Abstract
The notion that U.S. mothers with minor children are less happy and more depressed than nonmothers largely relies on data collected in the 1990s or earlier. Although the coronavirus disease 2019 pandemic brought much attention to the stressfulness of parenting, we lack knowledge of how mothers fared relative to nonmothers in the 2000s and 2010s, before the pandemic. The authors investigate trends in the parenthood gap in happiness, depression, and self-rated health among women aged 18 to 59 years, using the 1996 to 2018 General Social Survey (n = 13,254) and the 1997 to 2018 National Health Interview Survey (n = 263,110). Results indicate that twenty-first-century mothers with younger children were better off than nonmothers on two measures, reporting less depression and better health. Mothers’ “depression advantage” grew across this time. However, mothers with older children reported less happiness than nonmothers, a continued trend from the 1990s. The study underscores the importance of examining various well-being indicators across the changing contexts of parenting.
Widespread school and daycare closures and subsequent burdens on mothers in supervising remote learning and providing additional care for children during the coronavirus disease 2019 (COVID-19) pandemic brought much public and scholarly attention to the demands of parenting and their implications for well-being. Parents, especially mothers, reported high stress and poor mental health during the pandemic in the United States and other countries (Adams et al. 2021; Elder and Greene 2021; Hart and Han 2021; Kerr et al. 2021; Lai et al. 2022; Montazer et al. 2022; Patrick et al. 2020; Russell et al. 2022; Singletary et al. 2022; Yan et al. 2022; Zamarro and Prados 2021). Most of these studies used data that were collected only after the pandemic hit (with some exceptions, e.g., see Montazer et al. 2022) and often lacked comparative data to those not in the parenting role. To better understand the situation of parents, it is important to know how mothers were faring compared with nonmothers before the onset of the pandemic.
The notion that mothers in the United States are worse off (e.g., more distressed or less happy than women without children) is widespread (McLanahan and Adams 1987; Simon and Caputo 2019). These ideas are drawn largely from studies using data collected in the 1990s or earlier (McLanahan and Adams 1987; Simon and Caputo 2019; Umberson, Pudrovska, and Reczek 2010). Since the 1990s, parenting norms and behaviors have changed (Altintas 2016; Cha and Park 2021; Nomaguchi and Milkie 2019) and happiness and health among American adults worsened (Beck et al. 2014; Dutta and Foster 2013; Twenge et al. 2019), suggesting the merit of investigating trends in the parenthood gap in happiness and health in the 2000s and 2010s.
This article renews and updates knowledge of the parenthood gap in happiness, depressive symptoms, and self-rated health among U.S. women aged 18 to 59 years prior to the pandemic, using the 1996 to 2018 General Social Survey (GSS) and the 1997 to 2018 National Health Interview Survey (NHIS). Research on parenthood and well-being has increasingly suggested the importance of investigating various aspects of well-being, defining it broadly to include subjective well-being (e.g., happiness), mental health (e.g., depression), and physical heath. Each of these indicators has implications for mothers’ subsequent childbearing, and for child well-being, and thus for the society as a whole (e.g., Tosi and Goisis 2021; Turney 2012; for a review, see Nomaguchi and Milkie 2020). Yet prior research tends to examine only one of them (e.g., Glass et al. 2016). By examining the three outcomes, this study provides a fulsome picture of how having minor children in the household is related to adults’ health and subjective well-being in the twenty-first century, prior to the COVID-19 pandemic. Because research shows that parents’ subjective well-being and mental health are better when children are very young than when children are school age or teenage (Meier et al. 2018; Nomaguchi 2012), we divide parental status into two groups by the age of the youngest child.
The Parenthood Gap in Happiness and Health: Theoretical Perspectives
As the demands-rewards perspective posits, parenting involves both challenges and rewards (Nomaguchi and Milkie 2020). The demands of caring for children—physical, financial, and mental investments and effort—can be a major stressor, especially when children have special needs (Sellmaier 2021) or when parents lack resources, like being a single parent, having difficulties in arranging childcare, or facing weak institutional support for raising children (Glass, Simon, and Andersson 2016; Pollmann-Schult 2018; Stier and Kaplan 2020). Parenting can generate other (or secondary) stressors, such as financial strain, partnership strain, and work-family conflicts (Blanchflower and Clark 2021; Pollmann-Schult 2014). The rewards of parenting, rooted in the idea of role enhancement theory (Moen, Kelly, and Huang 2008), refer to aspects of parenting that allow adults to gain prestige, resources, and social identity. Children bring joyful and meaningful experiences that can enhance parents’ happiness and health (Musick, Meier, and Flood 2016; Nelson et al. 2013). Young children can strengthen adults’ ties with their parents and siblings (Aldrich, Nomaguchi, and Fettro 2022; Min et al. 2022) and broaden adults’ social networks in the larger community (Nomaguchi and Milkie 2003). The parenting role fosters a sense of responsibility in adults for maintaining a routine and healthy lifestyle, such as eating balanced meals and reducing alcohol use, which can promote better health (Reczek et al. 2014; Simon and Caputo 2019). Scholars agree that the parenthood gap in happiness and health depends largely on the context of parenting, including the age of children, partnership status, socioeconomic status, and country context, in large part because the context influences levels of demands and rewards of parenting as well as resources that help parents cope with the demands (for reviews, see Nomaguchi and Milkie 2020; Umberson et al. 2010).
The life-course perspective on parental well-being contends that historical time is a key context that shapes levels and types of demands and rewards of parenting (Nomaguchi and Milkie 2020; Umberson et al. 2010). As economic circumstances change, the meaning of childhood changes, and so do social norms regarding what parents are expected to do for and with children (Hays 1996). Parenting strains involve individuals’ subjective perception that they feel it difficult to meet socially expected demands of the parenting role (Pearlin 1983). Thus, as parenting norms change, the levels and types of parenting strains change over time. Much recent research on the parenthood gap in health and well-being using data collected in the 2010s came from European counties, most of which focused on happiness (e.g., Dotti Sani 2022; Pollmann-Schult 2018; Preisner et al. 2020; Radó 2020). In contrast, recent research examining the U.S. context used data collected in the mid-1990s (Simon and Caputo 2019) or in the mid-2000s (Glass et al. 2016), or pooled data collected from the early 1970s with data from the late 2010s together (Blanchflower and Clark 2021). The goal of this work is to renew knowledge about the parenthood gap in health and well-being among U.S. women by examining the trends from the late 1990s to the late 2010s, prior to the pandemic.
Trends in the Parenthood Gap in Happiness and Health from the 1990s to the 2010s
To conceptualize and predict potential change in the parenthood gap in happiness, depression, and self-rated health from the late 1990s to the late 2010s, we rely on two lines of research: (1) trends in parenting norms and behaviors and (2) overall trends in happiness, depression, and self-rated health among American adults. As we describe below, together, recent findings in these areas of research suggest the possibility that some advantages in mothers’ well-being compared with nonmothers’ might have emerged in the 2010s.
Trends in Parenting Norms and Behaviors
Researchers agree that parenting norms changed from the 1970s to the 1990s toward childrearing that is more child centered, expert guided, and emotionally and physically labor intensive, which Hays (1996) called “intensive mothering.” Some suggest that these parenting emphases may have intensified or widened across groups throughout the 2000s and the 2010s (e.g., Faircloth 2014). Intensive mothering ideology holds mothers accountable for how children will turn out, emphasizing that children’s talents and skills must be carefully cultivated by their primary caregivers in order for them to thrive physically, cognitively, emotionally, and socially (Faircloth 2014). The ideology posits that mothers’ ample, one-on-one attention to each individual child is essential (Milkie and Warner 2014; Wall 2010). Although such beliefs are common among parents across the class spectrum (Hays 1996; Ishizuka 2019), childrearing practices and enactment of the ideology differ for middle- versus working-class mothers (Hays 1996; Lareau 2003). Consistent with intensive mothering norms, maternal time spent in developmental childcare activities with young children increased from the 1970s to the 2010s, with some time lag by education level. Mothers with a bachelor’s degree increased their time more rapidly than mothers without a bachelor’s degree from the 1970s to the early 2000s, which resulted in a widening gap in maternal childcare time by education (Altintas 2016; Sayer, Bianchi, and Robinson 2004). From the early 2000s to the late 2010s, mothers without a bachelor’s degree increased their childcare time steadily, whereas mothers with a bachelor’s degree decreased their childcare time slightly, narrowing the gap to the level before the early 2000s (Cha and Park 2021; Prickett and Augustine 2021).
Reasons for the intensifications of intensive mothering ideology and behavior in the twenty-first century include the increase in perceived job insecurity among American adults, derived largely from the changing nature of careers and the job market (Milkie and Warner 2014; Nomaguchi and Milkie 2020). As the requirement of a bachelor’s degree to secure a decent job has increased, competition to get into universities has risen. In turn, parents, especially mothers, increasingly spend time with children in educational activities and organizing and attending children’s extracurricular activities, which they believe are critical for children to gain their chance to be admitted to “good” universities (Ramey and Ramey 2010). Studies tracking American adults’ values for children using the GSS have shown that American adults’ support for “hard work” relative to “autonomy” increased from 1986 to 2018, which arguably indicates that more American adults recognize that young people need to work hard to secure opportunities in the ever competitive world of schooling and jobs (Nomaguchi and Milkie 2019).
Many researchers assert that intensive mothering ideology undermines mothers’ well-being. Furedi (2002) contends that through the idea that parents are responsible for preventing any harms from happening to their children, the expert-guided intensive mothering culture normalized “parent bashing.” The ideology imposes a sense of pressure, anxiety, and guilt on mothers, who often feel as if they are not doing enough for their children (Faircloth 2014; Milkie, Nomaguchi, and Schieman 2019; Milkie and Warner 2014), and who must act as individualized safety nets to support children as they grow, given very weak institutional supports (Milkie and Warner 2014). Intensive mothering’s emphasis of self-sacrifice as a virtue of a “good” mother encourages mothers across the class spectrum to prioritize their children’s well-being to the detriment of their own (Elliott, Powell, and Brenton 2015; Hays 1996; Wall 2010). Gunderson and Barrett (2017), using data from the 2004 to 2006 Midlife in the United States, find that mothers who spend more time providing children with emotional support are more likely than mothers who spend less time in this activity to report depressive symptoms. The ever pervasive intensive mothering ideology, thus, may mean that the parenthood gap in happiness and health might have widened during the 2000s and 2010s, compared with the 1990s.
At the same time, there are some signs of a retreat from intensive mothering ideology in public discourses in the late 2000s and the 2010s. The idea of “free-range” parenting (Skenazy 2009) advocates for less hands-on parenting, emphasizing the importance of recognizing children as resilient and letting them experience reasonable independence. When a public debate about how much supervision children need was sparked in the mid-2010s, many media outlets published opinions that supported free-range parenting (Wergin 2015). Law scholars have begun criticizing the current state and federal child neglect statues as highly biased toward intensive parenting ideology and thus discriminatory against parents who use a more hands-off approach to parenting (Manno 2016). In academic research, Christopher (2012) argues that the majority of full-time employed mothers in her study reject at least some intensive mothering ideals, constructing good mothering as delegatory of childrearing tasks while remaining responsible for children. As mentioned earlier, mothers with a bachelor’s degree decreased their childcare time slightly from the 2000s to the 2010s (Cha and Park 2021; Prickett and Augustine 2021). Using data from the 2010 to 2013 Well-Being module of the American Time Use Survey, studies report joy and fulfillment that mothers experience from being involved in children’s lives (Milkie, Wray, and Boeckmann 2021; Musick et al. 2016). For example, Musick et al. (2016) found that mothers are happier when they spend time in activities with children compared with those where children are not present. Taken together, these recent public discourses and research findings suggest that we might see some changes in the patterns in the parenthood gaps in happiness and health in the 2010s toward a motherhood advantage. As we will see in the next section, research on overall trends in happiness and health among American adults during the 2010s also indicates this possibility.
Overall Trends in Happiness and Health
To better understand trends in the parenthood gap in health and well-being, it is important to place them in the context of overall trends in health and well-being among American adults during the period. Studies produced consistent findings that on average American adults aged 18 and older reported lower levels of happiness in the 2000s and the early 2010s than they did before. Using the 1972 to 2010 GSS, Dutta and Foster (2013) showed that the average happiness rating increased from the 1970s to the 1980s, declined in the 1990s, and further declined in the 2000s. Using the 1972 to 2014 GSS, Twenge, Sherman, and Lyubomirsky (2016) showed that the declining trend in happiness was concentrated among adults aged 30 or older, while young adults aged 18 to 29 were happier in more recent years than earlier years. Twenge et al. also note that cohort effects constitute only a small portion of the trends in happiness and that period effects seem to be the primary contributor. Cummings (2020), using the 1972 to 2016 GSS, shows that the trends differed markedly by the intersection of gender and race, suggesting that it is important to examine subgroup differences in happiness trends.
Similarly, several studies show that the rates of depressive symptoms among American adults increased from the mid-2000s to the late 2010s. Using data from the 2005 to 2017 National Survey on Drug Use and Health, Weinberger et al. (2018) found that rates of major depressive episodes during the previous 12 months increased over time. Using the Kessler-6 Distress Scale (Kessler et al. 2012) to measure psychological distress during the previous 30 days in the 2005 to 2017 National Survey on Drug Use and Health, Twenge et al. (2019) found that major depression increased, especially among adolescents and young adults, and more among women than men. Other researchers, using data from the 2005 to 2016 National Health and Nutrition Examination Survey, also report that rates of depressive symptoms during the preceding two weeks increased among American adults (Kauffman et al. 2021; Yu et al. 2020). In contrast to Twenge et al. (2019), Yu et al. (2020) showed that the increase was concentrated among those aged 65 and older.
A worsening trend is also found in self-rated health from the 1990s to the 2000s. Using data from the 1993 to 2001 Behavioral Risk Factor Surveillance Surveys, Zack et al. (2004) reported that the percentage of Americans with fair or poor self-rated health increased from the late 1990s to 2001. Self-rated health worsened in most demographic groups, especially adults 45 to 54 years old, those with a high school diploma but no further degrees, and those with annual household incomes less than $50,000. Beck et al. (2014), using the 1972 to 2008 NHIS, found that the percentage of Americans with fair or poor self-rated health decreased from the 1970s throughout the 1990s and then increased in the 2000s. Research extending the examination of the trends in self-rated health among American adults to the 2010s is scarce.
Researchers speculate on the reasons for the increases in depressive symptoms and self-rated poor health and the decrease in happiness during the 2000s and 2010s. It is worth mentioning these speculations, because these ideas help us conceptualize whether the worsening trends in health and well-being in the 2000s and 2010s differed for mothers and nonmothers. Weinberger et al. (2018) point out the increases in perceived levels of stress, job loss, and financial strains, in part related to the 2008 recession. Hidaka (2012) argues that besides the increase in economic inequality, the increases in social isolation, sedentary lifestyle, poor diets, and sleep deprivation in recent decades may have contributed to poor physical and mental health. Twenge (2015), on the basis of the finding that the increase in depressive symptoms was largely concentrated to somatic symptoms rather than a direct item on depression, speculates that the pressures and lifestyles in contemporary U.S. culture may lead people to experience sleeping issues and difficulty concentrating, which seem to be driving the increase in depressive symptoms. Researchers who study young adults’ mental health find that social media, which quickly increased in terms of platforms and users from the mid-2000s to the mid-2010s (Auxier and Anderson 2021), is a key source of the increasing prevalence of depressive symptoms (Shensa et al. 2017).
Having minor children at home might prevent adults from getting into some of the circumstances that can lead to depression, such as social isolation and unhealthy lifestyles (Reczek et al. 2014; Simon and Caputo 2019), suggesting that mothers are less likely than nonmothers to be prone to these “unhealthy” daily habits that seem to be leading more American adults than before to feeling less happy and more depressed. Indeed, Herbst and Ifcher (2016), using the 1975 to 2007 GSS, found that parents’ happiness relative to nonparents increased over time, especially during the early 2000s, and that the narrowing parenthood gap in happiness was driven by the decline in nonparents’ happiness. We extend Herbst and Ifcher’s study by (1) focusing on more recent decades, (2) restricting the respondents to be younger than age 60 to reduce heterogeneities among nonmothers, and (3) examining depressive symptoms and self-rated health as well as happiness. We focus on mothers, whereas Herbst and Ifcher looked at “parents” without separating them by gender, because the parenting role and parental well-being differ by gender (Nomaguchi and Milkie 2020; Umberson et al. 2010).
The Present Study
On the basis of prior research suggesting changing parenting norms and behaviors during the recent two decades, as well as general worsening trends of health and well-being among American adults during the period, this study examines patterns in the parenthood gap in happiness, depression, and self-rated health. The focus is on well-being in the 2000s and 2010s compared with the 1990s, among U.S. women aged 18 to 59 years, using data from two major national data sets. Such knowledge may provide new insights into understanding social determinants of parenting strains and parental well-being. Drawing on foregoing discussions, we expect that rates of unhappiness, depression, and poor health increased since the 1990s, particularly in the 2010s, and the increases in such rates were less pronounced for mothers than nonmothers. Analyses control for age, race, education, marital status, employment status, family income, given prior findings of their associations with happiness, depression, and self-rated health, trends in these measures over time, and parenting strains and rewards, as reviewed above (e.g., Cummings 2020; Glass et al. 2016; Lamidi 2020; Nomaguchi and Milkie 2020; Simon and Caputo 2019; Twenge et al. 2019; Umberson et al. 2010; Zack et al. 2004).
Methods
Data are drawn from two major national surveys that have been collected over time in the United States, the GSS for analyses of happiness and the NHIS for analysis of depressive symptoms and self-rated health. We restrict the samples to adults younger than 60 years to make the comparison groups between parents and nonparents more relevant.
The GSS
The GSS is a nationally representative sample of adults 18 years of age or older, living in noninstitutional arrangements in the United States, conducted since 1972 by National Opinion Research Center at the University of Chicago. Until 2004, only English-speaking persons were included; since 2006, Spanish-speaking persons have been added to the sampling target population (Smith et al. 2019). Data for this study are drawn from 12 years of GSS conducted in 1996, 1998, 2000, 2002, 2004, 2006, 2008, 2010, 2012, 2014, 2016, and 2018 (n = 32,434). For the present analysis, we selected respondents aged 18 to 59 years (n = 24,204) and those who identified themselves as women (n = 13,254).
Happiness is measured by the question “Taken all together, how would you say things are these days—would you say you are very happy, pretty happy, or not too happy?” The response ranges were coded as 1 = not too happy, 2 = pretty happy, and 3 = very happy.
Parental status is measured as three categories, by using the information on the number of children the respondents had ever had and the age of youngest child living in the household, including (1) the youngest child is younger than 6, (2) youngest child aged 6 to 17, and (3) no child younger than 18. 1
Year of the Survey
We examined the survey year as dummy variables as well as a continuous variable (0 = 1996, 2 = 1998, . . ., 12 = 2018). The results were similar; we present results using the dummy variables.
Controls
The multivariate analyses control for factors that are related to both parental status and happiness. Age is measured in years, ranging from 18 to 59. Race is measured as three groups, including White (reference), Black, and other race. Education is measured as four categories: (1) less than high school, (2) high school diploma, (3) some college, and (4) bachelor’s degree or higher. Marital status includes (1) married (reference); (2) divorced, separated, or widowed; and (3) never married. Weekly hours of employment is measured as hours worked in the main job in the previous week. Family income is measured by the inflation-adjusted family income recoded by GSS investigators, coded in thousands.
The NHIS
The NHIS is a cross-sectional household interview survey of the noninstitutionalized population of the United States residing within the 50 states and the District of Columbia at the time of the interview, conducted by the National Center for Health Statistics since 1957. The questionnaire was revised in 1997 and again most recently in 2019. We thus use data from 1997 to 2018. From each family in the NHIS households (the vast majority of the households, 97 percent, had one family), one “sample adult” aged 18 years or older is randomly selected. For this study, we selected sample adults who were aged 18 to 59 (n = 482,087) and who identified as women, providing a sample size of 263,110.
Depressive symptoms are measured using the Kessler-6 Scale of Serious Depressive Symptoms, which is the sum of the six questions “During the past 30 days, how often did you feel (1) so sad that nothing could cheer you up, (2) nervous, (3) restless or fidgety, (4) hopeless, (5) that everything was an effort, (6) worthless? (0 = none, 1 = a little, 2 = some, 3 = most, 4 = all of the time).” We examine two types of dichotomous variables, (1) moderate or serious depression, where those who scored 5 or higher are coded 1 and those who scored lower than 5 are coded 0 (Prochaska et al. 2012), and (2) serious depression, where those who scored 13 or higher are coded 1 and those who scored lower than 13 are coded 0 (Kessler et al. 2012).
Self-rated health is measured by the question “Would you say your health in general is excellent, very good, good, fair, or poor?” Following prior research (Beck et al. 2014; Lamidi 2020; Zack et al. 2004), a dichotomous variable is created on which those who reported poor or fair health are coded 1, and those who reported good, very good, or excellent health are coded 0.
Parental status is measured as the presence and the age of youngest child in the sample adult’s family, including (1) the youngest child is younger than 5, (2) youngest child is aged 5 to 17, and (3) no child younger than 18. The NHIS uses a different cut point for the age of the young child from that used in the GSS for public use data. Like the GSS, the NHIS 1997 to 2018 public data do not provide information on whether the children living in the family are the sample adults’ children, nor the information about whether the sample adults have any children. We created the respondents’ parental status variable by using information about the sample adults’ family structure (i.e., one adult, no children younger than 18; multiple adults, no children <18; one adult, one or more children younger than 18; multiple adults, one or more children younger than 18), whether the sample adults’ families have anyone younger than 5 living in the household, and the sample adults’ relationship to the family reference person. The sample adults whose relationship with the family reference person were other than “self” or “spouse/partner”—child, grandchild, parent, grandparent, aunt or uncle, cousin, or unrelated—were coded 0 (nonparents), because the majority of them were children of the reference person who were aged 18 to 24, who seem to be young adults living in their parental home where children younger than 18 were also living, who were most likely their younger siblings. 2
Year of the Survey
We examine year of the survey as dummy variables as well as a continuous variable (0 = 1997, 1 = 1998, . . ., 21 = 2018). The results were similar; we present results using the dummy variables.
Controls
Age is measured in years, ranging from 18 to 59 years. Race/ethnicity is measured as four groups, including White (reference), Black, Hispanic, and other race. Marital status includes married (reference), cohabiting, divorced or separated, and never married. Education is measured as four groups: (1) less than high school, (2) high school diploma, (3) some college, and (4) a four-year college degree or more. Weekly hours of work for pay range from 0 to 60 (top-coded with the 95th percentile). Family income is recoded by the NHIS investigators, measured as (1) less than $35,000, (2) $35,000 to 74,999, and (3) $75,000 and more. In the 2007 to 2018 surveys, the last category was separated into two categories, $75,000 to 99,999 and $100,000 and more, but we collapsed these categories together to be consistent with the number of categories in the 1997 to 2006 surveys.
Analytical Plan
Missing data were imputed by using a multiple imputation technique in SAS (Berglund 2010; D’Agostino McGowan and Toll 2015). To examine trends in the association between parental status and happiness from 1996 to 2018, we use the pooled sample of the 1996 to 2018 GSS to conduct ordered logistic regression models with control variables. All analyses are weighted to adjust for the multistage probability design of the data across the years of surveys (Smith et al. 2019). To examine trends in the association between parental status and moderate to serious depressive symptoms or self-rated health from 1997 to 2018, we use the pooled sample of the 1997 to 2018 NHIS to conduct logistic regression models with control variables. Following National Center for Health Statistics (2019), all analyses are weighted using the sample adult weight which includes design, ratio, nonresponse and poststratification adjustments for sample adults. To adjust the sample design, the PROC SURVEYMEANS and PROC SURVEYLOGISTIC commands are used in SAS (D’Agostino McGowan and Toll 2015). For each dependent measure, two models are examined. Model 1 examines main effects of parental status and year of the survey with control variables. Model 2 adds interaction terms between parental status and year of the survey to model 1 to examine whether the association between parental status and happiness, depression, or self-rated health varies by year.
Results
The descriptive statistics for variables from the GSS and NHIS samples of U.S. women aged 18 to 59 are presented in Table 1. In the GSS sample (left columns), 20 percent had young children younger than 6, 22 percent had children aged 6 to 17 but no younger children, and 58 percent had no child younger than 18 living in the household. The distribution for happiness was 12 percent “not too happy,” 56 percent “pretty happy,” and 32 percent “very happy.” In the NHIS sample, 19 percent had at least one young child younger than 5, 26 percent had children aged 5 to 17 but no younger children, and 55 percent had no child younger than 18 living in the household. Twenty-two percent experienced moderate or serious depressive symptoms in the preceding 30 days, with four percent experiencing serious depression. About 10 percent reported fair or poor health.
Proportion or Mean (SD) for the Variables.
Note: General Social Surveys (GSS) 1996 to 2018 (12 surveys) are pooled; means are weighted. National Health Interview Surveys (NHIS) 1997 to 2018 (22 surveys) are pooled; means are weighted and adjusted for the survey design.
The GSS and the NHIS public use data employ different cut-points for “young children.”
Descriptive statistics for happiness, depression, and self-rated health by survey year are presented in Tables 2 and 3. The percentage distribution for reporting “very happy,” “pretty happy,” or “not too happy” among U.S. women aged 18 to 59 years did not show increasing or decreasing trends in the 2000s and 2010s, compared with those in the 1990s. In contrast, the proportions of U.S. women experiencing depressive symptoms and fair or poor health were higher in the 2010s compared with earlier decades.
Percentage Distribution for Happiness among U.S. Women Aged 18 to 59 Years: General Social Survey (n = 10,118).
Note: Data are weighted.
Proportions of U.S. Women Aged 18 to 59 Years Reporting Depression or Poor Health: National Health Interview Survey (n = 263,110).
Note: Data are weighted and adjusted for the design effects.
The results of the ordered logistic regression models for happiness are shown in Table 4. With the pooled sample, mothers whose youngest child was aged 6 to 17 years were less likely than women living with no minor children to score “very happy” compared with scoring “pretty happy” or “not too happy,” whereas mothers whose youngest child was younger than 6 did not significantly differ from nonmothers (Model 1). In supplemental analysis, we found that the odds of reporting being “very happy” were significantly lower for mothers whose youngest child was aged 6 to 17 compared with mothers with younger children. Unlike the findings of prior research (Dutta and Foster 2013; Herbst and Ifcher 2016; Twenge et al. 2016), we did not find a declining trend in happiness among American women from 1996 to 2018. Such inconsistencies could be due to the differences in the study samples: although we restricted the sample to women younger than 60, the prior studies did not. Model 2 shows that differences in happiness by parental status varied little across years. As shown in Figure 1, during the two decades, the adjusted percentage for reporting “very happy” was 29.7 percent for mothers whose youngest child was younger than 6 and 28.6 percent for mothers whose youngest child was aged 6 to 17, compared with 31.8 percent for nonmothers. The adjusted percentage of reporting “not too happy” compared to “very happy” or “pretty happy” was 10.8 percent for mothers whose youngest child was younger than 6 and 11.3 percent for mothers whose youngest child was aged 6 to 17, compared with 9.9 percent for nonmothers.
Ordered Logistic Regression Models Predicting Happiness among U.S. Women Aged 18 to 59 Years: General Social Survey 1996 to 2018.
Note: Happiness is measured as 1 = not too happy, 2 = happy, and 3 = very happy. Omitted reference categories are no child younger than 18 in the household, year 1996, parental status × year 1996, White, married, and high school diploma. Data are weighted. N = 263,110.
p < .05. **p < .01. ***p < .001.
Adjusted percentage distributions for reporting “very happy,” “pretty happy,” or “not too happy” by parental status among U.S. women aged 18 to 59 years: General Social Survey 1996 to 2018.
For depressive symptoms, the results from the logistic regression models are presented in Table 5. We examine the results for moderate or serious depression first (left columns). As Model 1 shows, mothers with young children had a lower rate of experiencing moderate or serious depressive symptoms in the previous 30 days than women not living with minor children. Mothers whose youngest children were aged 5 to 17 differed little from nonmothers in rates of moderate or serious depressive symptoms. Supplemental analysis indicates that the odds are significantly higher for mothers with older children than mothers with young children. The coefficients for the year variable suggest that the prevalence of moderate to serious depressive symptoms increased around 2007 to 2009, during the Great Recession, then declined in 2011 and 2012 back to the level in 1997, and then increased again in 2013 and further climbed up in 2017 and 2018. The interaction between year and parental status (Model 2) is significant for several years, especially during the 2010s, indicating that the trends varied by parental status. Predicted probabilities of experiencing moderate or serious depressive symptoms for the three parental status groups across 22 years from 1997 to 2018 are shown in Figure 2A. In 1997, the predicted probabilities of experiencing moderate or serious depressive symptoms differed little across the three groups with probabilities ranging from 0.219 to 0.233. After the increase in the prevalence of depression during the Great Recession, the decrease in the prevalence of depression in 2011 and 2012 was steeper for mothers than for nonmothers. From 2013 to 2018, the prevalence of depression increased less rapidly for mothers than for nonmothers. In 2017, the predicted probability of experiencing moderate or serious depressive symptoms was 0.289 for nonmothers, whereas it was 0.224 for mothers with young children and 0.243 for mothers whose youngest child was aged 5 to 17.
Logistic Regression Models Predicting Moderate or Serious Depressive Symptoms in the Past 30 Days among U.S. Women Aged 18 to 59: National Health Interview Survey 1997 to 2018.
Note: Omitted reference categories are no child <18 years in the household, year 1997, parental status × year 1997, White, married, and high school diploma. Data are weighted and adjusted for the design effects. N = 263,110.
p < .05. **p < .01. ***p < .001.
Predicted probability of reporting moderate or serious depressive symptoms in the past month: U.S. women aged 18 to 59 years, National Health Interview Survey 1997 to 2018.
Results for serious depression (right columns in Table 5) show that overall (Model 1), again, mothers with young children, but not mothers with children aged 5 to 17, were less likely than nonmothers to report having serious depression. The prevalence of serious depression did not change much during the Great Recession but increased during the late 2010s. The interaction model (Model 2) and Figure 2B suggest that the motherhood advantage became especially apparent during the 2010s. In 1997, the predicted probabilities of experiencing serious depressive symptoms were slightly higher for mothers whose youngest child was aged 5 to 17 (0.044) than nonmothers (0.040). During the 2007 to 2009 Great Recession, depression rates increased only for nonmothers. In the late 2010s, too, the increase in depression rates was more apparent for nonmothers than for mothers. In 2018, the predicted probability of experiencing serious depressive symptoms climbed up to 0.057 for nonmothers, whereas it was 0.035 for mothers with young children, and 0.037 for mothers whose youngest child was aged 5 to 17.
For self-rated health, from 1997 to 2018, overall, mothers, namely those whose youngest was younger than 5, were less likely than nonmothers to report fair or poor health as opposed to excellent, very good, or good health (Table 6, model 1). The coefficients for the survey year variable suggest an upward trend from 1997 to 2018 in the probability of reporting fair or poor health among American women aged 18 to 59. The interaction terms between year and parental status are almost all nonsignificant (Model 2), suggesting that the parenthood gap in self-rated health changed little over time. Figure 3 plots predicted probabilities of reporting fair or poor health from 1997 to 2018 for the three parental status groups. In 1997, the predicted probabilities of reporting fair or poor health, adjusted to control variables, were 0.095 for nonmothers, 0.060 for mothers whose youngest child was younger than 5, and 0.089 for mothers whose youngest child was aged 5 to 17. In 2018, the probabilities were 0.108 for nonmothers, 0.063 for mothers whose youngest child was younger than 5, and 0.087 for mothers whose youngest child was aged 5 to 17.
Logistic Regression Models Predicting Fair or Poor Health among Women Aged 18 to 59 Years: National Health Interview Survey 1997 to 2018.
Note: Omitted reference categories are no child <18 years in the household, year 1997, parental status × year 1997, White, married, and high school diploma. Data are weighted and adjusted for the design effects. N = 263,110.
p < .05. **p < .01. ***p < .001.
Predicted probability of reporting fair or poor health: U.S. women aged 18 to 59, National Health Interview Survey 1997 to 2018.
To summarize the findings, in the 2000s and 2010s, mothers living with minor children were on average less happy than women without minor children living in the household, other things equal, such as age, race/ethnicity, education, weekly hours of paid work, and relationship status. Nevertheless, depression rates were lower for mothers, namely those whose youngest children were preschool aged, than nonmothers, with the gap widening during the late 2010s. Mothers with young children were also less likely than women without minor children to report fair or poor health.
Discussion
Using data from the GSS and NHIS, this study updates and renews prior knowledge of the parenthood gap in happiness and health in the United States by chronicling the trends from the late 1990s to the late 2010s, using a proper comparison group. The diffusion and the intensification of intensive mothering ideology during the period had led to concerns about its emotional costs for mothers (Furedi 2002; Gunderson and Barrett 2017; Wall 2010). Some studies, however, have painted a more positive picture of mothers’ well-being (Herbst and Ifcher 2016) especially in the 2010s (e.g., Musick et al. 2016). Research on general patterns prior to the pandemic, when mothers’ stressors received a great deal of attention, has been scarce. We find that on average mothers fared better than nonmothers in terms of depressive symptoms and self-rated health in the 2000s and 2010s. Although there are likely some mothers who were overwhelmed by the high standards of time, emotional, and financial investment in raising children set by intensive mothering ideology, other mothers were able to resist the cultural pressure, and still others may have enjoyed the legitimacy of being very involved in children’s lives (Lankes 2022). What determines these differences in mothers’ experiences in the era of intensive mothering culture is beyond the scope of the present analysis. We need future research that investigates explanations for the “motherhood advantage” in mental health and self-rated health during the 2000s and 2010s, and what shielded mothers from worsening mental and physical health that nonparents experienced during that period.
Two specific periods are worth consideration. First, we find that whereas both mothers and nonmothers had increases in depression rates during the 2007 to 2009 Great Recession, mothers recovered more quickly than nonmothers in the subsequent few years. During the Great Recession, the federal government expanded funding to increase Head Start enrollment through the American Recovery and Reinvestment Act of 2009. Some research shows that the availability of Head Start was related to a lower rate of falling into poverty during the Great Recession and a faster recovery from poverty after the recession (Scarborough et al. 2021). Second, consistent with prior research (Kauffman et al. 2021; Twenge et al. 2019), we find that overall depression rates increased throughout the 2010s, but the increase was concentrated on nonmothers. One potential source of such divergent trends in depression rates for mothers and nonmothers from 2010 to 2018 is the decrease in child poverty since 2009 (Burns, Fox, and Wilson 2022). The authors speculate that the expansions to social safety net, such as earned income tax credit, child tax credit, the Supplemental Nutrition Assistance Program, and free school lunches, as the major source for the decline in child poverty. In particular, the sharp decline in child poverty rates from 2019 to 2021 are considered to be due largely to the expansion to child tax credit. Together, these findings call for future research that investigates the role of federal funding to help children during the pandemic, such as COVID-19 stimulus payments and the expansion of the child tax credit as part of the American Rescue Plan Act (U.S. Department of the Treasury 2022), in helping mothers recover from worsened mental health, which was reported by several studies (e.g., Montazer et al. 2022; Zamarro and Prados 2021), during pandemic times.
Mothers were also better off than nonmothers on self-rated health, especially those whose youngest children were younger than school-age, controlling for parents’ age. Research suggests there are health benefits of having children, that is, taking care of one’s health in order to function well for children and prioritizing health as part of role modeling for children (Reczek et al. 2014). These factors may be part of the small advantage we observe over this time period of mothers over nonmothers. Research and public discourses on the parenthood gap in health and well-being have largely focused on mental health, like depression, and subjective well-being, namely happiness, while paying much less attention to physical health as a well-being indicator. More research should examine multiple indicators of health and well-being beyond depression and happiness, including physical health.
We find that mothers with school-age children were less happy “taking all things together” than women in their age groups who did not live with minor children, which is consistent with data from the 1990s and earlier. Mothers’ relative unhappiness compared with nonmothers may be inherent in the question that seemingly asks them to include children’s lives in their calculus of happiness. Mothers may be sizing up “all things together” for themselves and their children, which includes future uncertainties for their offspring that may not impinge on nonmothers as much. Mothers may not be as happy as otherwise they might be if even one child is not doing their best (Milkie and Warner 2014). Another possibility for the continued happiness disadvantage for mothers relative to nonmothers is that having children exacerbates unequal divisions of labor within the household (Yavorsky, Kamp Dush, and Schoppe-Sullivan 2015), which contributes to the lack of leisure time mothers observe comparing themselves with male partners and with women without children (Sayer 2016). Government and workplace policies like guaranteed paid parental leaves and other provisions supporting better quality work conditions could certainly alleviate some of the stressors on mothers. Notably such resources benefit everyone, not just parents (Glass et al. 2016). The lack of support for raising children in the United States may make mothers less happy than nonmothers (Glass et al. 2016). If U.S. mothers did have supports that are extremely common in other countries, they might have been happier than nonmothers, and perhaps even more advantaged in terms of better physical health and lower levels of depression. Although parenthood can bring resources to adults, these tend to be unequal on the basis of individual parents’ socioeconomic characteristics, geographical locations, and other factors that are beyond their will and efforts. To make sure all parents have enough resources to cope with the demands of raising children, they need institutional support, such as paid leave, quality childcare, community recreational facilities, as well as other societal supports that will serve their children in the future such as health care coverage and affordable college tuition.
It may seem contradictory that twenty-first century mothers are slightly more advantaged on the symptom-based (depression) and physical components of well-being but slightly disadvantaged in a subjective well-being measure that inherently has them comparing themselves with others, maybe fathers. This apparent contradiction, and the notable social change that women not living with children are relatively worse off than in the past, are worthy of future research attention. Moreover, the mechanisms through which motherhood status relates to strains and well-being deserve future analysis. Besides the role of the declining trends in child poverty that we mentioned above, selections into parenthood and the growing number of people living alone or outside of coresidential partnerships merit investigation as a source of the new patterns of the parenthood gaps in health and well-being.
The present analysis has limitations. We should note that in both the GSS and the NHIS, parental status was measured as the presence of children younger than 18 in the household, without specifying whether the respondents were primary caregivers of these children. It is also important to keep in mind that some of the nonmothers in the sample were future mothers of minor children and “past” mothers of minor children; they were just nonmothers at the time surveyed—they did not have minor children in the household that year. Both the GSS and the NHIS provide cross-sectional data, which do not allow us to track the same respondents over time. We focused on mothers, in part because we measured parental status on the basis of residential status, and residential fathers are more selective than residential mothers due to unequal patterns of children’s living arrangements after parental separation (Raley and Sweeney 2020). Future research should examine trends in health and well-being among fathers. Finally, although we focus on three important aspects of well-being, including others might provide a more fulsome picture of differences across parents and nonparents. For example, including gaps in the sense of purpose or meaning, for which parents are likely better off compared with nonparents (Nelson et al. 2013), and perhaps increasingly so, might be of interest.
This study provides unique knowledge about the parenthood gap in three important well-being measures prior to the COVID-19 pandemic. Many studies have reported increases in stress and poor mental health among mothers with minor children during the onset of the pandemic, but most of these studies did not have data collected before the pandemic hit. The present findings show that mothers fared better than nonmothers on two measures, depression and self-rated health, before the pandemic and that their advantage increased over the decade prior to the pandemic. Future research is warranted to understand whether and how experiences during the COVID-19 pandemic may have reshaped the trajectories of the health and well-being of women without children in the home, of mothers, and of children.
Footnotes
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research is supported by the Center for Family and Demographic Research, Bowling Green State University, which has core funding from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (P2CHD050959). An earlier version of the article was presented at the 2022 annual meeting of the American Sociological Association in Philadelphia, August 6 to 9.
