The United States is in the midst of a critical conversation about parental leave policies because it remains one of two out of 185 countries that does not offer a national paid maternity leave policy (Addati et al. 2014). Fully 71 of those countries that provide a national paid maternity leave policy also extend their policies to fathers (Addati et al. 2014). According to data from the U.S. Current Population Survey (CPS), conservative estimates suggest that in each month in 2015, there were an average of 299,861 women on maternity leave and 21,703 men on paternity leave (Zagorsky 2017). Unfortunately, U.S. policymakers who are involved in these critical conversations have very little research to inform their decisions. Most research on parental leave has occurred outside the United States in countries such as Sweden and Norway that have generous national paid leave policies and much different demographic profiles (Bergqvist and Saxonberg 2017). Moreover, the research that does exist on this topic tends to focus more on maternity leave relative to paternity leave, even though fathers in the United States are more involved in childrearing than ever before (Lamb 2010).

At the transition to parenthood, when new family relationships are emerging, the time parents take off from work after the birth of a child is one important factor that influences the quantity and quality of their involvement in childrearing (Nepomnyaschy and Waldfogel 2007; Petts and Knoester 2018; Pragg and Knoester 2017), which in turn may impact their child’s development and familial relationships (Petts and Knoester 2019). During early childhood, parental involvement is key for establishing parent-infant attachment (Ainsworth 1979) and supporting children’s social, emotional, and cognitive development (Easterbrooks and Goldberg 1984; Flouri 2005; Shannon et al. 2006). This time period is often described as a “critical period” for the development of family relationships, such that the patterns established may persist for years (Doherty et al. 2006, p. 438; Pleck 1997). Indeed, fathers who took 2 or more weeks of leave from work after the birth of their child were more involved in direct childcare (e.g., changing diapers) at 9-months postpartum (Nepomnyaschy and Waldfogel 2007). Furthermore, longer paternity leaves have been linked to fathers’ increased involvement in providing care for and engaging in developmental tasks with their child both during infancy and during the child’s first few years of life (Petts and Knoester 2018).

For mothers, shorter maternity leave has been linked to a reduction in the sensitivity they show toward their infant, greater marital dissatisfaction, and increased depression and anxiety (Clark et al. 1997; Gjerdingen and Chaloner 1994; Hyde et al. 2001). Mothers who returned to work full-time within 12 weeks of the birth of their child were more likely to reduce the amount of time they breastfed and the number of well-baby visits their child attended, and they reported higher levels of child externalizing behavior when their child was 4-years-old (Berger et al. 2005). Beyond the parent-child relationship, Petts and Knoester (2019), using data from the Fragile Families and Child Wellbeing Study, reported that in the years following a child’s birth, longer paternity leave was positively associated with parental relationship quality and coparenting quality. Given this growing body of literature linking parental leave and later parent-child and familial outcomes, surprisingly little is known about predictors of leave-taking duration for new parents.

The current study examined how numerous demographic and attitudinal measures including financial characteristics (e.g., access to paid leave, household income), demographics (e.g., age, education), identity-relevant characteristics (e.g., parental role importance, parental role beliefs), and job characteristics (e.g., job satisfaction, work shift) are associated with the length of parental leave among new mothers and fathers. Our study improved upon previous research on parental leave-taking by (a) measuring the length of maternity and paternity leave continuously in days, (b) including a paid leave variable that reflects the proportion of leave that was paid, and (c) examining mothers and fathers in the same couples and within the same model. Our research provides scholars and policymakers with a better understanding of which mothers and fathers are taking the longest leaves from work and if the length of leave is based solely on access to any source of paid leave (because we were unable to examine specific sources of paid leave in the present study) or if expectant parents’ characteristics also influence leave-taking. This information, in turn, can be used to inform policy and/or interventions to increase the support for and use of parental leave at the transition to parenthood.

Parental Leave in the United States

The 1993 Family and Medical Leave Act (FMLA) provides up to 12 weeks of unpaid, job-protected leave to eligible employees after the birth of a child (U.S. Department of Labor n.d.). FMLA represented an important shift in the opportunity for more mothers and fathers in the United States to take parental leave than ever before. It is estimated that 73.6% of employees work at FMLA eligible worksites. Unfortunately, FMLA eligibility requirements further reduce the number of FMLA eligible employees to 59.2% (Klerman et al. 2012). Beyond the eligibility requirements, it is often financially unrealistic for new parents to take unpaid time off from work. As a result, expectant parents may rely on one partner’s income for a period of time or one partner, usually the mother, may opt to stay home with the child (Kramer et al. 2015).

Because the United States continues to lack a national paid leave policy, other sources of paid parental leave have emerged. Although not relevant to the state in which data for the current analyses were collected, it is important to note that some states (and Washington D.C.) have enacted their own state-level paid leave policies (Brainerd 2017). Additionally, individual workplace policies are sometimes more generous than FMLA, but even then, these policies are only offered by a small percentage of employers and there are often caveats (Bureau of Labor Statistics 2018). Of the small percentage of workplaces that offer more generous policies, they tend to favor mothers over fathers (Zagorsky 2017) and higher-earning over lower-earning employees (e.g., Starbucks; Redden 2017). As a result, most parents, especially fathers and those in entry-level and hourly jobs, remain unable to access paid leave time through a state-level policy or their employer to adjust to parenthood, and they may only have access to unpaid leave or a limited number of sick and/or vacation days hence shortening or eliminating leave. In contrast, a growing body of literature conducted in states with state-level leave policies has documented that leave-taking increases when paid leave is enacted (Bartel et al. 2018; Rossin-Slater et al. 2013).

The Impact of FMLA

Research that spanned the implementation of FMLA suggests that mothers’ leave-taking was impacted more than fathers’ (Waldfogel 1999; Han and Waldfogel 2003). For mothers, FMLA coverage was associated with increased leave-taking in the first 3 months after the birth, but not in the month of the birth, suggesting that women were likely utilizing some sort of paid leave (e.g., accrued vacation leave or sick days) during the birth month, and then accessing unpaid leave via FMLA after they had used up their paid leave days. From 1994 to 2015, the number of men taking a week or more of paternity leave increased three-fold. In each month, 67 more men took a week or more of paternity leave than in the month prior. Yet in 2015, only 10% of employed fathers reported taking a week or more of paternity leave after the birth of a child (Zagorsky 2017). Given limitations of the CPS data, this is likely an underestimate of the percentage of fathers who take leave because these data do not capture fathers who took less than a week of paternity leave (Zagorsky 2017). More detailed, community-level data from a sample of 550 fathers suggests that 91% of them took at least 1 day of leave, with a mean of 5 days and a mode of 2 days (Hyde et al. 1993).

Not surprisingly, the gendered pattern of parental leave-taking remains robust, even though more women have entered the workforce and dual-earner couples have become increasingly common (Bureau of Labor Statistics 2010; Fitzgerald and Harmon 2001; Toossi 2002). More mothers than fathers take parental leave, and mothers consistently take longer leaves than fathers (Armenia and Gerstel 2006; Zagorsky 2017). Data from the Early Child Longitudinal Study-Birth Cohort indicates that by 3 months postpartum, 67% of mothers who were employed prior to the birth of a child had returned to work; 89% by 9 months postpartum (Han et al. 2008). Using the same data source, Petts and Knoester (2018) reported that 88% of fathers in their sample took paternity leave with an average length of leave of 1.53 weeks. Similarly, in a survey of just over 1000 fathers with at least one child under the age of 18 asking about their most recently born child, 75% of fathers indicated that they took one week or less off from work and 16% did not take any time off from work (Harrington et al. 2011). However, it is surprising, given the robust gendered pattern of leave-taking and individual workplace policies favoring mothers over fathers, that fathers are more likely to report taking paid leave than mothers (Lenhart et al. 2019; Zagorsky 2017).

Previous Research on Predictors of Leave-Taking

One set of studies that has examined differences between parents who take leave and those who do not has focused primarily on the roles of demographic factors. Klerman and Leibowitz (1994) reported that older mothers and mothers with higher educational attainment were more likely to report taking paid leave offered by their employer. But, when controlling for educational attainment, mothers who earned higher wages were more likely to take unpaid leave. Similar findings persisted across two decades after the implementation of FMLA: Women who took at least a week of maternity leave remained qualitatively different (i.e., higher educational attainment, more likely to be married, approximately 2.4 years older, and non-Hispanic White) when compared with all mothers who gave birth in the same time period (Zagorsky 2017). Fewer studies have examined demographic predictors of the duration of parental leave. Across the implementation of FMLA (1987–1994), the length of leave for both mothers and fathers increased, but the greatest increase was for college-educated, married mothers and college-educated fathers (Han et al. 2009). More recent evidence indicates racial and educational differences in the length of maternity and paternity leave, with more highly educated fathers, mothers who did not complete high school, and Black mothers taking the longest leaves (Petts 2018; Petts et al. 2018).

Other investigations have included attitudinal and social-contextual factors as predictors of parental leave duration. Feldman et al. (2004) examined predictors of the length of maternity and paternity leave in a community sample of 98 dual-earner couples recruited when their first-born child was 3–5 months-old in 1996–1998. Differences between mothers who took longer (≥ 12 weeks) and shorter leaves (< 12 weeks) were examined. Shorter leaves for mothers were predicted by higher career centrality, lower infant preoccupation, higher maternal depression, and the negative impact of the birth on the marriage and on maternal self-esteem. Longer leaves for fathers, measured in days, were predicted by higher family salience, employer’s positive reaction to childbirth, planned pregnancies, and more marital support. The major limitation of this prior study was that the predictors of the length of maternity and paternity leave were collected several months after the child was born, presumably after the majority of mothers and fathers returned to work. Thus, parental perceptions may have already been influenced by their early experiences as dual-earner couples with newborns.

More recently, Petts et al. (2018) drew from two national datasets to examine measures of father identity salience and commitment as predictors of paid paternity leave-taking and duration. They reported that first-time fathers and fathers who were involved in parenting activities prior to the child’s birth were more likely to take paid leave and to take longer periods of paid leave. Moreover, fathers who endorsed positive attitudes about fatherhood were also more likely to take longer periods of paid leave. In their study, paid leave-taking was treated as a dichotomous variable; the length of paid leave, as a categorical variable (i.e., no leave; 1 week or less of paternity leave; more than one, but no more than 2 weeks of paternity leave; and more than 2 weeks of paternity leave). The current study improves upon previous research on parental leave-taking by measuring the length of maternity and paternity leave continuously in days, including a paid leave variable that reflects the proportion of leave that was paid and examining mothers and fathers in the same couples and within the same model.

Identity-Relevant Predictors of Leave

Drawing from identity theory (Stryker 1980; Stryker and Serpe 1982), the roles with which an individual identifies (e.g., parent, partner, worker, friend) can be organized into a salience hierarchy. Identity theorists postulate that the more salient an identity, the more likely an individual’s behavioral choices will reflect that identity. At the transition to parenthood, expectant parents likely identify with multiple roles (e.g., husband/wife/partner, mother/father, worker, friend). The role that is the most salient to the parent’s identity should result in behavioral choices that favor that role. For example, if an expectant father identifies his most important role as worker, he may be less likely to take advantage of parental leave, whereas a father who identifies his most salient role as parent may be more likely to take leave. Feldman et al. (2004) found that mothers with higher career salience took shorter leaves and fathers with higher family salience took longer leaves. However, parents in their study reported role identity after making leave-taking decisions. Thus, it was possible that postpartum experiences with work or parental leave may have precipitated shifts in identity salience. Therefore, reexamination of parental role identity in relation to leave-taking across the transition to parenthood is warranted.

As pointed out by Hyde et al. (1993), men’s role salience is complex because of the shifting expectations of fathers in U.S. society. Traditionally, being a good provider was synonymous with being a good father (Pleck 1981). Fathers who identify with the good-provider role may be less likely to take paternity leave unless it is paid because of their desire to provide financially for their families. However, fathers are now expected, and often desire, to be more directly involved in childrearing (Lamb 2010). Thus, multiple roles may be salient for fathers, making other identity-relevant attitudes also important for understanding leave-taking behavior. For instance, a father’s gender role ideology may be another important factor in his decision to take time off from work after the birth of a child. Indeed, in a longitudinal study across the transition to parenthood, fathers who took the longest leaves from work reported gender role attitudes that were more egalitarian, in addition to reporting higher family salience (Hyde et al. 1993).

Mothers’ role salience also has become increasingly complex as more mothers have entered the workforce (Fitzgerald and Harmon 2001; Toossi 2002), yet their roles as primary caregivers for children remain pervasive. Intensive mothering (the dominant mothering ideology in the U.S.) is mothering that is wholly child-focused (emotionally and financially), labor intensive, and the sole responsibility of mothers (Hays 1996). Douglas and Michaels (2004, p. 4) termed these demanding and high parenting expectations for mothers as “New Momism,” and Lee et al. (2012) suggested that a new cultural phenomenon of perfectionistic parenting may be emerging. Thus, mothers, even (or especially) those who work full-time, may be subject to unrealistically high parenting standards, and these perfectionistic standards may affect their leave-taking behavior over and above role salience. For example, mothers who feel held to unrealistically high standards for parenting may take longer leaves from work in order to attempt to meet those standards.

Another important consideration for understanding parental leave-taking involves the development of expectant parents’ identities across the transition to parenthood. Petts et al. (2018) found that greater involvement in early parenting activities prior to the child’s birth was associated with longer paid paternity leaves. One way to capture identity development is the extent to which the pregnancy was planned. Parents who have planned a pregnancy to a greater extent may be more likely to take longer leaves because they have had more time to contemplate and anticipate parenthood, resulting in more well-developed identities as parents. Indeed, Feldman et al. (2004) reported that greater pregnancy plannedness was associated with longer leaves for new fathers. In addition, another aspect of parental identity development may involve parents’ growth in confidence in their ability to parent their child successfully. Expectant parents who have already developed high parenting self-efficacy—those who believe themselves to be highly capable of caring for their child—will likely take longer leaves, given that low self-efficacy has been linked to lower task persistence and perceptions that tasks are challenging (Jerusalem and Mittag 1995).

Job Characteristics

Lastly, given the current study’s focus on dual-earner couples, other important predictors related to parents’ jobs may influence parental leave. Although it might seem like new parents who are highly satisfied with their jobs would want to return to work as soon as possible after a child’s birth, either because they enjoy working or because they are afraid of losing a good job, previous research has found that job satisfaction is also related to family-friendly workplace policies, such as those that support parental leave. For example, maternal job satisfaction is greater when mothers are satisfied with their organizations’ family leave policies (Brown et al. 2002). Although far fewer studies have included fathers, there is evidence that family-friendly policies are related to fathers’ job satisfaction as well (Saltzstein et al. 2001). Thus, expectant parents’ job satisfaction may be related to longer leave-taking after a child’s birth, as parents take advantage of the family-friendly workplace supports that they were anticipating.

As an exploratory variable, we also examine how parental work schedules (i.e., working a standard day shift or not) are related to parental leave. Numerous studies have reported that working non-standard work schedules is a strategy sometimes used by parents to split childcare with their partners (Han 2004; Millward 2002) and that non-standard work schedules are more common among individuals who are economically disadvantaged (Leupp et al. 2019). Thus, parents who work non-standard work schedules may be least likely to have access to parental leave, yet no known previous research has explored the link between non-standard work schedules and parental leave.

The Current Study

Given the limited effects of FMLA on leave-taking, especially for fathers, identifying characteristics of parents who take longer leaves from work after the birth of their first child may provide important information for policymakers and advocates for parental leave. In our study we use longitudinal data from dual-earner different-gender couples who were followed over their transition to parenthood to examine if new mothers’ and fathers’ leave duration is more strongly associated with (a) financial characteristics (i.e., access to any type of paid leave, household income) or if (b) demographic and (c) identity-relevant characteristics of expectant parents as well as their (d) job characteristics are also important.

Focusing on financial characteristics, we hypothesized that a higher proportion of paid leave would be positively associated with longer leaves for both mothers and fathers, but we expected this effect to be stronger for fathers given the historical salience of their roles as financial providers (Lamb 2010). Mothers and fathers with higher annual household incomes were expected to report longer leaves from work. Regarding demographic characteristics, we hypothesized that mothers would take longer leaves than fathers. Mothers and fathers with higher levels of education and those who were older were also expected to report taking longer leaves from work.

Turning to identity-relevant characteristics, we hypothesized that fathers who indicated that parenting was their most important role, held more egalitarian gender attitudes in their parenting role beliefs, and did not endorse ideas about maternal essentialism would take longer leaves from work. Similarly, we hypothesized that mothers who indicated that parenting was their most important role and those who felt subjected to perfectionistic parenting standards would take longer leaves from work. Additionally, we predicted that mothers and fathers who expected themselves to be good parents and who reported the pregnancy as more planned would take longer leaves from work. Finally, looking at job characteristics, we hypothesized that higher job satisfaction—especially for mothers—would be associated with longer leaves from work post-birth, given links between job satisfaction and family-friendly workplace policies. Finally, we examined work shift (i.e., a regular daytime shift or not) as an exploratory variable.

Method

Participants

Data were drawn from a longitudinal study of 182 dual-earner different-gender couples living in a large U.S. Midwestern city who were studied across their transition to parenthood beginning in 2008–2009. In order to participate in the study, couples had to be married or cohabiting, working full-time with plans to return to work after the birth of their child, at least 18 years-old, expecting their first child, and the biological parents of the child. Participants were recruited via childbirth education classes, newspaper advertisements, flyers, and participant referrals. At the time of recruitment, couples were informed that the study consisted of “four phases to gather information about their beliefs and expectations for parenthood, personality characteristics, time use, and family relationships.” The present study was reviewed and approved by the sponsoring university’s Institutional Review Board.

Our analytic sample included 130 couples in which both partners were working prior to the birth, returned to work after the birth, and reported the length of time they took off from work after their child’s birth as well as the extent to which their leave was paid. Participants’ ages ranged from 18 to 50 with a mean age for mothers of 28.81 (SD = 3.70, range = 18–41) and for fathers of 30.86 (SD = 4.58, range = 20–50). Eighty-eight percent (n = 115) of couples reported that they were married and 12% (n = 15) reported that they were cohabiting. Couples reported a median annual household income of $82,787 (range = $22,000–$238,000). The sample was highly educated with 79% (n = 103) of mothers and 68% (n = 88) of fathers reporting that they had obtained at least a bachelor’s degree. A majority of participants, 85% of mothers (n = 110) and 86% (n = 112) of fathers, identified as White. A small percentage of the sample identified as African American (5% or n = 6 mothers and 5% or n = 6 fathers), Asian American or Pacific Islander (4% or n = 5 mothers and 5% or n = 6 fathers), Hispanic/Latin American (3% or n = 4 mothers and 1% or n = 1 father), and other (2% or n = 2 fathers). A small percentage of mothers (4%; n = 5) and fathers (1%; n = 1) identified as multiracial. There were no statistically significant differences on key demographic characteristics (i.e., race, education, income, marital status, and age) between the analytic sample and the larger sample.

Procedure

Expectant parents independently completed questionnaires during their third trimester of pregnancy. After the birth of their child, the new parents participated in three follow-up assessments at 3, 6, and 9 months postpartum. At each follow-up, parents reported if they were currently on maternity or paternity leave and when they expected to return to work. If parents were not on maternity or paternity leave, they reported the length of their leave from work and to what extent it was paid and/or unpaid.

Parental Leave Measures

Length

The length of leave that mothers and fathers reported they took off from work after the birth of their child was reported in days and weeks. This variable was coded as the number of days a mother or father took off from work (i.e., 1 week was coded as 7 days). If there were discrepancies in a parent’s leave reporting at the different time points, we used the earliest report of leave time.

Paid or Unpaid

Parents also reported the number of days and weeks of their leave that were paid and unpaid. Paid leave was coded as the proportion of leave that was paid (number of paid leave days divided by the total number of days of leave taken).

Prenatal Measures

Parenting Role Importance

Mothers and fathers were asked to rank “…the roles, in order of importance, of the woman (or man) of the family” from most important (1) to least important (5). The five roles included: parent, spouse/companion, worker/professional, economic provider for the family, and household and/or home caretaker (Perry-Jenkins et al. 1992). Dummy codes were created to indicate whether mothers and fathers did or did not rank the parental role as their most important role (1 = most important; 0 = not the most important). A majority of participants rated either parent (36.9% of fathers or n = 48 and 37.7% or n = 49 of mothers) or spouse/companion (54.6% of fathers or n = 71 and 63.8% of mothers or n = 83) as their most important role.

Parenting Role Beliefs

To measure mothers’ and fathers’ gender attitudes about parenting, we used the 26-item Beliefs Concerning the Parental Role Scale (e.g., “Men should share with childcare such as bathing, feeding, and dressing the child”; Bonney and Kelley 1996; α = .82 for mothers and α = .86 for fathers) in which parents rated the extent to which they agreed with statements on a scale from 1 (disagree strongly) to 5 (agree strongly). Scores were averaged, and higher scores indicated more egalitarian gender attitudes.

Maternal Essentialism

The following three items from the Survey of First-Time Mothers were used (Beitel and Parke 1998) to assess mothers’ and fathers’ views about maternal essentialism: (a) “Mothers are instinctively better caretakers than fathers”; (b) “Fathers have to learn what mothers are able to do naturally in terms of childcare”; and (c) “Mothers are naturally more sensitive to a baby’s feelings than fathers are.” Statements were rated from 1 (disagree strongly) to 5 (agree strongly) (α = .86 for mothers and α = .84 for fathers). Scores were averaged, and higher scores reflected greater maternal essentialism consistent with the notion that mothers have a natural advantage over fathers in parenting.

Parenting Self-Efficacy

Parenting self-efficacy was measured using 10 items that assessed expectant mothers’ and fathers’ perceptions of how good they expect themselves to be at handling different situations with their infant on a scale of 1 (not good at all) to 4 (very good) (e.g., “When your baby is upset, fussy, or crying how good will you be at soothing him or her?”; modeled after Teti and Gelfand 1991; α = .85 for mothers and α = .86 for fathers). Scores were averaged, and higher scores reflected stronger parenting self-efficacy. (The full scale is available as an online supplement.)

Plannedness of Pregnancy

Fathers and mothers reported the extent to which their pregnancy was planned or unplanned on a scale from 1 (not at all planned) to 7 (planned), with the mid-point of 4 as neutral. Mothers’ and fathers’ responses were correlated at .89, so an average score was used in the analyses, and higher scores reflected greater plannedness of the pregnancy.

Maternal Societal-Oriented Parenting Perfectionism

Four items from the Multidimensional Parenting Perfectionism Questionnaire (Snell et al. 2005) were used to measure maternal societal-oriented parenting perfectionism. Mothers reported the extent to which these items—(a) “Only if I am a ‘perfect’ parent will society consider me a good parent”; (b) “Most people expect me to always be an excellent parent”; (c) “In order for people to accept me, I have to be the greatest parent in the world”; and (d) “Most people expect me to be perfectionistic when it comes to being a parent”—were 1 (not at all characteristic of them) to 5 (very characteristic of them) (α = .82 for mothers). Scores were averaged, and higher scores reflected greater perfectionism.

Job Satisfaction

Mothers and fathers reported their current job satisfaction on a single 4-point scale (“All in all, how satisfied would you say you are with your job?”; 1 = not at all satisfied; 4 = very satisfied). Higher scores indicated greater job satisfaction.

Regular Daytime Work Shift

Mothers and fathers answered the following question, “Which of the following best describes the hours you usually work at your main job?” Possible responses were: a regular daytime shift (6 AM – 6 PM), a regular evening shift (any time between 2 PM – midnight), a regular night shift (any time between 9 PM – 8 AM), a rotating shift (changes periodically from days to evenings or nights), a split shift (consisting of two distinct periods each day), and some other schedule. Mothers’ and fathers’ work schedules were coded as dichotomous variables: 0 = not a regular daytime shift; 1 = a regular daytime shift (6 AM- 6 PM).

Demographic Characteristics

Mothers and fathers reported their level of education. Possible responses were: less than high school, high school or GED, vocational or tech program, some college, Associate’s degree, Bachelor’s degree, Master’s degree, and Doctorate degree or equivalent. Mothers’ and fathers’ education were coded as dichotomous variables (0 = less than a bachelor’s degree; 1 = at least a bachelor’s degree). Mothers’ and fathers’ ages (in years) were calculated using their dates of birth and the dates they completed the questionnaire during the third trimester. Mothers’ and fathers’ races/ethnicities were coded as dichotomous variables (0 = non-White and 1 = White). Annual household income for each couple was calculated by taking the average of mothers’ and fathers’ reports of their annual household income in thousands of U.S. dollars. To have an interpretable coefficient, we divide this number by 1000 prior to analyses.

Analytic Plan

First, we conducted preliminary analyses of means, standard deviations, and correlations among all variables. Next, we performed Structural Equation Modeling (SEM) with observed variables in IBM SPSS AMOS Version 25.0 to examine predictors of the length of time that mothers and fathers took off from work after the birth of their first child. Mothers and fathers were analyzed in the same model because they were members of couples and thus their data were not independent. Mothers and fathers in the same couple are distinguishable, and one strength of SEM is that it readily handles distinguishable dyads (Ledermann and Kenny 2017). Furthermore, SEM allows for parameters to be estimated using Full Information Maximum Likelihood (FIML; Arbuckle 1996), which is a preferred method for handling missing data (Johnson and Young 2011). The covariances among all exogenous variables and among disturbance terms of the endogenous variables were estimated. Multiple fit indices were used to assess the fit of the model based on recommendations from Hu and Bentler (1999), including the Chi-square test, the root-mean-square-error of approximation (RMSEA), the Comparative Fit Index (CFI), and the Tucker Lewis Index (TLI). The fit of the model is considered good when the Chi-square test is non-significant (p > .05), the RMSEA is less than .06, and the CFI and TLI are higher than .95.

Results

Preliminary Analyses

Table 1 shows the means, standard deviations, and correlations of all variables for mothers and fathers. Mothers reported an average of 66.76 days of maternity leave with a standard deviation of 26.39 days and a range from zero days to 168.00 days. The median length of leave for mothers was 70.00 days and the mode was 84.00 days. Fathers reported an average of 13.77 days of paternity leave with a standard deviation of 14.54 days and a range from zero days to 84.00 days. The median length of leave for fathers was 10 days and the mode was zero days. A paired samples t-test revealed that mothers’ and fathers’ lengths of leave differed significantly, t(129) = 20.37, p < .001, Cohen’s d = 1.79.

Table 1 Descriptive statistics and correlations of phase 1 predictors and outcome variables
Full size table

Regarding predictors of leave duration, mothers reported that a higher average proportion of their leave was paid (M = .59, SD = .46) compared to fathers (M = .49, SD = .78). Mothers’ (r = .26, p < .01) and fathers’ (r = .42, p < .01) proportions of their leaves that were paid were positively associated with their own lengths of leave from work. Both mothers and fathers reported high levels of expected parenting self-efficacy (scores of 3.38 and 3.32 on a 4-point scale), high levels of egalitarian attitudes for parenting role beliefs (4.30 for both mothers and fathers on a 5-point scale), and high levels of job satisfaction (3.23 and 3.13 on a 4-point scale). Mothers reported average levels of parenting perfectionism (2.49 on a 5-point scale). Fathers endorsed maternal essentialism ideals more highly than mothers (3.22 and 2.56 on a 5-point scale). For mothers, parenting role beliefs and maternal essentialism were highly correlated (r = −.59, p < .01), but this correlation was weaker for fathers (r = −.23, p < .01).

SEM Analyses

The structural model is shown in Fig. 1. The model fit well, with χ2(21) = 21.89, p = .406, CFI = 1.00, TLI = .97, and RMSEA = .02. The model explained 30% of the variance in father’s length of leave and 21% of the variance in mother’s length of leave. When mothers (B = 10.55, β = .19, p = .03) and fathers (B = 8.45, β = .46, p < .001) had a higher proportion of their leave paid, they took longer leaves from work. Older fathers took longer leaves from work (B = .72, β = .23, p = .005). Fathers who more highly endorsed maternal essentialism took shorter leaves from work (B = −2.63, β = −.20, p = .014). Mothers who were more satisfied with their jobs took shorter leaves from work (B = −10.28, β = −.29, p < .001). When couples reported a higher annual household income, mothers took longer leaves from work (B = .19, β = .28, p = .004). When expectant parents reported that their pregnancy was more planned, fathers took shorter leaves from work (B = −1.48, β = −.21, p = .015).

Fig. 1
figure 1

Structural model of predictors of the length of maternity and paternity leave. The correlations among the exogenous variables and the disturbance terms are not shown. Unstandardized regression coefficients are reported for all significant paths. M = mothers; F = fathers. *p < .05. **p < .01. ***p < .001

Full size image

Discussion

The Family and Medical Leave Act represented an important shift in the United States because the government sought to “support families in their efforts to strike a workable balance between the competing demands of the workplace and home” (Commission on Family and Medical Leave 1996, p. xii). However, the employees covered by and the impact of FMLA on leave-taking have been limited. Mounting empirical evidence has linked shorter leaves for mothers to numerous negative outcomes for children and families (Berger et al. 2005; Clark et al. 1997; Gjerdingen and Chaloner 1994; Hyde et al. 2001), whereas longer leaves for fathers have been linked to numerous positive outcomes (Nepomnyaschy and Waldfogel 2007; Petts and Knoester 2019; Rossin 2011). The current study examined if longer leave-taking for mothers and fathers is primarily influenced by financial characteristics or if demographic, identity-relevant, and job characteristics of expectant parents also play a role in their length of leave from work after the birth of their first child. The findings show that—among relatively socioeconomically advantaged new parents—financial characteristics, especially paid leave, are important for leave duration for both parents. In addition, identity-relevant and demographic characteristics mattered for the length of paternity leave, whereas job characteristics were relevant to the length of maternity leave.

Financial Characteristics

As we anticipated, the lengths of leave for mothers and fathers were associated with financial factors. For both mothers and fathers, having higher proportions of leave days that were paid was associated with longer leaves. For every one unit increase in the proportion of leave that was paid, mothers’ leave increased by 10.55 days and fathers’ leave increased by 8.45 days. In addition, greater annual household income was associated with longer leaves for mothers but not for fathers. As household income increased by one unit, mothers’ leave increased by .19 days. These findings suggest that incentivized leave (i.e., leave that is paid) seems particularly important for longer leaves, and they are not surprising given a growing body of literature linking paid leave to longer leaves (Bartel et al. 2018; Rossin-Slater et al. 2013). Additionally, the finding for mothers regarding the role of higher annual household income is consistent with Klerman and Leibowitz’s (1994) finding that when controlling for educational attainment, women with higher wages were more likely to take unpaid leave. Taken together, these findings confirm that financial factors are critical to understanding parental leave-taking behavior, even among relatively socioeconomically advantaged new parents.

Demographic Characteristics

As hypothesized, mothers took significantly longer leaves than fathers. Building on findings that have compared mothers and fathers who take leave with mothers and fathers who do not take leave (Han et al. 2009; Klerman and Leibowitz 1994; Zagorsky 2017), we hypothesized that mothers and fathers who were highly educated and older would report taking longer leaves from work. However, only fathers’ age was a significant predictor of leave duration, such that older fathers took longer leaves. For every one-year increase in fathers’ age, their length of leave increased by .72 days. We speculate that older fathers compared with younger fathers may be more established in their jobs/careers and may be more able to take paternity leave. However, one possible explanation for the general lack of findings related to demographic characteristics is that the length of leave, compared with a dichotomous measure of leave-taking used in previous studies, is less influenced by demographic characteristics. Thus, individual demographic characteristics may be particularly important for understanding who takes leave and who does not, but less important when examining the length of leave among mothers and fathers who do take leave.

Identity-Relevant Characteristics

Situated in identity theory (Stryker 1980; Stryker and Serpe 1982), with consideration of how men’s and women’s role salience has become increasingly complex (Douglas and Michaels 2004; Fitzgerald and Harmon 2001; Hyde et al. 1993; Pleck 1981; Toossi 2002), we hypothesized that parents who reported that parenting is their most important role, the pregnancy was more planned, and expected themselves to be more efficacious parents would take longer leaves from work. We also expected that fathers who held more egalitarian gender attitudes in their parenting role beliefs and did not endorse ideas about maternal essentialism and mothers who felt subjected to perfectionistic parenting standards would take longer leaves. However, only two of these predictors were significant: fathers’ beliefs about maternal essentialism and pregnancy plannedness. For every one unit increase in fathers’ beliefs about maternal essentialism, fathers’ length of leave decreased by 2.63 days, and for every one unit increase in pregnancy plannedness, fathers’ length of leave decreased by 1.48 days. The lack of findings for identity-relevant variables was unexpected, and it indicates that the other predictors included in the current analyses may be more important to parental leave duration. Given that much of the previous research that included gender attitudes and role importance was conducted approximately 15 years prior to when the data in the current study were collected (with the exception of Petts et al. 2018), gender attitudes and role importance may be less critical currently in determining which mothers and fathers take leave and who takes longer leaves. However, analyses should be replicated in a more diverse and representative sample of new parents in the United States who may have greater variation in their gender role attitudes and role importance than the current sample.

Our hypothesis that fathers who held more egalitarian gender attitudes would take longer leaves from work was not supported. Fathers’ beliefs about parental roles did not predict their leave duration. However, fathers who endorsed ideas about maternal essentialism took shorter leaves from work. It may be the case that the maternal essentialism measure is a more sensitive assessment of fathers’ attitudes about the caregiving and nurturing roles inherent to caring for newborn infants than more general measures about parent and gender roles. This finding has important implications for prevention and intervention efforts because fathering has been shown to be malleable at the transition to parenthood (Doherty et al. 2006). Fostering fathers’ beliefs about their natural abilities to care for their newborn infants may directly influence the length of their paternity leave at the transition to parenthood, but additional research is needed to further substantiate this potential link.

We also found that the extent to which the pregnancy was planned was unrelated to mothers’ leave-taking and negatively associated with fathers’ leave-taking, such that when couples indicated that their pregnancy was more planned, fathers took shorter leaves from work. The finding for mothers is consistent with Feldman et al. (2004), who also found no association between pregnancy plannedness and mother’s length of leave. However, the finding for fathers is opposite of what Feldman and colleagues reported and contrary to our hypothesis. There are multiple possible explanations for this unexpected finding. First, Feldman et al. (2004) used a dichotomous measure of pregnancy plannedness (yes/no), whereas our measure used a 7-point scale. This may suggest that pregnancy plannedness, when measured on a scale, predicts paternity leave length differently. Additionally, we measured pregnancy plannedness during the third trimester of pregnancy, whereas Feldman et al. measured pregnancy plannedness after the birth of the child, which may suggest that the timing of when this question is asked matters. In the current sample, approximately 45% of couples indicated that their pregnancies were highly planned (7) and only 10% indicated that their pregnancies were not at all planned (1). Thus, the finding in the current study may also be a product of our sample, which consisted of primarily highly educated, middle class, White, married families, given that Finer and Henshaw (2006) suggest that approximately half of all pregnancies in the United States are unplanned and that planned pregnancies are more common among married women compared with cohabiting and noncohabiting unmarried women (Musick 2002).

Job Characteristics

Regarding job characteristics as predictors of duration of parental leave, we found no evidence that work shift was related to duration of leave in this sample. However, we urge future researchers with access to a more diverse sample of parents to consider exploring the link between work schedules and parental leave given that parents who work non-standard work schedules may be among the most disadvantaged and the least likely to have access to parental leave (Leupp et al. 2019). Moreover, only mothers’ (and not fathers’) job satisfaction was associated with their leave-taking. We found that mothers who were more satisfied with their jobs took shorter leaves from work, which was in the opposite direction than hypothesized. For every one unit increase in mothers’ satisfaction with their jobs, their length of leave decreased by 10.28 days. Although we had expected that greater job satisfaction would be linked to longer leaves, given that both mothers’ and fathers’ greater job satisfaction has been linked to family-friendly workplace policies (Brown et al. 2002; Saltzstein et al. 2001), it is not necessarily surprising that new mothers who are satisfied with their jobs may be eager to return to work, possibly because they enjoy their work or because they are afraid of losing a good job.

Future research on the role of job characteristics in parental leave-taking should take a more nuanced approach to assessing different components of job satisfaction, including specific (e.g., satisfaction with leave policies) and global aspects, and considering other factors that may moderate relations between job satisfaction and leave-taking, such as supervisor support. Moreover, given the high levels of education of parents in our sample, these parents were more likely to be pursuing employment as part of careers compared with the general population of new parents. Therefore, associations between job characteristics and parental leave may differ for expectant parents with lower levels of education and who are working but not pursuing careers.

Limitations and Future Research Directions

As we have emphasized throughout the discussion of findings, given the sociodemographic characteristics of our sample, the results of the present study are not generalizable to the larger population of new parents in the United States. In particular, given that the educational attainment and income levels of the current sample reflect a relatively advantaged sample of new parents, it is important to acknowledge that the parents in our sample may be more likely to have access to FMLA or additional leave benefits from their employers, such as paid leave. The current sample’s advantaged status likely influences their leave-taking behaviors in ways that are not applicable to the larger population of new parents in the United States. The couples who participated in our study were also those willing to voluntarily participate in an intensive longitudinal study of the transition to parenthood, and thus they may have been especially committed to parenthood compared with expectant parents in the general population.

Another limitation of the current study is that given the design of the questionnaire, we could not determine if parents were utilizing leave provided by FMLA, their employer, accrued sick and/or vacation days, or a combination of the three. Future research should collect more detailed information about the sources of paid and unpaid leave because they may differentially influence parental leave-taking and duration. In addition, the current sample participated in our study during the economic recession of 2008–9, which was characterized by many Americans losing their jobs (Elsby et al. 2010); thus, the leave-taking behavior of expectant parents during this time period may have been affected by these broader economic circumstances as well.

Finally, given the eligibility criteria for the current study, we excluded parents who did not report leave and those who were unemployed after their child’s birth. A limitation of the current data is that we are unable to determine why parents were unemployed post-birth, and we propose that more nuanced data are needed to fully understand if unemployment after the transition to parenthood is linked to factors such as the lack of availability of paid parental leave, employers penalizing new parents who take leave, or personal decisions.

Practice Implications

The length of maternity and paternity leaves is multiply determined, which has important implications for both policymakers and practitioners who work with expectant and new parents. Given that mothers’ and fathers’ leave-taking duration seems strongly influenced by financial characteristics, especially the availability of paid leave, U.S. policies that support paid parental leave for mothers and fathers may be key to increasing leave usage. For practitioners, given the role of identity-relevant characteristics in fathers’ leave-taking duration, attention to helping fathers further develop their parental identities in the realms of caregiving and nurturing may motivate more fathers to take longer leaves from work after their child’s birth.

Conclusions

The current study provides an important starting point for understanding how financial, demographic, identity-relevant, and job characteristics are associated with the length of maternity and paternity leave for new parents in dual-earner families in the United States. Taken together, the findings suggest that in order for U.S. families, including fathers and low-income mothers, to reap the benefits of parental leave, leave that is financially incentivized would be most beneficial. In addition, attention to the development of expectant fathers’ capabilities as caregivers and nurturers of newborn infants may encourage more fathers to spend valuable time at home developing close and lasting relationships with their children.