The Impact of Telework on Conflict between Work and Family: A Meta-Analytic Investigation

James Kunz, MA, MS, Colorado State University; Hannah Finch, BS, Colorado State University; Joshua Prasad, PhD, Colorado State University; Kiplin Kaldahl, MS, NORC at the University of Chicago

Purpose/Objectives
In this paper, we aim to conduct a meta-analysis evaluating the empirical evidence linking telework and work-family conflict (WFC). We hypothesize a negative relationship between telework and WFC as depicted in previous works (Allen et al., 2013; Gajendran & Harrison, 2007). Telework should be associated with reduced work-interference-with-family (WIF) and family-interference-with-work (FIW). We hypothesize these relationships will vary based on gender, telework measurement, and national sample.

Background
Telework is a flexible work arrangement (FWA) where individuals work away from their central organization, typically from home (Allen et al., 2015). There is a common assumption telework should be beneficial for managing one’s work and non-work domains due to the flexibility telework affords (Gajendran & Harrison, 2007). Gajendran and Harrison (2007) evaluated the mediating effect of WFC in the relationship between telework and work-related outcomes and found small, but significant beneficial effects of telework on WFC. Allen et al. (2013) also found beneficial small, significant effects between FWAs and WIF, and insignificant effects between FWA and FIW.

Nonetheless, a meta-analysis investigating the telework/WFC relationship has not been published since 2007 (Gajendran & Harrison, 2007). Further, this work did not distinguish between WIF and FIW, despite the consideration of these as separate constructs within WFC research (e.g., Allen et al., 2013; Frone, Russell, & Cooper, 1992). With the increase in teleworking both prior and as a result of the COVID-19 pandemic (Global Workplace Analytics, 2020), there is likely more to be discovered about this relationship through recent studies and potential moderators. Gender may moderate the telework/WFC relationship as women and men may have differential reasons for teleworking (Mann & Holdsworth, 2003; Olson & Primps, 1984). Further, given the inconsistent measurement and conceptualization of telework across literatures (Allen et al., 2015; Beckel & Fisher, 2021), many results may depend on measurement approach. Finally, national sample might change the telework/WFC relationship due to reduced accessibility of work-family practices in the United States compared to other nations (Earle, Mokomane & Heymann, 2011). As COVID-19 declines, telework is likely to remain. This work is necessary to clarify our understanding of how telework relates to WFC.

Methods
We conducted a literature search using MetaBus (Bosco et al. 2015) and construct keywords gathered from previous work on telework and/or work-life conflict. Our search yielded 241 articles. We excluded studies that did not mention telework and work-life conflict, or did not have a relevant effect size, resulting in 30 articles. Two independent coders reviewed each study, and corrected discrepancies (κ> .80).

We followed Schmidt and Hunter’s (2015) random-effects meta-analytic methods to analyze the studies obtained. This involved calculating true population correlations (𝑝ˆ) that corrected for sampling error and measurement error in FIW and WIF outcomes, using artifact-distribution methods to estimate unreported outcome reliabilities.

We did not obtain enough published studies to examine moderators of the telework/FIW relationship. We anticipate obtaining enough unpublished studies to conduct these analyses. Time for reanalysis should be minimal as all syntax has been prepared except for planned publication bias analyses.

Findings
See Table 1 for results. The observed relationship between telework and WIF is negative and exhibits substantial heterogeneity. Conversely, the relationship between telework and FIW is positive, though nonsignificant, with meaningful heterogeneity. Gender moderated the relationship between telework and WIF. Majority Female samples exhibited a positive, yet nonsignificant relationship, whereas majority Male samples produced a significantly negative relationship. Further evaluation of heterogeneity reveals substantial variability among Female samples, suggesting additional moderators. Effects were more consistent among males. Measurement also impacted study findings. Continuous measures of telework exhibited a weak relationship with WIF and substantial effect size heterogeneity, whereas dichotomous measures yielded a negative relationship that was consistent. Nationality of sample did not impact the relationship between telework and WIF. US samples and non-US samples exhibited comparable results to the overall meta-analytic findings for WIF.

Discussion
Our results highlight the nuanced relationship between telework and work-family conflict. There is a persistent and beneficial effect of telework on WIF. However, our results differ from prior work by way of a positive, non-significant relationship between telework and FIW.

The conducted moderator analyses help to expand on the field’s current understanding of telework and work-family outcomes. Gender was found to have a slight positive, non-significant, effect on WIF for females, and a negative effect for males. This difference may be explained through findings that women are more likely to engage in telework to attend to family demands (Mann & Holdsworth, 2003). Further, our analyses underscore how variation in telework measurement methods forces different interpretations of results. We urge future research to align measurement methods with theory so the operationalization of telework captures its nuance rather than just serve as a proxy for good workplaces.

Conclusions
Gajendran and Harrison (2007) lamented the fragmented, conflicting telework literature at the time of their meta-analysis. The literature remains challenging to navigate, yet our meta-analysis reveals that outcome operationalization, gender, and measurement methods help clarify why conflicting results persist. We propose that focus and consensus around these points will promote the productivity of the telework literature, especially regarding the impact of telework on work-family conflict.