary to expand nonwork besides family because people dedicate themselves to other dimensions too, such (Keeney, Boyd, Sinha, Westring, & Ryan, 2013; Oliveira, 2017). Only two studies were found about WNWC (Keeney et al., 2013; Oliveira, 2017), not focused on teleworkers. It enhances the need to analyze whether the time spent teleworking at home interferes with the desired dedication to different nonwork dimensions, as well as whether there are telework or teleworkers' characteristics that contribute to this interference.

Brazilian teleworkers increased to 8.6m people in May 2020 due to the pandemic (IBGE, 2020), and should keep growing due to organizations' financial advantages, mainly by reducing accommodation costs (Daniels, Lamond, & Standen, 2001). This context emphasizes the need to contribute academically, expanding the investigation from work-family conflict (WFC) (Greenhaus & Beutell, 1985) to WNWC (Keeney et al., 2013; Oliveira, 2017) for teleworkers. This step is essential to support managers in improving teleworkers' well-being and, consequently, society's well-being. People are more productive and satisfied when they can dedicate themselves to different nonwork dimensions (Bailyn, 2006; Hirschi, Herrmann, Nagy, & Spurk, 2016)--and telework at home can make this dedication difficult. Concerning the Covid-19 pandemic, three studies explored only WFC and indicated difficulties to reconcile different roles at home (Lemos, Barbosa, & Monzato, 2021; Pandini and dos Santos Pereira, 2020; Leite & Lemos, 2020).

Due to the lack of WNWC studies for teleworkers (Oliveira & Pantoja, 2020), to fill the gap described, the problem presented is: "Which characteristics of telework and teleworkers are associated with the work-nonwork conflict (WNWC)?". To answer this question, the article's objective is to analyze the relationship between telework and teleworkers' characteristics and the WNWC, investigating time spent in the nonwork dimensions and the dimensions more affected in the WNWC.

Theoretical background

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 (Nakrosiene, Buciuniene, & Gostautaite, 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, & Martinson, 2003) and WFC (Russell, O'Connell, & McGinnity, 2009). So, it can also consume time desired for other nonwork dimensions, originating the first hypothesis:

HI. For teleworkers, there are other nonwork dimensions in the WNWC which are affected as much as family.

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:

H2a-h. Telework intensity is negatively associated with WNWC.

H3a-h. Previous telework experience is negatively associated with WNWC.

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...