A theoretical perspective of work–family conflict

Much of the research on work and family issues has relied on the occupational stress perspective and role theory [19]. According to role theory, roles are the result of the expectations of others about the appropriate behavior of someone in a particular position. When various members of the role set hold different role expectations for the focal person, they may impose pressures that result in psychological conflict. WFC has been defined as “a form of interrole conflict in which the role pressures from the work and family domains are mutually incompatible in some respect” [23] and consists of two broad dimensions: work-to-family conflict (hereafter WFC; i.e., work interfering with family) and family-to-work-conflict (FWC; i.e., family interfering with work) [24]. Drawn on role theory, time is a finite resource that cannot be expanded by engaging in multiple roles, an hour devoted to work represents an hour that will not be available in the family domain. The competing concept between work and family roles suggests that when people devote longer hours to work, they are less likely to meet family demands. Thus, work or family domains can impinge on each other, and interference will be experienced in two ways: bringing work home or life outside of work negatively affecting work performance.

WFC has been treated as an outcome of stress. On the other hand, it often has been considered a potential source of stress that has adverse effects on affective work outcomes such as organizational commitment [25–27], job satisfaction [4, 28, 29], and intention to quit [30]. Moreover, family characteristics (e.g., marital status, parental status, having a working spouse) may also lead to higher levels of FWC [31]. Individuals who are married and have children typically have more family role responsibilities. When confronted with antecedents such as work domain stressors, married parents may experience greater WFC than those who are single or without parental responsibilities [32].

Gender egalitarianism, family characteristics, and work–family conflict

Understanding the gender-role attitudes (traditional or egalitarian) of individuals is critical to predicting how much conflict they experience in negotiating work and family demands. From the perspective of social role theory [33], gender roles have been considered as “stereotypes produced by the sexual division of labor and societal expectations and through socialization and the formation of gender roles, the behaviors of men and women generally support and sustain the division of labor” [34]. Korabik et al. [35] posited that women with traditional gender-role attitudes might experience higher levels of WFC if they have paid employment, compared to those with more egalitarian attitudes about work and life. Similarly, men with traditional gender-role attitudes might experience higher levels of FWC if they spend time taking care of children or performing domestic tasks, compared to those with more egalitarian attitudes.

Based on GLOBE’s study, gender egalitarianism, referring to how societies divide roles between women and men, is generally conceived of as a precursor to WFC [7, 16, 20]. The more gender egalitarian a society, the less it relies on biology to determine gender roles. Using random samples collected from 10 European countries that participated in the European Social Survey in 2005, Fahlén [16] found that a greater gender gap in WFC perceptions among individuals in countries with weaker support of work–family policies and more traditional gender norms. Similarly, Rajadhyaksha and Velgach [36] found that in India, individuals with more traditional ideologies experienced more WFC. Moreover, gender had moderating effects on the relationship between gender ideology and FWC. Women with traditional attitudes experienced greater FWC than egalitarian women and men and traditional men. A comparative study by Poelmans et al. [37] examined the relationship between gender egalitarianism and WFC among 835 parents with full-time jobs in Canada, Spain, Taiwan, and the United States. Findings suggested that parents with traditional gender-role attitudes experienced greater WFC than those with egalitarian attitudes. Ruppanner and Huffman’s [7] multilevel study using ISSP data found that women’s empowerment at the country level decreased mothers’ FWC, whereas fathers in more gender-empowered countries tended to have higher levels of WFC. Thus, the first hypothesis was as follows:

1. H1: Parents in countries with more egalitarian gender practices will report lower levels of WFC.

Gender egalitarianism and family characteristics are generally conceived of as a precursor to work–family conflict. Family characteristics (e.g., marital status, having a working spouse) may associate with bidirectional work–family conflict [31]. Individuals who are married and have children typically have more family role responsibilities. When confronted with antecedents such as work domain stressors, married parents may experience greater WFC than those who are single or without parental responsibilities [32]. Fathers in dual-earner couples experienced significantly greater WFC than mothers, but other studies have found that women report higher levels of WFC than men [17]. In addition, family characteristics may lead to higher levels of FWC [31]. However, previous studies showed mixed results, which might be related to inconsistency in samples and measurements. It remains unclear whether gender egalitarianism moderates the relationship between family characteristics and work–family conflict. Thus, the second hypothesis was as follows:

1. H2. The association between family characteristics (i.e., marital status, having a spouse with full-time employment) and parents’ work–family conflict will be weaker in countries with more egalitarian gender roles.

Family-friendly policies and policy dimensions

Scholars have used institutional theory to investigate the diffusion of work–family practices among organizations [38, 39]. Institutional theory proposes that organizations are influenced by normative pressures arising from external sources such as the government [21]. National family-friendly policies are considered institutional factors that lead to consequences and norms for employers and workers to encourage conformity in terms of work such as parental leave following the birth or adoption of a child [40]. Thus, working parents in most countries today are entitled to different types of leave known by various labels. The most common types of leave are maternity, parental, and paternity leave. Maternity leave aims to protect working women during pregnancy and recovery from childbirth. By safeguarding employment and income security during and after maternity, maternity protection contributes to the health and well-being of women and their babies and the realization of gender equality and women’s empowerment [41]. Parental leave is gender-neutral, job-protected leave designed to allow employed parents to care for small children at home [42]. Parental leave has been considered an effective strategy for increasing involvement of women in the labor market, promoting gender equality and active participation in childcare, facilitating work–family reconciliation, and increasing the national fertility rate [43]. Paternity leave also allows men to spend a few days or weeks with family after childbirth or adoption with job protection [42].

Allen et al. [40] pointed out that it is challenging to compare national parental leave policies that govern work–family time due to differences such as time allotted for leave and percentage of salary paid. For example, parental leave provisions vary in terms of four main dimensions: (a) length of total continuous leave available (maternity, paternity, parental, and childcare leave; typically ranging from 9 months to 3 years and not always with pay); (b) entitlement—a mix of family and individual entitlements with the remainder divided equally between individuals (which may or may not be transferable); (c) income replacement—flat, means-tested, partial rates, or a combination of methods; and (d) flexibility, or how leave can be used (at times chosen by the employee, in one segment or shorter blocks of time, on a full-time or part-time basis, or the ability to take additional leave in special circumstances) [44]. Compared with the United States, which provides only 12 weeks of unpaid parental leave [9], Swedish parental leave offers the most generous combination of time and financial support to parents [45]. It provides 480 days of leave per child up to the age of 8 or until they complete their first year of school, a high income replacement rate (80%) for most of the leave period, and significant flexibility in terms of using leave in more than one block and on a part-time or full-time basis [9].

Various forms of national parental leave policies feature economic, cultural, and social differences [9]. Previous research has suggested that economic development is an important factor that may affect the work–family interface [46–49]. Stier et al. [15] and Allen et al. [40] used gross domestic product as an indicator of national economic development to compare WFC and FWC across countries by economic context. Economic prosperity has been viewed as a resource to lessen WFC, such as higher pay for workers to purchase groceries, and help manage work–family roles. Similarly, economic prosperity may influence organizations to offer family-friendly practices [40].

A wide variety of cross-national research has investigated the effects of parental policies on maternal and child health, well-being of preschool children and their parents, labor market attachment among women, and gender roles in families. Previous empirical studies using macro data from Organisation for Economic Co-operation and Development countries found that extended parental and maternity leave had positive effects on child health outcomes. Using data from 18 countries from 1969 to 2000, Tanaka [10]) found that the length of job-protected paid leave decreased infant mortality rates. Recent sociological research has studied the gender division of housework in terms of national context [11–13]. Using 44 time-use surveys between 1965 and 2003 from 20 countries combined with original national-level data, Hook [50] employed a multilevel model to examine the relationship between national context and household labor (e.g., laundry, caring for sick relatives) among men. He found that housework labor time among men increased with national levels of employment of women. The effects of children on men’s housework labor time depended on women’s national employment hours, the length of available parental leave, and the eligibility of men to take parental leave, indicating that particular public policies affect men in specific household situations. Campaña et al. [51] used data from the Harmonized European Time Use Survey and the European Values Study in the 2000s and 2010s to examine the gender gap in time allocation in European countries. They found that more generous paternity and parental leave policies imply that men may be more involved in care responsibilities, which may reduce the gender gap in unpaid work and potentially even paid work, because women could devote more time to their jobs.

Family-friendly policies are believed to ease the burden of negotiating work and family demands and lead to lower levels of WFC. The International Policy Center for Inclusive Growth and UNICEF in 2020 [52] reviewed maternity, paternity, and parental policies in the workplace among 24 countries in Latin America and the Caribbean. They found that extended leave policies not only increase the likelihood of breastfeeding but also improve mothers’ stress and gender equality. However, few studies have examined the effects of parental policies on WFC using cross-national random samples. Using ISSP data in 2002, Stier et al. [15] found that childcare availability and to a certain extent, maternity policies helped women and men reduce their sense of conflict. They suggested future studies use each policy’s measure to examine the effects of family-friendly policies across countries.

1. H3: Parents in countries with more family-friendly leave policy bundles—characterized by (a) higher flexibility of use, (b) longer length of leave, and (c) higher income replacement—will report lower levels of WFC.
2. H4: The association between family characteristics (i.e., marital status, having a spouse with full-time employment) and parents’ work–family conflict will be weaker in countries with more family-friendly leave policy bundles—characterized by (a) higher flexibility of use, (b) longer length of leave, and (c) higher income replacement.

Organization type and WFC

Some organizational characteristics may moderate the relationship between family-friendly public policy bundles and work–family conflict. According to institutional theory, organizations are influenced by normative pressures arising from external sources such as the government [21]. Under some conditions, these pressures may cause organizations to be guided by legitimated environment [53]. During recent decades, governments have strongly encouraged employers to provide family-friendly programs, such as childcare and flexible scheduling [54]. Social welfare professionals and other researchers have also urged organizations to proactively engage in key work–family practices to help employees balance their work and family roles [55, 56]. Hence, these social and political forces form various institutional pressures on organizations to adopt family-friendly practices in the workplace. This congruence is likely to be strong for public organizations, which will be more likely to conform to pressure to adopt family-friendly practices.

For example, in the United States, federal and state governments have taken leading roles in promoting childcare and flexible workplace options in the public sector. In 1974, the federal government adopted a flexible scheduling policy, becoming the first major organization to adopt work–family programs [57]. Subsequently, California mandated that all state agencies provide childcare information and referral services to their employees, whereas other states such as Michigan and New York introduced initiatives that fostered public support of childcare for state employees [58]. Similarly, in 2004 approximately 45% of Australian women employed in the public sector had access to paid maternity leave through industrial awards or workplace agreements [59]. Similarly, in Asia, only public employees had access to unpaid parental leave before Taiwan’s government introduced the Gender Equality in Employment Act in 2002 [60]. Using regression analysis on data collected from 993 human resources managers in Spain, Pasamar and Alegre [38] found that normative pressure determined the level of adoption of work–family practices and the use of these benefits by workers. They suggested the need to examine whether an institutional environment conduces work–family balance by adoption of work–family policies and promotes use of these benefits, particularly in the public sector.

1. H5. Compared to private organizations, the association between public organizations and parents’ WFC will be stronger in countries with more family-friendly leave policy bundles—characterized by (a) higher flexibility of use, (b) longer length of leave, and (c) higher income replacement.
2. H6. The association between public organizations and parents’ WFC will be stronger in countries with more egalitarian gender roles.

The majority of the literature reviewed here has shown that individuals’ WFC is affected at the macro (social policies) and micro (individuals and families) levels. However, the association between family-friendly policies and WFC among working parents remains unclear. Furthermore, discrepancies in the results of literature examining the effects of gender egalitarianism and gender on WFC could be attributed to differences in methodologies (e.g., only managerial employees, number of participating countries). Some studies found no difference in WFC between men and women [61]. Byron [61] conducted a meta-analysis of 61 published studies that suggested gender has no effects on either WFC or FWC. Due to inconsistencies regarding the moderating effects of gender in the literature, the present study performed statistical analysis separately by gender rather than including it as a control variable to more thoroughly explore similarities and differences between fathers and mothers. Recognizing these gaps in the international work–family literature, the present study examined multilevel effects of family-friendly policies and gender egalitarianism on WFC among working parents by gender, using ISSP data collected in 2005 from national probability samples in 24 countries in Africa, Asia, Europe, and North and South America.

Data

This study used the secondary data collected by the International Social Survey Program (ISSP) from random samples of 6,878 employees in 24 countries located in Africa, Asia, Europe, and North and South America to investigate the work–family interface in terms of organizational work–family initiatives, work demands, work–family conflict, and job-related outcomes. The ISSP’s official data archive (https://search.gesis.org/research_data/ZA4350) is the Zentralarchiv at the University of Cologne in Germany. The ISSP is an annual social survey conducted in 40 countries throughout the world. Using a multistage, stratified, probability-proportional-to-size sampling method, the data were collected via a cross-national collaboration in which independent institutions replicated survey questions among random samples in their country.

This study used 2005 ISSP data: Work Orientation modules as publicly available secondary data (de-identified data) that have been stripped of all identifying information and cannot be linked back to the subjects. Data analysis had not involved human subjects; therefore, IRB review was not required for this study. In recent decades, data collected by ISSP have been used widely by other social scholars [11, 13, 27, 45, 62]. The present study used GLOBE’s cultural dimensions of gender egalitarianism as the theoretical perspective [63] that classifies countries into eight clusters: Anglo, Confucian Asia, Eastern Europe, Germanic Europe, Latin America, Latin Europe, Nordic Europe, and Southern Asia. Although the 2005 ISSP data on work orientations were collected from 31 countries, seven countries (i.e., Bulgaria, Cyprus, Czech Republic, Dominican Republic, Flanders, Latvia, and Norway) in the ISSP data were excluded from the analysis because these countries were not included in the GLOBE study.

Based on GLOBE’s theoretical perspective, the 24 countries were ranked based on mean gender egalitarianism practice scores [64]. Hence, ISSP data used in the present study were collected in 24 countries: Australia, Denmark, Finland, France, Germany, Great Britain, Hungary, Ireland, Israel, Japan, Mexico, New Zealand, Norway, the Philippines, Portugal, Russia, Slovenia, South Africa, South Korea, Spain, Sweden, Switzerland, Taiwan, and the United States.

Although ISSP program collected data relevant to work orientation in 1989, 1997, 2005, and 2015, not all 24 countries participated in four-wave data collection. Only 9 countries in western societies participated in the very first wave of data collection in 1989 while South Korea only participated in the 3^rd wave of data collection in 2005. In addition, ISSP survey items were somewhat different across 4 waves. Work-family conflict as a variable of interest in this study was not included in the ISSP surveys in 1989 and 1998. It is difficult to pool data from four waves and make cross-national comparisons. Considering that South Korea is a country of research interest for the author, the 2005 ISSP data were used for this study.

Due to hypotheses drawn on theoretical perspectives of this study, the sample was restricted to participants who had children in the household and full-time jobs. Data were collected from one member of each household. A random sample of 6,878 parents (4,015 fathers and 2,863 mothers) was included in the analysis.

Measurement

Independent variables.

Individual-level variables. Organization type. Organization type was assessed with a single item measuring whether respondents worked for a private or public employer or were self-employed. Responses were coded as 1 = public sector and 0 = private sector or self-employed.

Family characteristics. Marital status (1 = married, 0 = other) and dual-earner family status (1 = having a spouse with full-time employment, 0 = other) were dichotomous variables.

Control variables. Two covariates, job position and education, were included in the models. Job position (1 = supervisor, 0 = other) was measured as a dichotomous variable. Education was measured in years.

Macro-level variables. Family-friendly public policy. Comprehensive reviews of government-mandated family-friendly policies (i.e., maternity, paternity, parental, and childcare leave) enacted in 2005 in 24 countries were conducted. Because Allen et al. [40] and Moss and O’Brien [9] indicated that national parental leave policies differ in time allotted for leave, flexibility, and percentage of salary paid, three variables were created to measure these three dimensions of family-friendly public policy: flexibility of use, length of leave, and income replacement rate. Qualitative information specified in the parental leave’s statutory statement varies regarding how leave can be used. Thus, based on evaluation criteria suggested by Deven and Moss [65] for reviewing family policies across 19 countries, flexibility of use in this study was measured as 1 = no flexibility; 2 = leave can be taken full-time or part-time; 3 = leave can be taken in one block or several blocks of time; 4 = leave can be taken for a shorter period with a higher income replacement rate or for a longer period with a lower income replacement rate; 5 = leave can be transferred to nonparents; and 6 = leave can be taken at any time until a child reaches a certain age (see Table 1). Higher scores indicated higher levels of freedom with how parents using leave policies and taking time off from work.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Country-level descriptive statistics (N = 6,878).

https://doi.org/10.1371/journal.pone.0291127.t001

The unit (e.g., calendar days, weeks, months, years) used for length of leave specified in each country’s written statement regarding statutory parent leave varies greatly. For example, in the United States, the statutory statement sets out 12 weeks for family and medical leave, whereas in South Africa, the statement specifies 4 months for maternity leave. Due to inconsistent units for length of leave ranging from more than 2 months (the Philippines) to 3 years (France) and a lack of the exact numbers of days in most countries’ written statement, a variable was created for total length of leave to enable cross-national comparisons. Length of parental leave (i.e., weeks, months, and years) in certain countries were converted into days. Thus, length of leave was measured as total statutory postnatal leave as 1 = less than 180 days; 2 = between 180 and 360 days; 3 = between 361 and 540 days; 4 = between 541 and 720 days; and 5 = more than 720 days. Higher scores indicated longer leaves and generosity of the policy.

Moss and O’Brien [9] pointed out that across countries, the format of payments also varies (i.e., flat-rate, mean-tested, paid for only part of the leave period, or a combination of these). A variable for national income replacement rate was created to measure each country’s financial support for parents taking parental leave in 2005. To enable cross-national comparisons, the rate of income replacement indicated how well each nation financially supported parents who took leave, ranging from 15.38% (Taiwan) to 105.45% (Slovenia) rather than the wage replacement (depending on individuals’ income) specified in the statutory statement. Thus, the rate of income replacement was calculated as average income replacement provided by each government to each parent who took parental leave during 2005 divided by the per capita gross domestic product of each country (see Table 1) [66]. Economic development (indicated by per capita gross domestic product) was embedded in the income replacement formula, based on Allen et al.’s [40] meta-analytic study using gross domestic product as an indicator of national economic development, to compare cross-national differences in WFC and FWC across economic contexts.

Gender egalitarianism. The present study used GLOBE’s cultural dimensions of gender egalitarianism [63], defined as how societies divided roles between women and men. According to GLOBE’s gender egalitarianism practice score, each country in the present study was classified into one of eight clusters: Anglo, Confucian Asia, Eastern Europe, Germanic Europe, Latin America, Latin Europe, Nordic Europe, and Southern Asia (see Table 1). The Eastern Europe cluster scored highest on gender egalitarianism practice (M = 3.84), whereas the Germanic Europe cluster scored lowest (M = 3.14). Higher scores indicated more egalitarian gender practices.

Data analysis

Hierarchical Linear Models (HLM) allowed for multilevel analysis of WFC using individual-level (level-1) and country-level (level-2) variables simultaneously, including tests for cross-level interactions between national family-friendly policies and organization type, marital status, and the effects of dual-earner families on WFC. This allowed for separate error terms and thus, correct estimations of standard errors at both individual and country levels. Hierarchically structured data is nested data where groups of units are clustered together in an organized format. Because the results regarding gender differences in the literature have been inconsistent, data analyses were conducted separately by gender to better understand the influence of gender on the effects of family-friendly policies on WFC (both WFC and FWC). Data analysis strategy follows a certain procedure of HLM to present the results by model. Presenting Model 1 to Model 4 step by step helps the readers understand whether the effects of key individual-level factors differ by country and whether those differences are explained by country-level variables.

The individual-level model was specified as follows: in which Yij was the level of WFC or FWC for parent i in country j; β0j was the level-1 intercept; β1j was the effect of public organization; β2j was the effect of marital status (married); β3j was the effect of having a spouse with full-time employment; βkj X were the slopes for k control variables X; and rij was the level-1 error term.

The country-level model was specified as follows: in which G00 was the country-level intercept; G01 was the effect of flexibility of use on the intercept (β0j); G10 was the country-level intercept for the public organization slope; G11–G13 were the cross-level interaction effects of flexibility of use, length of leave, and income replacement rate, respectively, on β1j; G14 was the cross-level interaction effect of gender egalitarianism on β1j; G20 was the country-level intercept for the marital status (married) slope; G21–G23 were the cross-level interaction effects of the three dimensions of policies on β2j; G30 was the country-level intercept for the slope of having a spouse with full-time employment; G3–G33 were the cross-level interaction effects of the three dimensions of policies on β3j; and U0j–U3j were the country-level error terms, assumed to be normally distributed with mean zero and variance σ2. All individual- and country-level variables were centered on their grand means; the intercept indicated average levels of WFC or FWC for men and women in a country with average characteristics.

Sample

Table 2 shows individual-level (level-1) descriptive statistics for all 24 countries. Fathers averaged 40.11 years old, compared to 38.38 years old for mothers. Most participants—86% of fathers and 73% of mothers—were married. Thirty-five percent of mothers worked for public organizations compared to 22% of fathers. More fathers were supervisors compared to mothers (42% vs. 29%, respectively). Fathers and mothers received about the same length of education (12.18 years vs. 12.74 years, respectively). Thirty-seven percent of fathers had a spouse with a full-time job, whereas 66% of mothers had a spouse with a full-time job. The average score of WFC was 2.67 for fathers, compared to 2.63 for mothers. The average score of FWC was 2.11 for mothers, slightly higher than FWC scores of fathers (M = 2.09).

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Individual descriptive statistics of the sample (N = 6,878).

https://doi.org/10.1371/journal.pone.0291127.t002

Country-level (level-2) descriptive statistics are shown in Table 1. For WFC, British fathers reported highest levels of WFC (M = 3.26), whereas Taiwanese fathers had the lowest WFC (M = 2.18) among the 24 countries. Australian mothers had highest levels of WFC (M = 3.14), whereas Taiwanese women had the lowest WFC (M = 2.05). Fathers and mothers in Switzerland experienced the highest levels of FWC (fathers = 2.65, mothers = 2.67), whereas those in Russia had the lowest levels of FWC (fathers = 1.53, mothers = 1.64).

Family-friendly policy bundles—specifically, flexibility of use, length of leave, and income replacement—varied widely across 24 countries. Anglo countries (e.g., Great Britain, Ireland), Latin European countries (e.g., France, Portugal), Eastern European countries (e.g., Hungary, Slovenia), Germanic European countries (e.g., Germany), and Nordic European countries (e.g., Sweden) have implemented highly flexible parental leave policies. Parents in these countries could take leave at any time until a child reached a certain age. In terms of total statutory postnatal leave, France, Germany, Hungary, Russia, Spain, and Taiwan offered more than 2 years. In contrast, Israel, Mexico, the Philippines, South Africa, Switzerland, and the United States had the shortest parental leave, ranging between 67 and 120 days. Except for the United States, the parental leave policy in the latter countries had no flexibility. In terms of income replacement during parental leave, Nordic European countries such as Denmark (100%) and Sweden (82.19%) and Eastern European countries such as Hungary (103.68%), Slovenia (105.45%), and Russia (93.15%) offered high replacement rates, whereas Taiwan (15.38%) and the Philippines (18.36%) had the lowest replacement rates.

WFC as an outcome variable

Tables 3–5 present the results of four hierarchical linear models of WFC run separately for fathers and mothers. As shown in Table 3, the intraclass correlation (ICC) results indicate that between-country variance constituted 7.2% of the overall variance in WFC among fathers, whereas between-country variance constituted 5.6% of the overall variance in WFC among mothers. The between-country (level-2) variance in the intercept was statistically significant [67] for fathers and mothers in Model 1. Therefore, it was appropriate to examine the multilevel model posited. The intercept showed that fathers had an average WFC score of 2.72, compared to 2.70 for women.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Hierarchical linear model results for individual- and country-level effects on WFC among fathers and mothers (Model 1).

https://doi.org/10.1371/journal.pone.0291127.t003

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Hierarchical linear model results for individual- and country-level effects on WFC among fathers and mothers (Models 2 and 3).

https://doi.org/10.1371/journal.pone.0291127.t004

Download:

• PNG
  larger image
• TIFF
  original image

Table 5. Hierarchical linear model results for individual- and country-level effects on WFC among fathers and mothers (Model 4).

https://doi.org/10.1371/journal.pone.0291127.t005

Model 2 included the three main individual-level variables—organization type, marital status, and having a spouse with full-time employment. Table 4 shows that working in a public organization had a negative effect on fathers’ WFC (G10 = -.078, p = .045). With respect to family characteristics, being married increased levels of WFC among fathers across countries (G20 = .172, p = .002), but having a spouse with full-time employment had no effect on WFC among fathers. Conversely, intercept coefficients of the three main family characteristics variables at the individual level had no effect on WFC among mothers.

In Model 3, individual-level control variables were added. For the fathers’ model, coefficients of public organization (G10 = -.080, p = .023) and marital status (G20 = .138, p = .013) variables were allowed to vary across countries. Job position (G40 = .163, p = .001) and education (G50 = .017, p = .005) had significant effects on fathers’ WFC; being a supervisor and having higher levels of education increased levels of WFC among fathers. After controlling for job position and education, cross-national differences in the effects of public organization and marital status on WFC among remained significant. All individual-level variables explained 5.9% (.119 versus .112) of the variance in country-level intercepts among fathers. Thus, nearly 6% of the variation in fathers’ WFC across countries was accounted for by cross-national variations in organization type (public) and marital status (being married). However, in the mothers’ model, none of the three main individual-level variables (organization type, marital status, having a spouse with full-time employment) had significant effects on WFC. Instead, job position (G40 = .157, p = .002) and education (G50 = .022, p = .008) were associated with higher levels of WFC among mothers. All individual-level variables explained 1.5% of the variance in country-level intercepts among mothers.

Model 4 (as shown in Table 5) included the effects of parental leave policies and gender egalitarianism on the intercept and slopes of the three individual-level variables. For fathers, parental leave policies had significant effects on the intercept. Rate of income replacement had a negative effect on the intercept (G03 = -1.094, p = .008), whereas flexibility of use had a positive effect on WFC (G01 = .113, p = .010). That is, fathers in countries that offered more generous rates of income replacement during parental leave had lower levels of WFC, supporting Hypothesis 3c. In contrast, countries with more flexibility of use regarding parental leave had higher levels of WFC among fathers. Length of leave had no significant effect on WFC among fathers across countries. Therefore, Hypotheses 3a and 3b were not supported. In terms of gender egalitarianism, fathers in countries with more egalitarian gender practices experienced more WFC (G04 = 1.05, p = .003). Among mothers, only flexibility of use had a significant effect on WFC (G01 = .069, p = .026), suggesting that mothers in countries with more flexible parental leave policies experienced higher levels of WFC.

FWC as an outcome variable

Tables 6–8 present the results of four hierarchical linear models of FWC run separately by gender. As shown in Table 6, the ICC results indicate that between-country variance constituted 7.6% of the overall variance in fathers’ FWC, whereas between-country variance constituted about 8.1% of the overall variance in mothers’ FWC. The between-country (level-2) variance on the intercept was statistically significant [67] for analysis of variance models for both genders. The intercept showed that the average FWC score among fathers was 2.09, compared to 2.15 among mothers.

Download:

• PNG
  larger image
• TIFF
  original image

Table 6. Hierarchical linear model results for individual- and country-level effects on FWC among fathers and mothers (Model 1).

https://doi.org/10.1371/journal.pone.0291127.t006

Download:

• PNG
  larger image
• TIFF
  original image

Table 7. Hierarchical linear model results for individual- and country-level effects on FWC among fathers and mothers (Models 2 and 3).

https://doi.org/10.1371/journal.pone.0291127.t007

Download:

• PNG
  larger image
• TIFF
  original image

Table 8. Hierarchical linear model results for individual- and country-level effects on FWC among fathers and mothers (Model 4).

https://doi.org/10.1371/journal.pone.0291127.t008

Model 2 included the three individual-level variables: organization type, marital status, and having a spouse with full-time employment. Only marital status had a significant positive effect on FWC among fathers (G20 = .128, p = .029) and mothers (G20 = .145, p = .007). In other words, married fathers and mothers both experienced higher levels of FWC compared to single parents.

In Model 3, all individual-level control variables were added. The coefficient for marital status remained significant for FWC among fathers (G20 = .127, p = .029) and mothers (G20 = .166, p = .004). Education was the only control variable with significant effects on levels of FWC for both fathers (G50 = .015, p = .006) and mothers (G50 = .010, p = .044). Model 4 included effects of parental leave policies and gender egalitarianism on the intercept and slopes of the individual-level variables. The results of Model 4 suggest obvious differences between genders in terms of country-level effects on FWC. For fathers, flexibility of use significantly increased FWC (G01 = .063, p = .048), whereas rate of income replacement had a negative effect on the intercept (G03 = -.609, p = .036). In countries with more flexible parental leave policies, fathers experienced higher levels of conflict from the family domain to the workplace. In contrast, in countries that offered more generous rates of income replacement, fathers had lower levels of FWC. For mothers, none of the country-level variables had significant effects on FWC.

Implications for research and practice

The findings of the present study suggest the importance of considering workplace context and gender roles when implementing family-friendly policies. The results indicate that both fathers and mothers were likely to face costs associated with taking leave from work, which deserves attention in future research. Workplace characteristics influenced fathers more than mothers in terms of the effects of parental leave flexibility of use, length of leave, and income replacement rates on WFC. Negative consequences may be associated with absence from the workplace, especially among women. More research is needed to investigate relationships among workplace gender composition and size to reveal the mechanisms that link policy, work characteristics, and WFC. Furthermore, WFC among men and women may be significantly influenced by earlier parental leave-taking by colleagues, workplace-specific social networks, and hierarchies in the workplace. Family policies such as maternity and paternity leave might be different for salaried and self-employed workers. For example, do self-employed mothers use their time differently than their salaried counterparts? Researchers could explore whether there are some specific characteristics of leave policies (e.g., leave being full time or part time) that might reduce work-family conflict and the mechanisms that link parental leave policies, WFC, and consequences of taking leave to provide more nuanced information about how women may be negatively affected by taking a leave of absence.

The main objectives of family-friendly policies are to reduce the conflict between family roles and commitment to work, allow more women to join the labor force, and improve the fertility rate. The findings of present study suggest that implementing parental leave policies with high flexibility and higher rates of income replacement may help fathers with a working spouse or who are employed in the public sector to better balance their work and family demands. In this study, mothers generally were not protected by individual dimensions of parental leave policies across countries. Instead, societal attitudes toward gender played a key role in helping mothers reduce bidirectional conflicts between work and family roles.

Findings of this study provide implications for policy makers and employers, especially those in developing countries (e.g., Vietnam, India) characterized by increased involvement of women in the job market with traditional workplace characteristics, related to guiding their family-friendly leave policies to cater to the specific needs of women. Gender egalitarianism might be best achieved if both parents have access to workplace supports such as childcare that allow them both to seamlessly work long hours. To promote work–life balance among working parents, policy makers and regulators need to monitor and examine diffusion and patterns of parental leave use in the workplace by gender and evaluate the effectiveness of parental leave policies for policy implementation and improvement.