key: cord-0952090-llhk0iew authors: Schieman, Scott; Badawy, Philip; Hill, Daniel title: Did perceptions of supportive work–life culture change during the COVID‐19 pandemic? date: 2022-02-11 journal: J Marriage Fam DOI: 10.1111/jomf.12826 sha: eaf05a6edf806c495f70412c68ff2fdb6180610d doc_id: 952090 cord_uid: llhk0iew OBJECTIVE: This article examines whether perceptions of supportive work–life culture changed during the COVID‐19 pandemic—and if that depended on (1) working from home; (2) children in the household; and (3) professional status. We test for gender differences across the analyses. BACKGROUND: During normal times, the “ideal worker” is expected to prioritize the demands of their job and is penalized for attending to family/personal needs while on company time. But the organization and expectations of roles might have changed due to the COVID‐19 pandemic. Organizations could have become more empathic or reinforced norms about single‐minded devotion to work. METHOD: In September 2019, we collected data from a national sample of Canadian workers. Then, during a pivotal period of shocks to the economy and social life, we re‐interviewed these participants in June 2020. RESULTS: We discovered that overall perceptions of work–life culture became more positive. However, subgroup differences revealed this positive change was muted among employees: (1) who worked from home; (2) with children under age 6 at home; and (3) in professional occupations. We found no subgroup differences by gender. CONCLUSION: Our findings address speculation about whether employees perceived their employers as becoming more supportive of work–life fit early in the pandemic. Future research should determine (a) longer‐term change in work–life culture during and after the pandemic; and (b) whether the actual benefits of supportive work–life culture also changed or if it was “window dressing.” This direction suggests it should have more strongly reduced work–life conflict as the pandemic unfolded. Work and family are "greedy institutions" that demand full commitment and undivided attention (Coser, 1974) . The "ideal worker," however, is expected to prioritize the demands of their job over family needs (Williams et al., 2013) . The ideal worker confronts trade-offs between advancing in their job or directing time, energy, and attention to their family or personal life. These "competing devotions," as Blair-Loy (2003) observes, create a tension between work and nonwork spheres that underlines requirements for a singular focus on work, potentially at the expense of other roles. When employees perceive penalties for putting family or personal needs ahead of their jobs-violating ideal worker norms-this represents what some call "flexibility bias/stigma" (Cech & Blair-Loy, 2014; O'Connor & Cech, 2018) . By contrast, a workplace culture that rejects these norms is likely to be perceived as supportive of employees' desires and efforts to achieve better work-life fit. Moreover, the perceived supportiveness of work-life culture is consequential because of its consistent link to lower work-life conflict (Mennino et al., 2005; Thompson & Prottas, 2005; Voydanoff, 2005) . The abrupt societal changes brought on by the coronavirus pandemic 2019 (COVID-19) generated widespread role restructuring (Government of Canada, 2020). Among employees who worked from home, "company time" and "family time" amalgamated in unprecedented ways. And for those logistically unable to work from home, challenges related to family needs likely also escalated because of restricted access to daycare and schooling. In the present study, we ask: During the early months of the pandemic, did employees' perceptions of supportive work-life culture change? Three elements are relevant: (1) expectations that work be prioritized over family/personal life; (2) penalties for attending to family/personal life; and (3) employer sensitivity to family roles (Thompson et al., 1999) . Focusing on these features of work-life culture, we develop two scenarios: the enhanced resource versus disrupted resource hypotheses. The enhanced resource view posits that dramatic changes early in the pandemic-especially in work arrangements and family needs-compelled employers to adopt a more empathic and flexible mindset (Caprino, 2020) . Structural and cultural shifts might have relaxed ideal worker norms. This scenario predicts an increase in the perceived supportiveness of work-life culture. Alternatively, if the "greedy institution" of work is truly greedy, even a pandemic will not soften its demands. A durable greedy institution compels that-especially during uncertain timesemployers expect adherence to ideal worker norms. This disrupted resource view predicts stability or even a decrease in perceived supportiveness of work-life culture. Within this framework, we also assess three contingencies: (1) working from home; (2) children at home; and (3) professional status. And, given women's and men's divergent role experiences as the pandemic unfolded (e.g., Landivar et al., 2020; Qian & Fuller, 2020) , we also assess gender variations at each stage. To test these ideas, we analyze data from the Canadian Quality of Work and Economic Life Study (C-QWELS). In September 2019, months before the first official COVID-19 case was reported, we collected data from a representative sample of Canadian workers. Then in June of 2020, during a pivotal period of economic shocks and social restrictions, we reinterviewed these same study participants. Defining a supportive work-life culture Role theory is a long-standing theoretical perspective to understand the dynamics between work and nonwork life (Kahn et al., 1964) . The overarching proposition is that multiple roles (e.g., work, family) have the potential to produce competing expectations and obligations. Border theory sharpens the focus on the border between work and nonwork life, with the latter predominantly reflected in family-related care. A key assumption is that workers are motivated to achieve work-life fit, which implies high role functioning and low conflict (Clark, 2000) . While multiple role demands can elevate inter-role strain, macrolevel forces potentially alter role configurations and norms that change the quality of those demands-and resources one might use to cope (Been et al., 2016) . A supportive work-life culture represents one of those resources (Allen, 2001 ). Thompson et al. (1999) define supportive work-life culture as "the shared assumptions, beliefs, and values regarding the extent to which an organization supports and values the integration of employees' work and family lives" (p. 394). Scholarly interpretations based on this definition imply that workers who perceive a supportive work-family culture reject the idea that employees at their workplace who prioritize family needs over work needs are viewed unfavorably (O'Connor & Cech, 2018) . A supportive work-family culture does not compel employees to choose between advancing in their jobs or devoting attention to family. Norms about attending to family needs on company time might be unwritten or unspoken-but still consequential. Mennino et al. (2005, p. 112) characterize the dynamics as follows: "Inasmuch as workplace practices become embedded in the normative expectations associated with the employing organization, they become taken for granted; the informal environment of an organization reflects that culture. For example, the unspoken assumption in an organization that employees leave their family at the door reflects the family-unfriendly norm of separate spheres. Conversely, an organization where most of the employees routinely take time during regular work hours to tend to family or personal matters with no fear of reprisal demonstrates a more family-friendly workplace culture" (italics added). Ideal workers are expected to display single-minded devotion to the firm and minimize interference from family roles (Blair-Loy, 2003) . While family is typically characterized as the primary countervailing force to undivided work devotion, the work-life concept extends beyond family to include needs related to one's personal life (Kossek & Lee, 2017) . People often have interests or obligations beyond the parameters of the family. Employees therefore signal divided loyalties when they seek accommodations for family or personal needs. Such violations might be interpreted as a deficient commitment to work (Cech & Blair-Loy, 2014; Williams et al., 2013) . Border theory's proposition that psychological, physical, and temporal barriers separate work and nonwork domains provides a framework for the influence of workplace culture (Clark, 2000) . Work-life culture defines the ways that workers navigate the border. For example, a norm that one should not take care of family needs on company time demarcates psychological, physical, and temporal barriers by prioritizing work. A supportive work-life culture clarifies this border by emphasizing that, if one does prioritize family needs, there will not be penalties. In other words, it is acceptable-at least on occasion-to take care of family needs as they arise. As the COVID-19 pandemic unfolded, this dynamic became especially salient for many employees. Did overall perceptions of work-life culture change early in the pandemic? In the first half of 2020, the COVID-19 pandemic dramatically altered work, family, and personal life, as lockdown requirements restricted many to their homes (Douglas et al., 2020) . Throughout most of April and May of 2020, roughly 5 million Canadians worked most of their hours from home-and the majority of those shifted from nonremote arrangements before the pandemic (Statistics Canada, 2020). The "system shock" perspective-with its emphasis on the ways that crises can reorder parts of society-provides a rationale for predicting change in perceptions of work-life culture (Been et al., 2016; Donnelly & Proctor-Thompson, 2015) . However, countervailing forces complicate narratives about the potential direction and uniformity of that change. In our view, these complications compel a more deliberate dissection of the "life" aspect of the work-life culture concept; the pandemic laid bare how the work-life concept embodies both family and personal life. This motivates greater reflection on the ways that people with family-related needs (e.g., children at home) might have experienced change in worklife culture, and their comparison to employees without similar family responsibilities. Broad social restrictions meant that many aspects of personal life outside the home decreased for most people-parents and nonparents alike. Simultaneously, however, restrictions also limited access to the usual providers of child supervision (i.e., daycares, schools); in this way, family needs likely increased for many workers. Below, we articulate how these countervailing forces contribute to competing views of potential change in work-life culture: the enhanced resource versus disrupted resource hypotheses. The enhanced resource hypothesis Several interrelated themes provide a rationale for predicting that overall levels of the perceived supportiveness of work-life culture increased. The system shock perspective underscores the power of uncertainty, suggesting that it was in all parties' interests to adjust to rapid changes during the early months of the pandemic. In this period of role restructuring, employees might have felt empowered to demand a more compassionate work-life culture or experienced a new-found leverage to discuss their personal circumstances and flexibility needs (Carnevale & Hatak, 2020; Mihalache & Mihalache, 2021) . And employers-due to their own role challenges, increased empathy, or fear of employee turnover (or some combination of the three)-might have relaxed ideal worker norms and became more accommodating of those who chose to prioritize family or personal needs (Caprino, 2020) . While empirical evidence is limited, numerous media accounts have speculated about a cultural transformation during this unsettled time in which work-family role performance expectations-both employers' and employees'-comprised more manageable standards and supportive sentiments (Dizik, 2020; Miller, 2020a; Zhang et al., 2021) . One report in the United Kingdom that used retrospective data found that respondents perceived that their workplace culture became more supportive during the COVID-19 lockdown in May and June 2020 (Chung et al., 2020) . Other scholars have implied that organizational leaders recognized the importance of being more other-oriented in the face of pandemic-related adversities (Li et al., 2021) , suggesting a broader empathic-or "family-friendlier"-turn (Covert, 2020) . Collectively, if these ideas about enhanced resources are accurate, we should observe an increase in the overall perceived supportiveness of work-life culture from pre-pandemic levels into the peak period of initial economic shocks and lockdowns of June 2020. An alternative-the disrupted resource hypothesis-predicts stability or even a decrease in the perceived supportiveness of work-life culture. The system shock idea is applicable here toobut asserts a different direction. Sometimes crises can have the opposite effect from giving workers more leverage, especially in periods of economic turmoil (Miller, 2020b) . Given the uncertainties many employers faced (Lemieux et al., 2020) , core elements of the ideal worker norm might have increased in salience. Before the pandemic, the ideal worker norm was never especially harmonious with obligations related to parenthood (Blair-Loy, 2003; Williams et al., 2013) . And with shock waves coursing through the economy, some employers might have felt an impulse toward prioritizing business survival over a more supportive work-life culture. Role restructuring linked to working from home might have intensified managers' efforts to maintain employees' productivity and engagement (e.g., endless Zoom meetings). Therefore, attending to family or personal needs might have become more dissonant with requirements for "undivided attention." In this scenario, norms that discourage prioritizing family or personal needs over work should have remained intact or even intensified. System shock research suggests that the effects of macrolevel events on organizational support could be unsettling. For example, role restructuring can increase role ambiguity and conflict (Pearlin, 1983) . Some resources may be disrupted and then repurposed in ways to better serve employers-while not necessarily being more attentive to employees' personal needs. For example, Donnelly and Proctor- Thompson (2015) observe that a key work-life resourcehome-based teleworking-acquired new meanings during a crisis event (an earthquake) in New Zealand. The intent behind teleworking shifted from employee-driven concerns for worklife fit to employer-driven motives about workers' commitment, productivity, and ultimately the survival of the business. In another "crisis study" during the 2008 Great Recession, Been et al. (2016) found that while managers felt an obligation to enable work-life arrangements for employees, they enacted practices that allowed their organizations to reap greater economic gains by reducing costs and increasing efficiency. Both studies show how-during macrolevel events-the efficacy of work-life resources shifted from employee-centric to employer-centric. The disrupted resource hypothesis is therefore based on the idea that even a pandemic might not neutralize the durable greedy institution and soften the ideal worker norms associated with it. Some observers concur. According to Covert (2020) , despite difficulties in segmenting work and family/personal lives early in the pandemic, many employers still expected "business as usual," with employees engaged in work as if little has changed. Likewise, Gelles (2020) asserts that some employers sought to sustain a pre-pandemic work ethic-thereby maintaining productivity levels-by excessively monitoring employees. Davis and Green (2020) argue that some attempts to maintain productivity levels were successful because of the social restrictions imposed via lockdowns. Virus mitigation efforts restricted many of the usual activities outside the home (Government of Canada, 2020). In response, some employees may have increased work time (Davis & Green, 2020) . This scenario brings ideal worker values of constant availability and the prioritization of work into sharper relief. Before the pandemic, Reid and Ramarajan (2016) noted that "without clear direction, many employees simply default to the ideal-worker expectation, suppressing the need to live more-balanced lives." During the pandemic, Covert (2020) suggests complications linked to ideal worker standards deepened: "This is not just a mere distraction that we can compartmentalize during the workday. No one should be expected to be able to focus on work as if everything were normal during a completely abnormal time. None of us can be ideal workers right now. Yet for many of us, there is little choice but to put together a Rubik's cube of a schedule that allows us to keep working as much as we did before." Collectively, these ideas provide a rationale for predicting stability or a decrease in the perceived supportiveness of work-life culture. Potential contingencies: Working from home, children, and professional status The ideas outlined above about broad societal changes during the early months of the pandemic and ideal worker norms underscore at least three possible contingencies: (1) working from home; (2) children in the household; and (3) professional status. We ask: Do any observed overall changes in perceptions of work-life culture differ across these conditions? The pandemic offered a "natural experiment" that rapidly shifted more than one-third of employed Canadians into work-from-home arrangements during the early months of the pandemic (Statistics Canada, 2020). The scope and pace of this role restructuring likely created new challenges associated with the integration of work and home roles (Pearlin, 1983) . To the extent that this work-from-home shift reflected flexible work arrangements-albeit structurallyenforced-the enhanced resource view would predict a greater boost in perceived supportiveness of work-life culture. For this group, employers might have relaxed expectations for role segmentation-by default becoming more accommodating. But from durable greedy institution perspective, increased (and unexpected) work-home role integration might have allowed for ideal worker norms to be sustained or intensified among those working from home. While some personal activities associated with the "life" side of the work-life formulation were restricted because of lockdowns, the opposite occurred for parenting activities. This motivates our analysis of the presence (and ages) of children at home as a potential contingency. Schools across Canada closed mid-March of 2020 and most remained closed for the school year with a few exceptions across provinces. Quebec re-opened most elementary schools mid-May; British Columbia, Manitoba, and Prince Edward Island all reopened schools at the beginning of June in some capacity-either part-time, combining in-person and virtual learning, or scheduling online meetings with teachers (Ho, 2020; Juric, 2020; Reimer & Blunt, 2020) . With restricted access to daycares and schools, much of the care and education of children shifted to the home sphere, thereby increasing the risk of competing role expectations and obligations (Dizik, 2020; Johnstone, 2020; Thomason & Williams, 2020) . Younger children, in particular, are more dependent and require more supervision and attention-and this likely elevated requirements for "involved parenting" to new levels early in the pandemic (Erickson et al., 2010; Miller, 2020a; Qian & Fuller, 2020) . Within this context, employers might have recognized the needs of employees with younger children at home and responded with greater empathy, relaxed expectations for single-minded devotion to work, and renegotiated norms about attending to family needs on company time. The enhanced resource hypothesis predicts these processes should manifest in more favorable perceptions of work-life culture among employees with young children. But the durable greedy institution thesis-and its "business-as-usual" corollary-counters the optimism. The disrupted resource hypothesis implies that employees with young children might have had more challenges with role restructuring that neutralized any boost reflected in the enhanced resource view. It therefore predicts that employees with young children should be the group most likely to experience stability or a decrease in the perceived supportiveness of work-life culture. Prior theorizing and evidence identifies the divergent salience and experiences of ideal worker norms among professionals (Blair-Loy, 2003; Williams, 2000; Williams et al., 2013) . This underlines professional status as another potential contingency. On one hand, professionals have status advantages that may position them to more effectively demand or utilize work-life accommodations. From the enhanced resource view, these status advantages of professionals might have contributed to a weakening of the ideal worker norm and the associated expectation for undivided attention. That would predict professionals experienced a greater boost in the perceived supportiveness of work-life culture. On the other hand, the durable greedy institution/ business-as-usual view acts as a countervailing force. With role restructuring, ideal worker norms and the work devotion schema may have remained intact or even intensified. In this scenario, we would expect that professionals are the group most likely to experience stability or even a decrease in the perceived supportiveness of work-life culture. While we consider each of these conditions-working from home, children, and professional status-independently as potential contingencies, we also assess whether or not they interact with each other. For example, while working from home might function as its own contingency, we see its possible overlap with the effects of children and professional status. Among those who worked from home, taking care of family needs while on company time might have become necessary-especially for those with young children. "Leaving family at the door" was simply not possible. Likewise, remote work tends to be more logistically feasible for professionals (Gallacher & Hossain, 2020) . Moreover, as Williams et al. (2013) observe: "Professionals are expected to arrange their lives to ensure unlimited availability to work unencumbered by family responsibilities" (p. 213). To maintain status, professionals with children might have therefore continued to behave in accordance with the ideal worker norm. At each stage of our analyses, we evaluate the potential for these intersecting contingencies. To add further complexity, several interrelated ideas inject gender as another potential influence. First, consistent with the family devotion schema (Blair-Loy, 2003) , traditional orientations to family as a primary sphere for employed women have endured (Dernberger & Pepin, 2020) . Second, women continue to shoulder the lion's share of family-related demands (Perry-Jenkins & Gerstel, 2020; but see Shafer et al., 2020) . Third, working mothers were more likely to exit employment early in the pandemic, as childcare and school closures resulted in reduced work hours to tend to family (Fuller & Qian, 2021; Landivar et al., 2020; Qian & Fuller, 2020) . Fourth, prior research shows that professional women experience conflicting pressures of the work and family devotion schemas (Blair-Loy, 2003; Cech & Blair-Loy, 2014) , and potentially greater bias for scheduling accommodations (Epstein et al., 1999) . In discourse about work-life culture, many observers fuse the presence of children at home with gendered work-family role dynamics. As Miller (2020a) contends: "If you're an involved parent, especially a mother, you're supposed to play down that fact at work or risk being penalized …Family caregiving has been considered a private matter, one that should not interfere with job responsibilities." But the pandemic elevated the so-called "private" challenges associated with competing devotions. Some observers in the media articulated the fear that parentsespecially mothers-might encounter greater penalties (Miller, 2020a) ; this suspicion underlines the claim that the perceived supportiveness of work-life culture might have diminished among working mothers. On the flipside, however, the pandemic might have compelled employers to reconsider the meanings around productivity and quality of work, resulting in a renewed recognition for flexible accommodations around family or personal needs. Egalitarian trends provide a counter-narrative to anticipating gender differences. While adherence to the family devotion schema-and its implications for divisions of parenting duties-continues to disproportionately affect working mothers (Ball & Daly, 2012; Collins, 2019) , some suggest that competing devotions might have intensified for working fathers due to their increased contributions to domestic labor (e.g., see Houle et al., 2017) . In fact, a Canadian study by Shafer et al. (2020) found that men's contributions to childcare and housework increased during the early months of the pandemic. Even before the pandemic, some observers had already suggested that the importance of a supportive work-life culture extended to professional fathers. According to Leschyshyn and Minnotte (2014) , "in the face of [ideal worker] norms, a supportive workplace culture might allow for professional fathers to experience a smoother navigation of the work-family interface …Professional workplaces characterized by supportive workplace cultures decrease the pressure of ideal worker norms, allowing fathers to more easily navigate work and family roles" (p. 444). Collectively, the egalitarian trend perspective predicts that observations related to the enhanced or the disrupted resource hypotheses should generalize in relatively similar ways to both women and men. To summarize, we test two competing perspectives about the ways that the perceived supportiveness of work-life culture might have changed during the early months of the COVID-19 pandemic. The enhanced resource hypothesis predicts an overall increase in supportiveness of work-life culture from September 2019 into June 2020. The diminished resource hypothesis predicts the opposite-that perceived supportiveness diminished during this period. We also evaluate if those competing predictions are contingent upon three conditions: (1) working from home; (2) children; and (3) professional status. At each stage, we test gender differences. The direction and degree of any observed gender differences remain open questions-as the traditional versus egalitarian views provide competing predictions about the continuing significance of gender as a source of inequality, especially early in the COVID-19 pandemic. To test our ideas, we analyze data from a nationally representative sample of Canadian workers in the Canadian Quality of Work and Economic Life Study (C-QWELS). The online survey was fielded from September 19 to 24, 2019 (N = 2524). We followed up June 17-24, 2020. All participants are members of the Angus Reid Forum (ARF) (see http://angusreid.org). The ARF is Canada's most recognized and engaged online community of adults managed by an industry-leading team of panel experts. Sample selection for the present study started with creating a balanced sample matrix of the Canadian population. Randomized samples of ARF members were then selected to match this matrix to ensure a broadly representative sample. For September, the response rate was 42 percent. Of the original 2524 participants, 73% were successfully retained in June. To achieve the analytical sample in the present study, first we deleted 744 cases who were not working full-time in September (N = 1780). Second, we deleted 467 cases that had dropped out by June (N = 1313) . Third, we deleted 55 cases who changed place of employment between September and June (N = 1265). Fourth, we deleted 16 cases that reported their gender as non-binary because of insufficient cases to test the focal relationships among this group (N = 1249). Fifth, we deleted 8 cases with missing values on study variables (N = 1241). The analytic methods described below combine the September sample with respondents from the June follow-up, while adjusting analyses for repeated observations of the same individuals (total number of groups = 1241; total number of observations = 2321). We weighted the data according to the most current gender, age, education, and region Census data to ensure broad representativeness of working Canadians. Tables S1 and S2 report the descriptive statistics and predictors of attrition, respectively. We measure the perceived supportiveness of work-life culture with three well-established items from the National Study of the Changing Workforce (NSCW) (Mennino et al., 2005; Thompson & Prottas, 2005) : "At my place of employment, employees who put their family or personal needs ahead of their jobs are not looked on favorably," "At my place of employment, employees have to choose between advancing in their jobs or devoting attention to their family or personal lives," and "There is an unwritten rule at my place of employment that you cannot take care of family needs on company time." Response choices are coded as follows: (1) "strongly agree," (2) "somewhat agree," (3) "somewhat disagree," and (4) "strongly disagree" We averaged the responses to these three items such that higher scores indicate a higher level of perceived supportiveness of the work-life culture at one's place of employment (alpha = .75). Working from home was assessed with the following item: "Since the onset of the COVID-19 pandemic, how often have you worked from home?" We compared those who said "most or all of the time/worked mainly from home" (coded 1) to those who reported "none/a little/some of the time." We use dummy-codes to assess the presence of children younger than age 18 at home by comparing those with no children living at home (0) to three groups who have children at home: (1) youngest child is under 6; (2) youngest child is 6-12 years-old; or (3) youngest child is a teenager. To code professional status, we adopted the occupation measure from the 2019 International Social Survey Programme (ISSP) and compare employees in professional and technical occupations to all others. Gender is coded men = 0 and women = 1. We included a set of sociodemographic characteristics and job qualities as controls to rule out their potential influence on our focal associations. Age is measured in years and marital status is coded married = 1 versus not married = 0. As part of membership in the ARF panel, all respondents were asked this question: "Would you say you are a member of a visible minority here in Canada (in terms of your ethnicity/race)?" Visible minority status is therefore self-categorized-a common approach in Canada (Little, 2016) . We coded it as follows: nonminority = 0, minority = 1. For education, we compare those with an undergraduate degree or higher to those with less than a university degree. For household income, we compare $50,000-$99,999 to the following: less than $25,000, $25,000-$49,999, $100,000-$149,999, $150,000-$199,999, $200,000 or more, and "do not know/refused." We adjust for salaried versus paid another way. We also control for long work hours and schedule control because of their potential influence on perceptions of work-life culture (Mennino et al., 2005; Thompson & Prottas, 2005; Voydanoff, 2005) . For long work hours, we contrast 50-plus hours per week with all others. For schedule control: "How much control do you have in scheduling your work hours?" Response choices are: (1) "none," (2) "very little," (3) "some," (4) "a lot," and (5) "complete." We employ mixed models that adjust analyses for repeated measures of the same individuals in September 2019 and June 2020. The basic form of the mixed model is: Within these equations, Y ij is the perceived supportiveness of work-life culture response variable for study participant i at time j, for i = 1…, n and j = September 2019 or June 2020 survey wave. The variable time indicates the survey wave, so that γ 10 indicates the average change in the perceptions from September to June. The random error terms R ij accounts for unobserved influences on perceived work-life culture due to time-varying factors. The mixed model also sets the intercept as random, which allows these perceptions to vary as a result of time-stable characteristics. The coefficient γ 0q represents the associations between perceived work-life culture and time-stable factors, while U 0j is a random error term that takes residual variation due to unobserved time-stable characteristics into account, and therefore adjusts standard errors for repeated observations of the same individuals over time. Interactions between time and each of the focal contingencies (e.g., children) test whether any observed change in perceptions between September and June differed across these conditions. We conduct analyses using Stata 16.1 with standard errors that are robust to heteroskedasticity (Hayes & Cai, 2007) . Did overall levels of the perceived supportiveness of work-life culture increase or decrease during the early months of the pandemic? The base model in Table 1 tests the bivariate association between survey wave ("June") and perceptions of work-life culture, representing the unadjusted change in overall levels of perceived supportiveness between September 2019 and June 2020. The June coefficient is positive and significant (b = .084, p < .001), indicating that-on average-the perceived supportiveness of work-life culture increased. In model 2, we demonstrate that the June coefficient holds steady net of adjustments for the focal independent variables and all controls (b = .074, p < .01); this reinforces the bivariate association shown in model 1. Among the focal independent variables, the only statistically significant pattern is observed for those who worked from home. The positive coefficient indicates that-on average-those who worked from home tend to perceive a more supportive work-life culture compared with those who did not work from home during the early months of the pandemic. Among the control variables, only long work hours and schedule control are significantly associated with perceptions of work-life culture. Employees who work long hours tend to perceive a less supportive work-life culture (b = À.122, p < .01), while those with greater schedule control tend to have more favorable perceptions (b = .098, p < .001). None of the other sociodemographic variables are associated with perceptions-importantly, this includes gender. Overall, women and men report similar perceptions of supportive work-life culture. Testing contingencies: Working from home, children, and professional status Working from home Does the observed overall increase in the perceived supportiveness of work-life culture differ depending on whether employees were working from home during the early months of the pandemic? Model 1 of Table 2 indicates that it does. We find a negative and statistically significant interaction between June and working from home (b = À.109, p < .05). To interpret this coefficient, we calculated predicted marginal effects (see Figure 1) . Those values indicate that the increase in the perceived supportiveness of work-life culture is stronger among employees who were not working from home (b = .126, p < .001); by contrast, among those who were working from home, we observe no significant change in perceptions (b = .017, p > .05). Does the increase in the perceived supportiveness of work-life culture differ depending on the presence of children at home? Model 2 of Table 2 indicates that it does. First, we find a negative and statistically significant interaction between June and youngest child under 6 (b = À.159, p < .05). Second, we did not observe significant interaction coefficients for June by either (a) youngest child is 6-12 or (b) or youngest child is a teenager. To interpret these patterns, we once again calculated predicted marginal effects (see Figure 2) . Those values indicate that the increase in perceived supportiveness is stronger among employees who do not have children at home (b = .113, p < .001). By contrast, perceptions of work-life culture held steady among employees with children at home-regardless of the age of the youngest child. Before moving on, we want to mention two important separate supplemental analyses that we conducted (not shown). First, in the interactions with June, only the contrast between employees with no children and those with a youngest child under 6 was statistically significant; moreover, analyses that compared each category of youngest child at home found no significant differences between them. Second, to assess if the contingent effect of children depended on working from home or not, we tested a 3-way interaction between June, working from home, and age of youngest child at home. That 3-way coefficient was not statistically significant. Does the overall increase in the supportiveness of work-life culture differ depending on professional status? Model 3 of Table 2 shows that it does. We find a negative and significant interaction between June and professional status (b = À.130, p < .01). Predicted marginal effects show T A B L E 2 Testing the interactions between survey wave (June) and working from home, children at home, professional status, and gender the increase in the perceived supportiveness of work-life culture is stronger among nonprofessionals (b = .134, p < .001); among professionals, we find no significant change in perceptions (b = .004, p > .05) (see Figure 3 ). As another side-note, we again conducted supplemental analyses to test 3-way interactions that include June by professional status along with either (a) working from home or (b) the age of youngest child. Our rationale for this step is that the contingent effects of working from home or age of youngest child differ by professional status. However, none of these 3-way coefficients were significant. Given the potential for gender's influence at each stage of our analyses, we evaluated genderbased contingencies-as well as its potential effects on the likelihood of stopping work or study attrition. First, we ask: Did the overall increase in the perceived supportiveness of work-life culture differ between women and men? Model 4 of Table 2 shows that it does not. We find a nonsignificant interaction between June and gender (b = À.037, p > .05). This suggests that women and men experienced a similar increase in perceived supportiveness from September to June. Next, we conducted additional analyses to address each of the following: (1) Did the 2-way interaction between June and working at home differ by gender? (2) Did the 2-way interaction between June and children differ by gender? (3) Did the 2-way interaction between June and professional status differ by gender? Answering these questions involves testing 3-way interactions between June, gender, and each of the respective other conditions (e.g., working from home). In none of these analyses did we observe any statistically significant gender-based contingencies (available upon request). Next we turn to evaluating gender's potential influence on stopping work or dropping out to gain insights about how this might bias our findings. Gender inequalities in employment and work experiences during the early months of the pandemic motivate these considerations (e.g., see Landivar et al., 2020; Qian & Fuller, 2020) . The first column in Appendix Table S2 shows that women were more likely than men to have stopped working by June (OR = 1.460, p < .05). Notably, each of the other baseline focal measures-including age of youngest child at homewere unrelated to the odds of stopping work. We then tested if the effects of the other focal measures on perceptions of work-life culture differed between women and men. Only the age of youngest child matters among men: Compared to men with no children at home, we observed reduced odds of stopping work among men with a youngest child age 6-12 (OR = .203, p < .01) or with a teenager (OR = .269, p < .05). Among women, the age of youngest child (relative to women with no children at home) was unrelated to the odds of stopping work. Our tests of gender-by-age of youngest child interactions were statistically significant-with those gender differences being driven by divergent patterns among men without children at home compared to men with a youngest child between 6 and 12 or a teenager. We also found that women were less likely than men to have dropped out of the study (OR = .803, p < .05). Overall, and with these patterns in mind, it is important to underscore that the multilevel model's treatment of data lost due to attrition depends on the assumption that missing data are "missing at random," in which missingness can be predicted by the model covariates like gender, children living at home, professional status, age, education, and other variables (Snijders & Bosker, 2012) . The inclusion of each of these variables in our models helps to ensure that the primary sources of attrition are either directly or indirectly taken into account. And finally, as a brief side-note, it is also noteworthy that baseline perceptions of supportive work-life culture had no significant effects on the odds of stopping work or dropping out of the sample among both women and men. For decades, scholars have assessed employees' perceptions of the level of supportiveness of work-life culture and linked these perceptions to outcomes like work-life conflict (Mennino et al., 2005; O'Connor & Cech, 2018; Thompson & Prottas, 2005; Voydanoff, 2005) . Using the same measure of the perceived supportiveness of work-life culture that has appeared in the National Study of the Changing Workforce, we surveyed Canadian workers to find out if these perceptions changed from the pre-pandemic levels (September 2019) into the early months of the pandemic (June 2020). We framed our ideas broadly around the "system shock" perspective because of its thesis that crises often reorder social roles. This provided our overarching rationale for the prediction that perceptions of work-life culture likely changed during the early months of the pandemic. However, we also articulated the ways that underlying countervailing forces might have complicated the narratives about the direction and uniformity of that change-and, at least in our data, this appears to have been the case. We proposed two alternative hypotheses: The enhanced resource hypothesis predicted that the abrupt changes during the early months of the pandemic would be accompanied by a softening of ideal worker norms and a "family-friendlier" mindset among employers. The disrupted resource hypothesis predicted a different direction: that perceptions of work-life culture became less supportive during the early months of the pandemic. Our rationale here was that the required readjustments and reorganization in a context of great uncertainty caused employers and organizational cultures to be less tolerant of competing devotions and therefore more restrictive around employees' efforts to deal with family/personal needs while on company time. We discovered that overall perceptions did indeed shift in a favorable direction-with the average perceived supportiveness of work-life culture increasing from September 2019 to June 2020. At first glance, this implies modest support for the enhanced resource hypothesis. But we discovered key subgroup differences which suggest that countervailing forces seem to offset that positive trend. This positive change in perceived supportiveness was muted among employees: (1) who worked from home; (2) with children under age 6 at home; and (3) in professional occupations. While we did not observe an absolute decrease in the perceived supportiveness of work-life culture among any group, it might be that the underlying countervailing dynamics articulated in the disrupted resource hypothesis contribute to this divergence in perceptions over time. In the early months of the COVID-19 pandemic, lockdown requirements restricted social life in many ways. Abrupt shocks to the organization of work and family roles compelled significant adjustments-likely magnified among those with greater domestic responsibilities. Family-related demands intensified because of restricted access to daycare and schooling. In this context, enacting penalties against employees who prioritized family needs over work needs might be deemed unfairly harsh. The optics would not be ideal. Any rules-written or unwritten-that discouraged care of family needs on company time likely receded into the background, at least somewhat. For many, especially those who worked from home, "leaving family at the door" was not possible. The thesis behind the enhanced resource perspective is that many employers should have recalibrated their expectations about the nature of the workfamily interface-especially for those with high work-home integration and young children at home. But we found the opposite, suggesting that there were countervailing forces during the early months of the pandemic. In uncertain and tumultuous times, it is plausible that some organizational cultures could have simultaneously softened ideal worker norms while still requiring a reasonable level of focus on the work role. Maintaining a single-minded devotion to work was likely unrealistic and a severe challenge for those with young children at home, but the survival of organizations in a volatile time of insecurity likely compelled organizations to require some productive capacity of their workers. Indeed, concepts like competing work-family devotions are irrelevant if one side of the seesaw (i.e., work) no longer exists. So while completely undistracted time and undivided attention is not possible for many employees, some quantity of undistracted time and undivided attention is necessary for adequate performance of the work role. While the greedy institution might have relaxed its greed during the early months of the pandemic, that let-up likely had its limits in the face of perceived or actual existential threats. This speaks to the idea of the durable greedy institution and the business-as-usual claims that some observers suggested endured as the pandemic unfolded. Professional status added another layer of complexity-showing that the increase in perceived supportiveness of work-life culture was dampened among professionals. This aligns with prior theorizing and evidence that identifies the divergent experiences of ideal worker norms and their associated challenges for time, energy, and attention among professionals (Blair-Loy, 2003; Williams, 2000) . Some scholars had already observed-before the pandemic-that professionals are often expected to organize their lives in ways that prioritize work over family or personal life (e.g., Williams et al., 2013) . The pandemic might have softened that dynamic to some degree-but new demands among professionals might have contributed to countervailing tides. Collectively, these patterns are consistent with the claim that employers' expectations that their employees adhere to ideal worker standards might have endured during the early months of the COVID-19 pandemic (Covert, 2020; Davis & Green, 2020; Gelles, 2020) . Before concluding, we discuss a few study limitations. First, we do not have actual measures of supportive work-life practices-only employees' perceptions. We were unable to find any published research since the onset of COVID-19 that documented changes in actual practices or policies, so we encourage researchers to examine this issue in greater detail. At the same time, we acknowledge that it might be challenging to find pre-pandemic baseline measures. We also recognize the complexity in the phrase "not on company time." Workplace cultures vary in what employees are allowed to do during time for which they are being compensated by their employer (e.g., calling a doctor versus watching a child). Some practices, such as flexible schedules, allow employees to meet family responsibilities while not on company time. But as we noted above, employees who take advantage of this flexibility might still experience bias. Another concern pertains to gender. We failed to detect gender differences in perceived worklife culture and the contingent effects of working from home, children, and professional status. One concern is our lack of data on housework and childcare. To the extent that women with children at home managed the lion's share of family-related demands during the pandemic, this might have increased the competing demands related to ideal worker and parent norms. It might have therefore elevated the salience of supportive work-life culture. Moreover, as working mothers experienced precarity during the pandemic (Landivar et al., 2020; Qian & Fuller, 2020) , they might have also feared penalties for prioritizing family needs. While expectations for singleminded devotion is challenging for working mothers during "normal times" (Collins, 2019) , this was more unrealistic given restrictions on daycare and school-and the associated upsurge in domestic responsibilities. Despite this, gender did not play a significant role in our results. More broadly, gender dynamics in work-family life are increasingly complex; for example, a Canadian study found that men increased their contributions to domestic needs during the pandemic (Shafer et al., 2020) . In sum, although we did not detect gender differences in our focal associations, we do not negate others that suggest divergences between women's and men's experiences early in the pandemic. More research is needed on the longer-term trends in the ways that women and men-across work-family arrangements-experience supportive work-life culture. Our findings address speculation about whether employees perceived employers as becoming more supportive of work-life fit. Future research should determine whether the actual benefits of supportive work-life culture changed. For example, did supportive work-life culture reduce work-life conflict more strongly as the pandemic unfolded? If not, employer empathy for an enhanced work-life culture might have felt more like "window dressing" and less of a substantive shift in a policy/practice-driven orientation to a more supportive work-life culture. Funding from the University of Toronto COVID-19 Action Initiative 2020 and Tri-Council Bridge funding supports this research (Scott Schieman, principal investigator). We are grateful to Alex Bierman for his extensive feedback and support. And we also thank the editor and the three reviewers for their helpful comments and suggestions. ORCID Scott Schieman https://orcid.org/0000-0003-2882-0417 Philip Badawy https://orcid.org/0000-0001-8772-7859 Family-supportive work environments: The role of organizational perceptions Dutch top managers and work-life arrangements in times of economic crisis Competing devotions: career and family among women financial executives How employees' expectations have changed through the pandemic: what leaders and HR officers need to know Employee adjustment and well-being in the era of COVID19: Implications for human resource management Consequences of flexibility stigma among academic scientists and engineers Working from home during the COVID-19 lockdown: Changing preferecnes and the future of work Making motherhood work: how women manage careers and caregiving It's ok to not be a perfect quarantine employee. The New York Times Three hours longer, the pandemic workday has obliterated work-life balance Gender flexibility, but not equality: Young adults' division of labor preferences How to work from home with your kids during coronavirus Disrupted work: Home-based teleworking = (HbTW) in the aftermath of a natural disaster Mitigating the wider health effects of covid-19 pandemic response The part-time paradox: time norms, professional life, family and gender Putting work and family experiences in context: Differences by family life stage COVID-19 and the gender gap in employment among parents of young children in Canada Remote work and employment dynamics under COVID-19: Evidence from Canada Coronavirus disease (COVID-19): Outbreak update Using heteroscedasticity-consistent standard error estimators in OLS regression: An introduction and software implementation When will school resume? What we know, province by province Changes in parents' participation in domestic tasks and care for children from 1986-2015. Spotlight on Canadians: Results from the General Social Survey COVID-19 school closures throw child care, work placements into doubt Some schools reopening to help P.E.I students who need individualized support, give respite to families Organizational stress: studies in role conflict and ambiguity Work-family conflict and work-life conflict Early signs indicate that covid-19 is exacerbating gender inequality in the labor force Initial impacts of the COVID-19 pandemic on the Canadian labour market Professional parents' loyalty to employer: The role of workplace social support How CEOS respond to mortality salience during the COVID-19 pandemic: Integrating terror management theory with regulatory focus theory Introduction to sociology Home-to-job and job-to-home spillover: The impact of company policies and workplace culture How workplace support for the COVID-19 pandemic and personality traits affect changes in employees' affective commitment to the organization and job-related well-being Three things lockdowns have exposed about working andparenting. The New York Times Could the pandemic wind up fixing what's broken about work in America? Not just a mother's problem: The consequences of perceived workplace flexibility bias for all workers Role strains and personal stress Work and family in the second decade of the 21st century COVID-19 and the gender employment gap among parents of young children Managing the high-intensity workplace Back to school in June: Manitoba students, staff headback to class for small group sessions The division of domestic labour before & during the COVID-19 pandemic in Canada Multilevel analysis: an introduction to basic and applied multilevel analysis What will work-life balance look like after the pandemic? Relationships among organizational family support, job autonomy, perceived control, and employee well-being When work-family benefits are notenough: The influence of work-family culture on benefit utilization, organizationalattachment, and work-family conflict Consequences of boundary-spanning demands and resources for work-to-family conflict and perceived stress Unbending gender: why family and work conflict and what to do aboutit Cultural schemas, social class, and the flexibility stigma Exploring public sentiment on enforced remote workduring COVID-19 Did perceptions of supportive work-life culture change during the COVID-19 pandemic?